{"id": "6c9feedfc3fd2bcf5e98fc45593cfff5ac905d7e207ef8afa93e0893111f7223", "sources": ["arxiv"], "title": "Understanding Key Features of Time Series Foundation Models from Epidemic Forecasting", "abstract": "Seasonal influenza infects millions of people and causes substantial morbidity and mortality in the United States each year, making accurate short-term forecasting a core public-health need. Reliable forecasts of epidemic time series can inform vaccination timing, hospital staffing, and resource allocation, yet the comparative behavior of modern forecasting architectures on infectious-disease surveillance data remains insufficiently characterized. We address this gap through a systematic evaluation of regional influenza forecasting using influenza-like illness surveillance and influenza-associated hospitalization time series under both temporal and spatial generalization settings for 1-4-week-ahead prediction. We compare classical neural network architectures, numerical transformer-based models, pretrained time series foundation models, and LLM-based forecasting approaches. Across tasks, we demonstrate that a mixture-of-experts model that fuses multiple pretrained forecasters achieves the strongest overall performance, indicating that heterogeneous pretrained representations provide complementary predictive information. Our results further show that numerical transformer-based models produce reliable forecasts, while pretraining provides the largest gains at longer horizons, particularly when the pretraining domain is mechanistically aligned with influenza dynamics. In contrast, LLM-based time series methods underperform relative to numerical forecasters in this setting. Finally, we examine hospitalization information as both an auxiliary covariate and a pretraining source. Hospitalization signals provide complementary improvements in selected settings and clarify when additional surveillance streams enhance the robustness of multi-horizon forecasting. These findings provide actionable guidance on model selection, pretraining strategy, and auxiliary-signal use for influenza preparedness.", "authors": ["Alireza Jafari", "Judy Fox", "Geoffrey C. Fox", "Madhav Marathe", "Aniruddha Adiga"], "categories": ["cs.LG"], "fields_of_study": [], "published_date": "2026-06-17", "url": "https://arxiv.org/abs/2606.19560", "pdf_url": "https://arxiv.org/pdf/2606.19560v1", "arxiv_id": "2606.19560", "doi": null, "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.35}
{"id": "0f62e3b10a488c5cfe8f3ce1e36f42cdb897e7274bb0ff99fc9ef4ed29d99a65", "sources": ["arxiv"], "title": "ForecastBench-Sim: A Simulated-World Forecasting Benchmark", "abstract": "Forecasting benchmarks for general-purpose AI systems usually inherit the constraints of the real world: outcomes resolve slowly, tail events are rare, and counterfactual questions are difficult to score. We introduce ForecastBench-Sim, a simulated-world forecasting benchmark built on game rollouts from Freeciv, a turn-based strategy game modelled on the Civilization series. Forecasters receive a fixed world report (a structured snapshot of the current game state) and answer questions about hidden future states; the benchmark then continues the simulation and scores forecasts. Because the world is simulated, the same setup can generate continuous or binary forecasting questions at arbitrary time horizons, paired intervention worlds for conditional or causal questions, and resolved examples of rare or disruptive outcomes. We describe the benchmark pipeline, question families, scoring protocol, and release artifacts, and report validation slices from model evaluations and an anonymized human pilot. ForecastBench-Sim is intended to complement real-world forecasting benchmarks by providing controlled, immediately resolvable tasks for studying probabilistic reasoning under dynamic world states.", "authors": ["Jaeho Lee", "Nick Merrill", "Ezra Karger"], "categories": ["cs.AI", "cs.CL", "cs.LG"], "fields_of_study": [], "published_date": "2026-06-17", "url": "https://arxiv.org/abs/2606.18686", "pdf_url": "https://arxiv.org/pdf/2606.18686v1", "arxiv_id": "2606.18686", "doi": null, "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.35}
{"id": "71eae1cfc0a5f8721ace544da8c3f024ed2d08def95c9cf1f8ed50947982ee43", "sources": ["arxiv"], "title": "Do Time Series Foundation Model Benchmarks Hide Regime-Dependent Failures? Evidence from Traffic Speed Forecasting", "abstract": "Standard benchmarks evaluate time series foundation models (TSFMs) using aggregate metrics, but these can mask severe failures in critical operating regimes. We introduce regime-stratified evaluation and apply it to three TSFMs on two standard traffic speed benchmarks. Traffic exhibits abrupt regime switching between free-flow and congested states, producing bimodal speed distributions during transitions. When we stratify by traffic regime, both accuracy and prediction-interval coverage degrade sharply during transitions: transition-regime MAE reaches 11 mph (versus 3 mph overall), and empirical coverage of 90% prediction intervals drops as low as 55%. These failures are invisible in aggregate metrics because free-flow observations dominate the sample. A simple historical conditional baseline (sampling from per-sensor training distributions) achieves better transition coverage than any TSFM, but has far worse overall accuracy. We propose bimodal mixture augmentation (BMA), a post-hoc method that combines TSFM forecasts with historical distributional knowledge, approaching the historical baseline's transition coverage while preserving the TSFM's accuracy. Our results suggest that TSFM benchmarks should incorporate regime-aware evaluation to surface failures that aggregate metrics hide.", "authors": ["Yingshuo Wang", "Xian Sun", "Lingdong Kong", "Wei Gao", "Yanhang Li", "Zhichao Fan", "Zexin Zhuang"], "categories": ["cs.LG"], "fields_of_study": [], "published_date": "2026-06-16", "url": "https://arxiv.org/abs/2606.18367", "pdf_url": "https://arxiv.org/pdf/2606.18367v1", "arxiv_id": "2606.18367", "doi": null, "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.35}
{"id": "84efca45803599bea40694f1971e37ec7bba670ac632bcd7e36bd7ba32eccb4f", "sources": ["arxiv", "semantic_scholar"], "title": "TimeVista: Exploring and Exploiting Vision-Language Models as Judges for Time Series Forecasting", "abstract": "High-quality time series forecasting is pivotal for real-world decision-making. However, traditional point-wise metrics often fail to reveal complex temporal patterns and align poorly with human intuitive preferences. While the ''LLM-as-a-Judge'' paradigm has revolutionized text evaluation by providing flexible, human-aligned judgment, its application to time series remains largely unexplored. In this paper, we leverage Vision-Language Models (VLMs) as judges for time series forecasting, harnessing their ability to comprehend time series plots grounded in textual information. Specifically, we propose a novel framework integrating micro- and macro-level judgments informed by contextual information to evaluate time series forecasting. To this end, we introduce TimeVista, a comprehensive VLM-as-a-Judge benchmark comprising 5563 time series samples paired with detailed evaluation rubrics. Extensive meta-evaluations demonstrate that VLMs are highly reliable judges, achieving significantly higher consistency with human preferences than conventional metrics. Building upon our benchmark, we comprehensively assess recent Time Series Foundation Models (TSFMs) under the VLM-as-a-Judge paradigm. Our results demonstrate that VLMs serve as robust and interpretable judges, providing a comprehensive, human-aligned standard for evaluating time series models.", "authors": ["Zhi Chen", "Yuxuan Wang", "Jialong Wu", "Yong Liu", "Haoran Zhang", "Xingjian Su", "Jianmin Wang", "Mingsheng Long"], "categories": ["cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2026-06-15", "url": "https://arxiv.org/abs/2606.16173", "pdf_url": "https://arxiv.org/pdf/2606.16173v1", "arxiv_id": "2606.16173", "doi": null, "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.35}
{"id": "4a528ac52154dc185b0cf8fb0896ac6e5c664c4ffa6c42c1e035c35f7150dfc8", "sources": ["arxiv", "semantic_scholar"], "title": "Quantum-classical hybrid models based on error correction for time series forecasting", "abstract": "Time series forecasting largely benefits from combining the strengths of different models, especially using a scheme where a model corrects another model by capturing supplementary patterns from forecasting errors. Concurrently, quantum models are providing a means to augment the classical capacity, including in time series forecasting, by acting alongside classical models in hybrid architectures. In this work, we propose the first forecasting system based on error correction that jointly uses quantum and classical models. Here, quantum models first extract patterns by exploring quantum phenomena, and classical models capture the remaining patterns from the quantum errors. Compared to classical single models and classical-classical hybrid models based on error correction, the complementary capacity that emerges from this quantum-classical system provided the best results in most of the addressed problems. Therefore, this work paves the way to introduce quantum models in established hybridization schemes for time series forecasting.", "authors": ["Jonathan H. A. de Carvalho", "Filipe C. de L. Duarte", "Fernando M. de Paula Neto", "Paulo S. G. de Mattos Neto"], "categories": ["quant-ph", "cs.LG"], "fields_of_study": ["Physics", "Computer Science"], "published_date": "2026-06-13", "url": "https://arxiv.org/abs/2606.15213", "pdf_url": "https://arxiv.org/pdf/2606.15213v1", "arxiv_id": "2606.15213", "doi": null, "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.35}
{"id": "8e31223be06f17e56eef817edb064faf2f7ae673f1dafd293526305187a25676", "sources": ["arxiv", "semantic_scholar"], "title": "Semantics-Enhanced Retrieval-Augmented Time Series Forecasting", "abstract": "Time series forecasting models often benefit from historical patterns. Inspired by Retrieval-Augmented Generation (RAG), recent research explored retrieving relevant historical time series segments to enhance forecasting. However, relying solely on time series similarity is often insufficient for retrieval under non-stationarity. To address this, we propose a multimodal approach: a \\textbf{S}emantics-\\textbf{E}nhanced \\textbf{R}etrieval-\\textbf{A}ugmented Time Series \\textbf{F}orecasting framework, SERAF. Unlike mainstream approaches that depend only on time series similarity, SERAF conducts dual retrieval over the time series and their self-generated textual descriptions. It retrieves two complementary sets of historical patterns and corresponding futures, which are selectively and jointly used to guide future predictions. Experiments across seven real-world datasets demonstrate the effectiveness of SERAF in bridging numerical and semantic views of time series compared with state-of-the-art baselines.", "authors": ["Shiqiao Zhou", "Zipeng Wu", "Holger Schöner", "Edouard Fouché", "IAG Wilson", "Shuo Wang"], "categories": ["cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2026-06-12", "url": "https://arxiv.org/abs/2606.14941", "pdf_url": "https://arxiv.org/pdf/2606.14941v1", "arxiv_id": "2606.14941", "doi": null, "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.35}
{"id": "51523f42010f1fcec9caf40068f2a3a388bc2159ba442d9efdd26a2c6ba25ea2", "sources": ["arxiv", "semantic_scholar"], "title": "APEX: A Network-Native Time-Series Foundation Model for Forecasting and Anomaly Detection for Wireless Edge Operations", "abstract": "Generic time-series foundation models transfer poorly to wireless network telemetry whose signals are bursty, zero-inflated, and coupled across protocol layers. We present APEX, a network-native, decoder-only transformer for forecasting enterprise AP telemetry, and evaluate it on DHCP degradation as a representative network task. APEX is pre-trained on 10-channel multivariate telemetry from ~4,500 production wireless networks (~100K AP time series, 34 metrics per AP), and is available as APEX-Large (269M, cloud) and APEX-Edge (10.5M, edge). On a 192-step (4-day) DHCP degradation benchmark, APEX-Large reduces MAE by 18% over the strongest foundation-model baseline (Toto) and 38% over SARIMA, with anomaly-detection F1 = 0.93, while APEX-Edge enables sub-second, privacy-preserving inference on AP-class edge hardware. These results suggest network-native pre-training is a practical foundation for proactive wireless operations.", "authors": ["Swadhin Pradhan", "Niloo Bahadori", "Peiman Amini"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2026-06-10", "url": "https://arxiv.org/abs/2606.11553", "pdf_url": "https://arxiv.org/pdf/2606.11553v1", "arxiv_id": "2606.11553", "doi": null, "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.35}
{"id": "8088d690c5e41e02896fb1339943e6b6c8e7326b3d440b93e8120bc11cb6cd36", "sources": ["arxiv", "semantic_scholar"], "title": "CITRAS-FM: Tiny Time Series Foundation Model for Covariate-Informed Zero-Shot Forecasting", "abstract": "Pretrained time series foundation models (TSFMs) have enabled zero-shot forecasting on unseen target series. However, existing TSFMs often incur high computational cost and provide limited support for diverse variable types, often failing to account for covariates that exogenously influence target variability. To address these challenges, we propose CITRAS-FM, a tiny 7M-parameter TSFM that supports univariate, multivariate, and covariate-informed zero-shot forecasting with real-time CPU inference. Built on a patch-based, decoder-only Transformer, CITRAS-FM introduces Shifted Attention into the cross-variate module to effectively exploit known covariates accessible throughout the forecast horizon. Moreover, to enable covariate-aware pretraining despite the scarcity of covariate-rich corpora, we propose CovSynth, which synthesizes realistic covariates from decomposed components of target series. Experiments on fev-bench, spanning 100 tasks across various settings, demonstrate that CITRAS-FM achieves state-of-the-art zero-shot accuracy among sub-10M TSFMs while delivering sub-0.1-second CPU inference, offering a strong balance between forecasting accuracy and real-time deployability.", "authors": ["Yosuke Yamaguchi", "Issei Suemitsu", "Yuki Kajihara", "Wenpeng Wei"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2026-06-09", "url": "https://arxiv.org/abs/2606.10798", "pdf_url": "https://arxiv.org/pdf/2606.10798v1", "arxiv_id": "2606.10798", "doi": null, "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.35}
{"id": "338b17ff3356047227f22accabe19c2a1249a3343fbee77f6c8a3c9f827b3705", "sources": ["arxiv", "semantic_scholar"], "title": "Does Normalization Choice Matter for Causal Large Time-Series Models?", "abstract": "Large models for time-series forecasting have been emerged as a promising paradigm for training models on heterogeneous collections of signals. These models typically rely on causal autoregressive architectures, where each observation is sequentially predicted from past. In practice, real-world time-series exhibit non-stationarities, which significantly influence predictive performance. To mitigate this, normalization is commonly employed. However, in efficient causal settings it might induce information leakage from future observations during training. Recent alternatives, including causal normalization and statistics computed from initial observations, have been proposed to address this issue, but their practical implications remain insufficiently understood. In this work, we evaluate normalization strategies for transformer-based large time-series models trained with patching and efficient causal strategy. We showcase that normalization choice significantly influences both training convergence and forecasting performance.", "authors": ["Samy-Melwan Vilhes", "Gilles Gasso", "Mokhtar Z Alaya"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2026-06-08", "url": "https://arxiv.org/abs/2606.09954", "pdf_url": "https://arxiv.org/pdf/2606.09954v1", "arxiv_id": "2606.09954", "doi": null, "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "ICLR 2026 Workshop: Time Series in the Age of Large Models, Apr 2026, Rio De Janeiro, Brazil", "quality_score": 0.55}
{"id": "235d0c175741471c5b415ad1bb73e7c101c55a1cd39cedda001777c5de06d76c", "sources": ["arxiv", "semantic_scholar"], "title": "Mix, Don't Pick: Why Synthetic Corpus Composition Matters for Time Series Foundation Model Pretraining", "abstract": "Choosing the wrong synthetic generator for time-series foundation model pretraining is costly: under identical training budgets, the best and worst generators produce up to a $2\\times$ gap in forecasting error, yet the field has no principled way to make this choice. The problem is compounded by the fact that generator rankings are not stable across architectures: across 11 generator families evaluated on Chronos-T5-Mini and Moirai-Small trained from scratch, we find that which generators are useful depends on the model architecture. Rather than solving the generator selection problem, we sidestep it: a simple equal-weight mixture of all generators matches or beats the best individual generator for both architectures, and composing this mixture with real data yields the strongest pretraining corpora overall. Synthetic pretraining is therefore a corpus composition problem, not a generator selection problem, and composition choices should be validated per model family rather than assumed to transfer.", "authors": ["Aaryan Nagpal", "Debdeep Sanyal", "Murari Mandal", "Dhruv Kumar", "Saurabh Deshpande"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2026-06-06", "url": "https://arxiv.org/abs/2606.09912", "pdf_url": "https://arxiv.org/pdf/2606.09912v1", "arxiv_id": "2606.09912", "doi": null, "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.35}
{"id": "a747331de032e2b3000c2385a2fac8e30e62765a27eebdec78b78fba179ccb06", "sources": ["arxiv", "semantic_scholar"], "title": "Time series Foundation Models based on Physics-Informed Synthetic Histories for Cold-Start Photovoltaic Forecasting", "abstract": "At commissioning time, Photovoltaic (PV) operators must forecast production before target-site observations are available, limiting the direct use of standard supervised forecasters. This cold-start setting is addressed with a zero-shot pipeline that generates a synthetic production history from plant metadata and meteorological covariates, enabling time-series foundation models (TSFMs) to forecast through inference-time conditioning. Five TSFMs are benchmarked against classical baselines under strict Cold-Start Baseline, Real Feedback, and Self-Forecast Feedback strategies. The evaluation spans $440$ PV sites across four datasets and diverse climate regimes. Covariate-aware foundation models outperform baselines by approximately $1.7-2\\times$: TabPFN-TS achieves the lowest error under Real Feedback (MAE $0.514$, RMSE $0.721$ $kWh$ ${kWp}^{-1}$ ${d}^{-1}$), while Chronos-2 is most robust under Self-Forecast Feedback. Performance is largely insensitive to the synthetic-history source, indicating that accuracy is driven more by the availability of plausible temporal context than by the specific generator.", "authors": ["Lorenzo Longarini", "Alessandro Rongoni", "Simone Silenzi", "Emanuele Frontoni", "Riccardo Rosati"], "categories": ["cs.LG", "eess.SP", "stat.ML"], "fields_of_study": ["Computer Science", "Engineering", "Mathematics"], "published_date": "2026-06-05", "url": "https://arxiv.org/abs/2606.07457", "pdf_url": "https://arxiv.org/pdf/2606.07457v1", "arxiv_id": "2606.07457", "doi": null, "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.35}
{"id": "23b3adf0a3090f80017f25e8d8d9b835f2752e86425120830ee7c49ebb816084", "sources": ["arxiv", "semantic_scholar"], "title": "GlucoFM-Bench: Benchmarking Time-Series Foundation Models for Blood Glucose Forecasting", "abstract": "Blood glucose forecasting models are foundational for modern diabetes management systems, as reliable short-term predictions can enable proactive interventions, support automated insulin delivery, and reduce the risk of hypo- and hyperglycemic events. From a modeling perspective, glucose forecasting poses unique challenges due to heterogeneous physiological dynamics across diabetes populations. Traditional machine learning and deep learning models have been extensively evaluated for glucose prediction, yet recent time-series foundation models (TSFMs) remain much less studied in this setting. To bridge this gap, we present GlucoFM-Bench, a comprehensive benchmark evaluating state-of-the-art TSFMs alongside supervised deep learning models for blood glucose forecasting. We assess eight representative architectures, including pre-trained TSFMs, time-series large language models, and task-specific deep learning models, across 15 publicly available diabetes-relevant datasets comprising 1,117 individuals with type 1 diabetes, type 2 diabetes, prediabetes, and no diabetes. Models are evaluated under zero-shot, few-shot, and full-shot protocols, with systematic variation in context length and prediction horizon. Across datasets, pre-trained TSFMs, especially Chronos-2 and TimesFM, show strong zero-shot and few-shot transfer, with the best zero-shot model performing within 5% of the best full-shot supervised model. Yet, when task-specific data are abundant, a lightweight LSTM remains strongest, outperforming TSFMs by 4--21% under full-shot training. Stratified analyses reveal persistent challenges in T1D cohorts and hypo-/hyperglycemic ranges, highlighting the need for evaluation beyond aggregate error metrics. Together, GlucoFM-Bench provides a standardized and reproducible foundation for evaluating, comparing, and improving foundation models for blood glucose forecasting.", "authors": ["Baiying Lu", "Zhaohui Liang", "Ryan Pontius", "Shengpu Tang", "Temiloluwa Prioleau"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2026-06-05", "url": "https://arxiv.org/abs/2606.06881", "pdf_url": "https://arxiv.org/pdf/2606.06881v1", "arxiv_id": "2606.06881", "doi": null, "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.35}
{"id": "eb11c82d98ce0c13ee337edf97a2693116969666b120b5f336a40d3558faa8ee", "sources": ["arxiv", "semantic_scholar"], "title": "TS-ICL: A Flexible Time-Indexed Foundation Model for Time Series via In-Context Learning", "abstract": "Foundation models mark a profound paradigm shift in time series modeling, with task-specific models being superseded by general-purpose zero-shot models. Yet, current approaches primarily focus on forecasting, while real-world time series are often irregularly and partially observed, requiring models that can jointly forecast, impute missing values, and handle degraded sampling conditions. To address these challenges, we introduce TS-ICL, a novel probabilistic In-Context Learning encoder--regressor Transformer that unifies forecasting and imputation. TS-ICL formulates time series tasks as timestamp-aligned regression and naturally incorporates covariates by training on synthetic dependency structures generated from a novel causal data prior. Empirically, TS-ICL achieves a new state-of-the-art in imputation, while remaining competitive with leading forecasting foundation models across both univariate and covariate-aware benchmarks. It shows particularly strong performance in forecasting with partially observed look-back windows.", "authors": ["Etienne Le Naour", "Tahar Nabil", "Adrien Petralia"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2026-06-04", "url": "https://arxiv.org/abs/2606.05878", "pdf_url": "https://arxiv.org/pdf/2606.05878v2", "arxiv_id": "2606.05878", "doi": null, "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.35}
{"id": "d198453f3a21e5f3765bc67b8838c8038da4117b5efce210f803d4e8195457c6", "sources": ["arxiv", "semantic_scholar"], "title": "TimeBlocks: Foundational and Continual Time-Series Blockbase -- Extended Version", "abstract": "The ongoing digitization has led to a proliferation of time-series data streams that monitor a variety of processes, from which valuable insights may be obtained. Further, the emergence of successful foundational language models begs the question of whether it is possible to achieve time-series models with the foundational properties of handling multiple tasks, while being sufficiently lightweight to allow real-time data stream processing. Existing foundational time-series models are often large and only effective in offline settings without stringent time and computational constraints, and where repeated model calibration is not needed. However, when applied to data streams, these models are ineffective due to their size and lack of support for continual calibration, which compromise their ability to deliver accurate real-time responses, their durability, and their deployability in hardware-limited settings. We propose TimeBlocks to enable versatile time-series processing by facilitating the efficient building of lightweight models suitable for multiple tasks under variable conditions. In particular, the method maintains a pool of interchangeable and modular model blocks that can be used to construct new time-series models. When presented with specific time-series data, a routing strategy iteratively selects the most suitable blocks to construct a lightweight and accurate model for the data. We equip TimeBlocks with a method called StreamCore to build a representative small subset of the data stream, which preserves a guaranteed approximation of the stream over time, enabling continual model calibration. An experimental study on multiple data sets and covering multiple tasks shows that TimeBlocks enables to build models capable of outperforming existing baselines.", "authors": ["David Campos", "Bin Yang", "Tung Kieu", "Lei Chen", "Chenjuan Guo", "Christian S. Jensen"], "categories": ["cs.LG", "cs.DB"], "fields_of_study": ["Computer Science"], "published_date": "2026-06-01", "url": "https://arxiv.org/abs/2606.02142", "pdf_url": "https://arxiv.org/pdf/2606.02142v1", "arxiv_id": "2606.02142", "doi": null, "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.35}
{"id": "33e34abccdbaaa3dfcb78a5a9c9d1c7b1e18961dfefeccc40708f4dfd166dba4", "sources": ["arxiv", "semantic_scholar"], "title": "Time Series as Language: A Universal Tokenizer for General-Purpose Time Series Foundation Models", "abstract": "While Next-Token Prediction (NTP) has unified LLM pretraining, its adaptation to unbounded, continuous time series (TS) remains open. To bridge the gap, we introduce UniTok, a universal tokenizer that transforms TS into discrete tokens, and UniTok-FM, a foundation model pretrained via NTP on these tokens. UniTok-FM is a general-purpose foundation model that supports zero-shot and prompt-boosted forecasting, as well as few-shot generation and classification via training-free in-context inference--a capability not achieved by prior works. Technically, UniTok is a vector-quantized autoencoder incorporating prefix normalization for scale stabilization, a progressive-resolution causal architecture for encoding and decoding, and a structure-preserving reconstruction loss for training. UniTok-FM adopts an off-the-shelf LLM architecture without TS-specific modifications. Instead of pretraining on isolated TS, it performs NTP on context windows formed by multiple series with similar patterns, aiming to capture their shared dynamics. Experiments on forecasting, generation, and classification show that a single unified UniTok-FM consistently outperforms statistical and supervised baselines, achieves competitive performance with task-specific foundation models, and uniquely enables training-free in-context inference across tasks.", "authors": ["Yunhao Zhang", "Ruiying Qi", "Jiale Zheng", "Jianfeng Zhang", "Lujia Pan", "Junchi Yan"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2026-05-31", "url": "https://arxiv.org/abs/2606.09861", "pdf_url": "https://arxiv.org/pdf/2606.09861v1", "arxiv_id": "2606.09861", "doi": null, "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.35}
{"id": "9e2edd3a20c3a730700b93873a85bbbdd9f05766af38fca8a5891b7b01134163", "sources": ["arxiv", "semantic_scholar"], "title": "KairosAgent: Agentic Time Series Forecasting with Fused Semantic Reasoning", "abstract": "Cross-domain multimodal time series forecasting is a challenging task, requiring models to integrate precise numerical comprehension, cross-domain semantic understanding, and effective multimodal fusion. Existing approaches either build Time Series Foundation Models (TSFMs) from scratch or leverage pretrained Large Language Models (LLMs). However, TSFMs often overlook semantic understanding and lack the ability to perform future-oriented semantic reasoning, and LLMs struggle with numerical comprehension and accurate quantitative forecasting. To overcome these limitations, we propose KairosAgent, a novel agentic framework for multimodal time series forecasting, including an LLM-based reasoner and a TSFM-based forecaster. KairosAgent unifies textual reasoning and numerical forecasting by dynamically invoking analytical tools to enhance the numerical understanding and semantic reasoning capabilities of LLMs. The reasoning results are subsequently fused into the TSFM pipeline, enabling more accurate and reliable future predictions. To further improve the reasoning, we curate a large-scale corpus of high-quality trajectories, alongside a reinforcement learning from forecasting paradigm with multi-turn refinement and turn-level credit assignment. Experiments demonstrate that KairosAgent achieves superior zero-shot forecasting performance while maximizing the utility of pretrained LLMs and TSFMs, presenting a promising direction for efficient and interpretable time series agents. The project page is at https://foundation-model-research.github.io/KairosAgent .", "authors": ["Kun Feng", "Ziwei Shan", "Yuchen Fang", "Yiyang Tan", "Sihan Lu", "Shuqi Gu", "Lintao Ma", "Xingyu Lu", "Kan Ren"], "categories": ["cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2026-05-28", "url": "https://arxiv.org/abs/2605.30002", "pdf_url": "https://arxiv.org/pdf/2605.30002v1", "arxiv_id": "2605.30002", "doi": null, "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.35}
{"id": "e4b2536b0fe4f721202b1ac28b188887b92fcec60eec13ffeac945e8035a6138", "sources": ["arxiv", "semantic_scholar"], "title": "AME-TS: Anchored Mixture-of-Experts for Time Series Forecasting", "abstract": "Time series forecasting models are increasingly scaled through large Transformer backbones, yet most existing approaches process all series through a shared dense computation path despite substantial heterogeneity in temporal structure. Mixture-of-Experts (MoE) offers a natural alternative by enabling conditional computation, but standard MoE routing leaves expert specialization weakly identified and often unstable during downstream adaptation. We propose AME-TS, a structure-guided sparse time series foundation model that aligns expert routing with interpretable temporal structure. AME-TS first uses a lightweight regime predictor to estimate series-level descriptors, including forecastability, seasonality, trend, and sparsity, and maps them to a soft structural prior over experts. This series-level prior guides token-level routing during training, encouraging structure-aligned specialization. On the GIFT-Eval benchmark, AME-TS delivers a strong accuracy-efficiency tradeoff across model scales: it substantially outperforms existing time series foundation models at small model scales and remains competitive with the strongest models at larger scales, while activating substantially fewer parameters through sparse routing. We further show that AME-TS learns more interpretable routing geometry and substantially more stable expert specialization than standard MoE during fine-tuning on the M5 dataset. These results suggest that structure-aware routing is an effective and reliable way to realize the benefits of sparse expert models for time series forecasting.", "authors": ["Rui Wang", "Renhao Xue", "Ray Razi", "Huan Song", "Hannah R. Marlowe"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2026-05-24", "url": "https://arxiv.org/abs/2605.25166", "pdf_url": "https://arxiv.org/pdf/2605.25166v1", "arxiv_id": "2605.25166", "doi": null, "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.35}
{"id": "816bfc39be3b916acf16044a560eb8a681f21766318e82d5a9b1e9d094bd5761", "sources": ["arxiv", "semantic_scholar"], "title": "Assessing the Operational Viability of Foundation Models for Time Series Forecasting", "abstract": "Time series forecasting drives operational decisions in areas like finance, transportation, and energy. While supervised learning approaches achieve strong performance, they require domain-specific training, feature engineering, and ongoing maintenance. Large-scale foundation models have recently emerged as a zero-shot alternative, avoiding task-specific training much like LLMs. In this work, we evaluate foundation models against standard supervised approaches. Rather than focusing solely on aggregate accuracy, we analyze performance across four operational regimes: periodic human-centric systems, physically constrained processes, stochastic financial markets, and heterogeneous demand forecasting. Our results characterize optimal deployment areas. Foundation models perform well in domains with transferable periodic structures and are efficient for cold-start or long-tail scenarios. Conversely, supervised specialists maintain higher precision in systems governed by strict physical constraints. In financial domains, newer foundation models are rapidly closing the performance gap with supervised specialists. We further quantify trade-offs in inference latency, data drift adaptability, and deployment constraints. Finally, we propose a Complexity Router that assigns each series to the optimal model class using empirical features. We demonstrate that this selective routing achieves higher accuracy and significantly lower inference costs compared to deploying a universal foundation model, providing a practical framework for balancing generalization and efficiency.", "authors": ["Kavin Soni", "Debanshu Das", "Vamshi Guduguntla"], "categories": ["cs.LG", "cs.AI", "stat.AP", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2026-05-23", "url": "https://arxiv.org/abs/2605.24381", "pdf_url": "https://arxiv.org/pdf/2605.24381v1", "arxiv_id": "2605.24381", "doi": null, "citation_count": 0, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/kavin-soni/timeseries-zeroshot-eval", "venue": null, "quality_score": 0.65}
{"id": "5d614b06036759653547d79410ec8970a9efaff83306926ed4955677c93430fc", "sources": ["arxiv"], "title": "Chronicle: A Multimodal Foundation Model for Joint Language and Time Series Understanding", "abstract": "Real-world time series come with text: metadata, descriptions, news, reports. Yet time series foundation models process numerical sequences in isolation, and the multimodal text-and-time-series models that attempt to bridge the two all adapt a pretrained language model post hoc, inheriting representations shaped without ever seeing temporal data. These models are also evaluated almost exclusively against other multimodal baselines, not against the strongest unimodal foundation models in either domain, leaving open whether joint training is needed at all. We present Chronicle, a compact 324M-parameter decoder-only transformer trained from scratch on natural language and time series within a single unified architecture. Both modalities share the same transformer blocks, attention mechanism, and residual stream; the bulk of pretraining uses unimodal batches so cross-modal capability emerges purely from shared parameters, with a short alignment stage that interleaves the two. To our knowledge, Chronicle is the first model jointly pretrained on text and time series from scratch, and the first multimodal model evaluated against dedicated foundation models in both domains. It matches Gemma-3-270M-PT on 19 NLU tasks, sets a new bar for frozen-embedding time series classification on 24 UCR/UEA datasets, and produces multimodal forecasts on Time-MMD that beat every supervised fusion baseline, all from a single backbone.", "authors": ["Paul Quinlan", "Jeremy Levasseur", "Qingguo Li", "Xiaodan Zhu"], "categories": ["cs.LG", "cs.AI", "cs.CL"], "fields_of_study": [], "published_date": "2026-05-18", "url": "https://arxiv.org/abs/2605.20268", "pdf_url": "https://arxiv.org/pdf/2605.20268v1", "arxiv_id": "2605.20268", "doi": null, "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.35}
{"id": "96eba68f54fe844378f53c7f54d1eadffd814011a9e5839a3a50edb790a04f76", "sources": ["arxiv", "semantic_scholar"], "title": "Empirical evaluation of Time Series Foundation Models for Day-ahead and Imbalance Electricity Price Forecasting in Belgium", "abstract": "Recent advances in Time Series Foundation Models (TSFMs) promise zero-shot forecasting capabilities with minimal task-specific training. While these models have shown strong performance across generic benchmarks, their applicability in volatile, complex electricity markets remains underexplored. Addressing this gap, this study provides a systematic empirical evaluation of several TSFMs, specifically Chronos-2 and Chronos-Bolt (developed by Amazon), and TimesFM 2.5 (provided by Google), for forecasting Belgian day-ahead and imbalance electricity prices. For both considered markets, Chronos-2 in ARX mode produces the most accurate forecasts. Compared with the best ensemble prediction from other machine learning methods, Chronos-2's Mean Absolute Error (MAE) is 5% lower for the day-ahead market. In contrast, the model yields 10% higher MAE predicting imbalance prices across all forecast horizons, except for the two-hour-ahead horizon. Moreover, we find that TSFMs exhibit genuine zero-shot forecasting skills but still struggle under extreme market conditions.", "authors": ["Chi Bui", "Maria Margarida Mascarenhas", "Arnaud Verstraeten", "Hussain Kazmi"], "categories": ["eess.SY", "cs.LG"], "fields_of_study": ["Engineering", "Computer Science"], "published_date": "2026-05-16", "url": "https://arxiv.org/abs/2605.17045", "pdf_url": "https://arxiv.org/pdf/2605.17045v1", "arxiv_id": "2605.17045", "doi": null, "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.35}
{"id": "4d47028b83a40d304b240a66abfd251dc54660ac5f485f8f02dcbda8592f01b8", "sources": ["arxiv", "semantic_scholar"], "title": "EHR-RAGp: Retrieval-Augmented Prototype-Guided Foundation Model for Electronic Health Records", "abstract": "Electronic Health Records (EHR) contain rich longitudinal patient information and are widely used in predictive modeling applications. However, effectively leveraging historical data remains challenging due to long trajectories, heterogeneous events, temporal irregularity, and the varying relevance of past clinical context. Existing approaches often rely on fixed windows or uniform aggregation, which can obscure clinically important signals. In this work, we introduce EHR-RAGp, a retrieval-augmented foundation model that dynamically integrates the most relevant patient history across diverse clinical event types. We propose a prototype-guided retrieval module that acts as an alignment mechanism and estimates the relevance of retrieved historical chunks with respect to a given prediction task, guiding the model towards the most informative context. Across multiple clinical prediction tasks, EHR-RAGp consistently outperforms state-of-the-art EHR foundation models and transformer-based baselines. Furthermore, integrating EHR-RAGp with existing clinical foundation models yields substantial performance gains. Overall, EHR-RAGp provides a scalable and efficient framework for leveraging long-range clinical context to improve downstream performance.", "authors": ["Saeed Shurrab", "Mariam Al-Omari", "Dana El Samad", "Farah E. Shamout"], "categories": ["cs.IR", "cs.AI", "cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2026-05-12", "url": "https://arxiv.org/abs/2605.12335", "pdf_url": "https://arxiv.org/pdf/2605.12335v1", "arxiv_id": "2605.12335", "doi": null, "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.35}
{"id": "e3f48b98fdd1bbc2c2857608fabbd634a35dbb7a70a4340d8d15802143e361f1", "sources": ["arxiv", "semantic_scholar"], "title": "Benchmarking Transformer and xLSTM for Time-Series Forecasting of Heat Consumption", "abstract": "Obtaining an accurate short-term forecasting for heat demand is an essential part of operating district heating networks cost-efficient and reliable. Heat consumption time series at the building level are highly dependent on exogenous variables such as outdoor temperature and individual usage patterns, making forecasting in this context a challenging task. Thus, this paper benchmarks novel Transformer-based and xLSTM architectures for short-term heat-demand forecasting. Using hourly data from 25 German buildings (2017-2025), we compare three-hour and 24-hour forecasting horizons relevant for intraday control and day-ahead scheduling. We establish a multi-building benchmark that tests whether models trained on pooled, heterogeneous building data are able to generalize across diverse building stock. The results show that the xLSTM achieves the lowest RMSE (19.88 kWh for three-hour, 21.47 kWh for 24-hour forecasts), while the Temporal Fusion Transformer attains the best MAE (9.16 kWh for three-hour forecasts). As xLSTMs and Transformers require long training times and have a huge number of trainable parameters, their sustainability remains questionable. Therefore, this paper further investigates the trade-off between predictive accuracy and computational resource demand of the evaluated forecasting models. The findings indicate that also low-parameter models like a traditional fully-connected network achieve good predictive results, highlighting that marginal accuracy gains of the novel prediction models come at substantial resource expense for this use case.", "authors": ["Marja Wahl", "Daniel R. Bayer", "Sven Rausch", "Marco Pruckner"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2026-05-10", "url": "https://arxiv.org/abs/2605.09722", "pdf_url": "https://arxiv.org/pdf/2605.09722v1", "arxiv_id": "2605.09722", "doi": "10.1109/SusTech67720.2026.11536443", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "IEEE Conference on Technologies for Sustainability", "quality_score": 0.55}
{"id": "48fb1a04cfec538439d88a9420f88d9f66df4a1053e6897dddb6060cc4033852", "sources": ["arxiv", "semantic_scholar"], "title": "Multivariate Financial Forecasting using the Chronos Time Series Foundation Models", "abstract": "Using Chronos-2, an open-source time-series foundation model, we evaluate pretrained time-series models for economic and financial forecasting with an emphasis on whether multivariate (MV) inputs improve accuracy relative to univariate (UV) baselines. The study covers two panels -- the Magnificent-7 equities and U.S. Treasury interest rates -- as well as a combined panel, using rolling monthly evaluations from 2000--2025. We vary input window lengths and forecast horizons and report RMSE and MAPE. Across datasets, MV forecasts consistently outperform UV forecasts, with especially strong gains for interest rates and meaningful improvements for equities. Series-level comparisons show MV improvements in every case, and error dispersion is generally lower under MV inputs. We also provide parameter-heatmap and time-series visualizations. However, mixing time series across equity and interest rate markets reduces forecast accuracy, indicating that adding noisy context degrades model performance. Overall, the results indicate that foundation models can leverage cross-series information to improve forecast accuracy in finance, and that the benefits are strongest when related series are modeled jointly under disciplined rolling protocols. Other than using an open-source foundation model, this paper also showcases how AI may be used for financial research.", "authors": ["Sanjiv R Das", "Tarang Goyal", "Mohini Yadav"], "categories": ["q-fin.ST", "cs.AI"], "fields_of_study": ["Economics", "Computer Science"], "published_date": "2026-05-08", "url": "https://arxiv.org/abs/2605.21504", "pdf_url": "https://arxiv.org/pdf/2605.21504v1", "arxiv_id": "2605.21504", "doi": null, "citation_count": 0, "influential_citation_count": 0, "has_code": true, "code_url": null, "venue": null, "quality_score": 0.65}
{"id": "a45dff8ddca734b1b4f792e41c4e78b95d040527068a14536c5d96b278068aa0", "sources": ["arxiv", "semantic_scholar"], "title": "Superposition Is Not Necessary: A Mechanistic Interpretability Analysis of Transformer Representations for Time Series Forecasting", "abstract": "Transformer architectures have been widely adopted for time series forecasting, yet whether the representational mechanisms that make them powerful in NLP actually engage on time series data remains unexplored. The persistent competitiveness of simple linear models such as DLinear has fueled ongoing debate, but no mechanistic explanation for this phenomenon has been offered. We address this gap by applying sparse autoencoders (SAEs), a tool from mechanistic interpretability, to probe the internal representations of PatchTST. We first establish that a single-layer, narrow-dimensional transformer matches the forecasting performance of deeper configurations across commonly used benchmarks. We then train SAEs on the post-GELU intermediate FFN activations with dictionary sizes ranging from 0.5x to 4.0x the native dimensionality. Expanding the dictionary yields negligible downstream performance change (average 0.214%), with large portions of overcomplete dictionaries remaining inactive. Targeted causal interventions on dominant latent features produce minimal forecast perturbation. Across all evaluated settings, we observe no empirical evidence that the analyzed FFN representations rely on strong superposition. Instead, the representations remain sparse, stable under aggressive dictionary expansion, and largely insensitive to latent interventions. These results demonstrate that superposition is not necessary for competitive performance on standard forecasting benchmarks, suggesting they may not demand the rich compositional representations that drive transformer success in language modeling, and helping explain the persistent competitiveness of simple linear models", "authors": ["Alper Yıldırım"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2026-05-06", "url": "https://arxiv.org/abs/2605.05151", "pdf_url": "https://arxiv.org/pdf/2605.05151v1", "arxiv_id": "2605.05151", "doi": null, "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.35}
{"id": "9d82239cda9829f10466586c6f12f894193745919af62e50a5cbff830bb30a44", "sources": ["arxiv", "semantic_scholar"], "title": "Explainable Load Forecasting with Covariate-Informed Time Series Foundation Models", "abstract": "Time Series Foundation Models (TSFMs) have recently emerged as general-purpose forecasting models and show considerable potential for applications in energy systems. However, applications in critical infrastructure like power grids require transparency to ensure trust and reliability and cannot rely on pure black-box models. To enhance the transparency of TSFMs, we propose an efficient algorithm for computing Shapley Additive Explanations (SHAP) tailored to these models. The proposed approach leverages the flexibility of TSFMs with respect to input context length and provided covariates. This property enables efficient temporal and covariate masking (selectively withholding inputs), allowing for a scalable explanation of model predictions using SHAP. We evaluate two TSFMs - Chronos-2 and TabPFN-TS - on a day-ahead load forecasting task for a transmission system operator (TSO). In a zero-shot setting, both models achieve predictive performance competitive with a Transformer model trained specifically on multiple years of TSO data. The explanations obtained through our proposed approach align with established domain knowledge, particularly as the TSFMs appropriately use weather and calendar information for load prediction. Overall, we demonstrate that TSFMs can serve as transparent and reliable tools for operational energy forecasting.", "authors": ["Matthias Hertel", "Alexandra Nikoltchovska", "Sebastian Pütz", "Ralf Mikut", "Benjamin Schäfer", "Veit Hagenmeyer"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2026-04-30", "url": "https://arxiv.org/abs/2604.28149", "pdf_url": "https://arxiv.org/pdf/2604.28149v1", "arxiv_id": "2604.28149", "doi": "10.1145/3744255.3811724", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.35}
{"id": "74819adb3eb194673d10944f2c79a8cad2c5672ad301233d4ecb97cb761096f2", "sources": ["arxiv", "semantic_scholar"], "title": "FETS Benchmark: Foundation Models Outperform Dataset-specific Machine Learning in Energy Time Series Forecasting", "abstract": "Driven by the transition towards a climate-neutral energy system, accurate energy time series forecasting is critical for planning and operation. Yet, it remains largely a dataset-specific task, requiring comprehensive training data, limiting scalability, and resulting in high model development and maintenance effort. Recently, foundation models that aim to learn generalizable patterns via extensive pretraining have shown superior performance in multiple prediction tasks. Despite their success and strong potential to address challenges in energy forecasting, their application in this domain remains largely unexplored. We address this gap by presenting the Foundation Models in Energy Time Series Forecasting (FETS) benchmark. We (1) provide a structured overview of energy forecasting use cases along three main dimensions: stakeholders, attributes, and data categories; (2) collect and analyze 54 datasets across 9 data categories, guided by typical stakeholder interests; (3) benchmark foundation models against classical machine learning approaches across different forecasting settings. Foundation models consistently outperform dataset-specific optimized machine learning approaches across all settings and data categories, despite the latter having seen the full historic target data during training. In particular, covariate-informed foundation models achieve the strongest performance. Further analysis reveals a strong correlation between predictive performance and spectral entropy, performance saturation beyond a certain context length, and improved performance at higher aggregation levels such as national load, district heating, and power grid data. Overall, our findings highlight the strong potential of foundation models as scalable and generalizable forecasting solutions for the energy domain, particularly in data-constrained and privacy-sensitive settings.", "authors": ["Marco Obermeier", "Marco Pruckner", "Florian Haselbeck", "Andreas Zeiselmair"], "categories": ["cs.LG", "cs.AI", "cs.CE"], "fields_of_study": ["Computer Science"], "published_date": "2026-04-24", "url": "https://arxiv.org/abs/2604.22328", "pdf_url": "https://arxiv.org/pdf/2604.22328v1", "arxiv_id": "2604.22328", "doi": "10.48550/arXiv.2604.22328", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.55}
{"id": "dfb96aa68551405b0f376ad5ffb264c5e647e2ed9a243d63341d3faf496da722", "sources": ["arxiv", "semantic_scholar"], "title": "Empirical Assessment of Time-Series Foundation Models For Power System Forecasting Applications", "abstract": "Accurate forecasting of electric load and renewable generation is essential for reliable and cost effective power system operations. Recent advances in transformer based and foundation machine learning models, driven by large scale pretraining, increased available data and computation, in addition to architectural innovations, have shown promise in time series forecasting across multiple domains. However, their application to power system forecasting tasks remains largely underexplored. This work presents a comprehensive, empirical benchmark of state of the art time series foundation models, transformer architectures, and deep learning baselines for solar, wind, and load forecasting using the high resolution ARPAE PERFORM dataset for the Electric Reliability Council of Texas (ERCOT) grid. Eight core capabilities are assessed, including zero shot performance, fine tuning efficiency, multivariate input and output handling, horizon sensitivity, generalization to unseen sites, probabilistic forecasting, and context window effects. Models evaluated include TimesFM, Chronos Bolt, MoiraiL, MOMENT, Tiny Time Mixer, Temporal Fusion Transformer, PatchTST, TimeXer, LSTM, and CNN. The manuscript aims to provide clear guidance on when foundation models can provide enhanced renewable and load forecasting capabilities and when other approaches remain the more practical choice for power system operations.", "authors": ["Muhy Eddin Za'ter", "Bri-Mathias Hodge"], "categories": ["eess.SY"], "fields_of_study": ["Computer Science", "Engineering"], "published_date": "2026-04-23", "url": "https://arxiv.org/abs/2604.22077", "pdf_url": "https://arxiv.org/pdf/2604.22077v1", "arxiv_id": "2604.22077", "doi": "10.48550/arXiv.2604.22077", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.55}
{"id": "ec54466fe7741e99b979b1413d921dfdb2cbe8ec73297d1c7cee8eeb2eac723c", "sources": ["arxiv", "semantic_scholar"], "title": "WaveMoE: A Wavelet-Enhanced Mixture-of-Experts Foundation Model for Time Series Forecasting", "abstract": "Time series foundation models (TSFMs) have recently achieved remarkable success in universal forecasting by leveraging large-scale pretraining on diverse time series data. Complementing this progress, incorporating frequency-domain information yields promising performance in enhancing the modeling of complex temporal patterns, such as periodicity and localized high-frequency dynamics, which are prevalent in real-world time series. To advance this direction, we propose a new perspective that integrates explicit frequency-domain representations into scalable foundation models, and introduce WaveMoE, a wavelet-enhanced mixture-of-experts foundation model for time series forecasting. WaveMoE adopts a dual-path architecture that jointly processes time series tokens and wavelet tokens aligned along a unified temporal axis, and coordinates them through a shared expert routing mechanism that enables consistent expert specialization while efficiently scaling model capacity. Preliminary experimental results on 16 diverse benchmark datasets indicate that WaveMoE has the potential to further improve forecasting performance by incorporating wavelet-domain corpora.", "authors": ["Shunyu Wu", "Jiawei Huang", "Weibin Feng", "Boxin Li", "Xiao Zhang", "Erli Meng", "Dan Li", "Jian Lou", "See-Kiong Ng"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2026-04-12", "url": "https://arxiv.org/abs/2604.10544", "pdf_url": "https://arxiv.org/pdf/2604.10544v1", "arxiv_id": "2604.10544", "doi": "10.48550/arXiv.2604.10544", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.5408}
{"id": "9508de7e24e50eba0c8bbe7e149cb1ec7e417cb770b4b29759a9c8e4e018bde7", "sources": ["arxiv", "semantic_scholar"], "title": "Zero-shot Multivariate Time Series Forecasting Using Tabular Prior Fitted Networks", "abstract": "Tabular foundation models, particularly Prior-data Fitted Networks like TabPFN have emerged as the leading contender in a myriad of tasks ranging from data imputation to label prediction on the tabular data format surpassing the historical successes of tree-based models. This has led to investigations on their applicability to forecasting time series data which can be formulated as a tabular problem. While recent work to this end has displayed positive results, most works have limited their treatment of multivariate time series problems to several independent univariate time series forecasting subproblems, thus ignoring any inter-channel interactions. Overcoming this limitation, we introduce a generally applicable framework for multivariate time series forecasting using tabular foundation models. We achieve this by recasting the multivariate time series forecasting problem as a series of scalar regression problems which can then be solved zero-shot by any tabular foundation model with regression capabilities. We present results of our method using the TabPFN-TS backbone and compare performance with the current state of the art tabular methods.", "authors": ["Mayuka Jayawardhana", "Nihal Sharma", "Kazem Meidani", "Bayan Bruss", "Tom Goldstein", "Doron Bergman"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2026-04-09", "url": "https://arxiv.org/abs/2604.08400", "pdf_url": "https://arxiv.org/pdf/2604.08400v1", "arxiv_id": "2604.08400", "doi": "10.48550/arXiv.2604.08400", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.5374}
{"id": "4c443abf1ee6741bcbfdfbeb68da9cb192d4997791082e96c70b218ed9f7c621", "sources": ["arxiv", "semantic_scholar"], "title": "Dynamic Linear Coregionalization for Realistic Synthetic Multivariate Time Series", "abstract": "Synthetic data is essential for training foundation models for time series (FMTS), but most generators assume static correlations, and are typically missing realistic inter-channel dependencies. We introduce DynLMC, a Dynamic Linear Model of Coregionalization, that incorporates time-varying, regime-switching correlations and cross-channel lag structures. Our approach produces synthetic multivariate time series with correlation dynamics that closely resemble real data. Fine-tuning three foundational models on DynLMC-generated data yields consistent zero-shot forecasting improvements across nine benchmarks. Our results demonstrate that modeling dynamic inter-channel correlations enhances FMTS transferability, highlighting the importance of data-centric pretraining.", "authors": ["Annita Vapsi", "Penghang Liu", "Saheed Obitayo", " Aakriti", "Manoj Cherukumalli", "Prathamesh Patil", "Amit Varshney", "Nicolas Marchesotti", "Elizabeth Fons", "Vamsi K. Potluru", "Manuela Veloso"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2026-04-06", "url": "https://arxiv.org/abs/2604.05064", "pdf_url": "https://arxiv.org/pdf/2604.05064v2", "arxiv_id": "2604.05064", "doi": "10.48550/arXiv.2604.05064", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.534}
{"id": "17393d7a2fc863455b515f0b452bc60a87f330c5315153952d06cbba33b5afd0", "sources": ["arxiv", "semantic_scholar"], "title": "Forecast collapse of transformer-based models under squared loss in financial time series", "abstract": "We study trajectory forecasting under squared loss for time series with weak conditional structure, using highly expressive prediction models. Building on the classical characterization of squared-loss risk minimization, we emphasize regimes in which the conditional expectation of future trajectories is effectively degenerate, leading to trivial Bayes-optimal predictors (flat for prices and zero for returns in standard financial settings). In this regime, increased model expressivity does not improve predictive accuracy but instead introduces spurious trajectory fluctuations around the optimal predictor. These fluctuations arise from the reuse of noise and result in increased prediction variance without any reduction in bias. This provides a process-level explanation for the degradation of Transformerbased forecasts on financial time series. We complement these theoretical results with numerical experiments on high-frequency EUR/USD exchange rate data, analyzing the distribution of trajectory-level forecasting errors. The results show that Transformer-based models yield larger errors than a simple linear benchmark on a large majority of forecasting windows, consistent with the variance-driven mechanism identified by the theory.", "authors": ["Pierre Andreoletti"], "categories": ["stat.ML", "cs.LG", "math.PR", "math.ST", "q-fin.CP"], "fields_of_study": ["Computer Science", "Mathematics", "Economics"], "published_date": "2026-03-31", "url": "https://arxiv.org/abs/2604.00064", "pdf_url": "https://arxiv.org/pdf/2604.00064v1", "arxiv_id": "2604.00064", "doi": "10.48550/arXiv.2604.00064", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.5271}
{"id": "1897a463760b10ae19dd5deb41a1b962ac5044b2947ece095bc226440f5682ab", "sources": ["arxiv", "semantic_scholar"], "title": "Multimodal Forecasting for Commodity Prices Using Spectrogram-Based and Time Series Representations", "abstract": "Forecasting multivariate time series remains challenging due to complex cross-variable dependencies and the presence of heterogeneous external influences. This paper presents Spectrogram-Enhanced Multimodal Fusion (SEMF), which combines spectral and temporal representations for more accurate and robust forecasting. The target time series is transformed into Morlet wavelet spectrograms, from which a Vision Transformer encoder extracts localized, frequency-aware features. In parallel, exogenous variables, such as financial indicators and macroeconomic signals, are encoded via a Transformer to capture temporal dependencies and multivariate dynamics. A bidirectional cross-attention module integrates these modalities into a unified representation that preserves distinct signal characteristics while modeling cross-modal correlations. Applied to multiple commodity price forecasting tasks, SEMF achieves consistent improvements over seven competitive baselines across multiple forecasting horizons and evaluation metrics. These results demonstrate the effectiveness of multimodal fusion and spectrogram-based encoding in capturing multi-scale patterns within complex financial time series.", "authors": ["Soyeon Park", "Doohee Chung", "Charmgil Hong"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2026-03-28", "url": "https://arxiv.org/abs/2603.27321", "pdf_url": "https://arxiv.org/pdf/2603.27321v1", "arxiv_id": "2603.27321", "doi": "10.48550/arXiv.2603.27321", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.5236}
{"id": "804c75bbe5ac8f335b64711994778ec81f85e78b76dc78c14ba8f93ef0f61061", "sources": ["arxiv", "semantic_scholar"], "title": "QuitoBench: A High-Quality Open Time Series Forecasting Benchmark", "abstract": "Time series forecasting is critical across finance, healthcare, and cloud computing, yet progress is constrained by a fundamental bottleneck: the scarcity of large-scale, high-quality benchmarks. To address this gap, we introduce \\textsc{QuitoBench}, a regime-balanced benchmark for time series forecasting with coverage across eight trend$\\times$seasonality$\\times$forecastability (TSF) regimes, designed to capture forecasting-relevant properties rather than application-defined domain labels. The benchmark is built upon \\textsc{Quito}, a billion-scale time series corpus of application traffic from Alipay spanning nine business domains. Benchmarking 10 models from deep learning, foundation models, and statistical baselines across 232,200 evaluation instances, we report four key findings: (i) a context-length crossover where deep learning models lead at short context ($L=96$) but foundation models dominate at long context ($L \\ge 576$); (ii) forecastability is the dominant difficulty driver, producing a $3.64 \\times$ MAE gap across regimes; (iii) deep learning models match or surpass foundation models at $59 \\times$ fewer parameters; and (iv) scaling the amount of training data provides substantially greater benefit than scaling model size for both model families. These findings are validated by strong cross-benchmark and cross-metric consistency. Our open-source release enables reproducible, regime-aware evaluation for time series forecasting research.", "authors": ["Siqiao Xue", "Zhaoyang Zhu", "Wei Zhang", "Rongyao Cai", "Rui Wang", "Yixiang Mu", "Fan Zhou", "Jianguo Li", "Peng Di", "Hang Yu"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2026-03-27", "url": "https://arxiv.org/abs/2603.26017", "pdf_url": "https://arxiv.org/pdf/2603.26017v1", "arxiv_id": "2603.26017", "doi": "10.48550/arXiv.2603.26017", "citation_count": 4, "influential_citation_count": 0, "has_code": true, "code_url": null, "venue": "arXiv.org", "quality_score": 0.8075}
{"id": "9edf58766f9ca387bcccb8347cbea5e82ef6baacd508a6008214584018edbec1", "sources": ["arxiv", "semantic_scholar"], "title": "IPatch: A Multi-Resolution Transformer Architecture for Robust Time-Series Forecasting", "abstract": "Accurate forecasting of multivariate time series remains challenging due to the need to capture both short-term fluctuations and long-range temporal dependencies. Transformer-based models have emerged as a powerful approach, but their performance depends critically on the representation of temporal data. Traditional point-wise representations preserve individual time-step information, enabling fine-grained modeling, yet they tend to be computationally expensive and less effective at modeling broader contextual dependencies, limiting their scalability to long sequences. Patch-wise representations aggregate consecutive steps into compact tokens to improve efficiency and model local temporal dynamics, but they often discard fine-grained temporal details that are critical for accurate predictions in volatile or complex time series. We propose IPatch, a multi-resolution Transformer architecture that integrates both point-wise and patch-wise tokens, modeling temporal information at multiple resolutions. Experiments on 7 benchmark datasets demonstrate that IPatch consistently improves forecasting accuracy, robustness to noise, and generalization across various prediction horizons compared to single-representation baselines.", "authors": ["Aymane Harkati", "Moncef Garouani", "Olivier Teste", "Julien Aligon", "Mohamed Hamlich"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2026-03-25", "url": "https://arxiv.org/abs/2603.24207", "pdf_url": "https://arxiv.org/pdf/2603.24207v1", "arxiv_id": "2603.24207", "doi": "10.48550/arXiv.2603.24207", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.5202}
{"id": "df8309b1f27e72fb87c2fb10f663336f22191772d07d0ed64e12ba0d6dee3ba8", "sources": ["arxiv", "semantic_scholar"], "title": "Forecasting with Guidance: Representation-Level Supervision for Time Series Forecasting", "abstract": "Nowadays, time series forecasting is predominantly approached through the end-to-end training of deep learning architectures using error-based objectives. While this is effective at minimizing average loss, it encourages the encoder to discard informative yet extreme patterns. This results in smooth predictions and temporal representations that poorly capture salient dynamics. To address this issue, we propose ReGuider, a plug-in method that can be seamlessly integrated into any forecasting architecture. ReGuider leverages pretrained time series foundation models as semantic teachers. During training, the input sequence is processed together by the target forecasting model and the pretrained model. Rather than using the pretrained model's outputs directly, we extract its intermediate embeddings, which are rich in temporal and semantic information, and align them with the target model's encoder embeddings through representation-level supervision. This alignment process enables the encoder to learn more expressive temporal representations, thereby improving the accuracy of downstream forecasting. Extensive experimentation across diverse datasets and architectures demonstrates that our ReGuider consistently improves forecasting performance, confirming its effectiveness and versatility.", "authors": ["Jiacheng Wang", "Liang Fan", "Baihua Li", "Luyan Zhang"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2026-03-25", "url": "https://arxiv.org/abs/2603.24262", "pdf_url": "https://arxiv.org/pdf/2603.24262v1", "arxiv_id": "2603.24262", "doi": "10.48550/arXiv.2603.24262", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.5202}
{"id": "7670097b8e92b6505edfd80fb410cde87c0f5aa48a1efb2b2b0b274006ec5be4", "sources": ["arxiv", "semantic_scholar"], "title": "CoRA: Boosting Time Series Foundation Models for Multivariate Forecasting through Correlation-aware Adapter", "abstract": "Most existing Time Series Foundation Models (TSFMs) use channel independent modeling and focus on capturing and generalizing temporal dependencies, while neglecting the correlations among channels or overlooking the different aspects of correlations. However, these correlations play a vital role in Multivariate time series forecasting. To address this, we propose a CoRrelation-aware Adapter (CoRA), a lightweight plug-and-play method that requires only fine-tuning with TSFMs and is able to capture different types of correlations, so as to improve forecast performance. Specifically, to reduce complexity, we innovatively decompose the correlation matrix into low-rank Time-Varying and Time-Invariant components. For the Time-Varying component, we further design learnable polynomials to learn dynamic correlations by capturing trends or periodic patterns. To learn positive and negative correlations that appear only among some channels, we introduce a novel dual contrastive learning method that identifies correlations through projection layers, regulated by a Heterogeneous-Partial contrastive loss during training, without introducing additional complexity in the inference stage. Extensive experiments on 10 real-world datasets demonstrate that CoRA can improve TSFMs in multivariate forecasting performance.", "authors": ["Hanyin Cheng", "Xingjian Wu", "Yang Shu", "Zhongwen Rao", "Lujia Pan", "Bin Yang", "Chenjuan Guo"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2026-03-23", "url": "https://arxiv.org/abs/2603.21828", "pdf_url": "https://arxiv.org/pdf/2603.21828v1", "arxiv_id": "2603.21828", "doi": "10.48550/arXiv.2603.21828", "citation_count": 2, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.5179}
{"id": "15b22151dba2c0c6384838fcc8e08c2a456abbbf287962dfafbab4e0fb386d65", "sources": ["arxiv", "semantic_scholar"], "title": "A Foundation Model for Instruction-Conditioned In-Context Time Series Tasks", "abstract": "In-context learning (ICL) enables task adaptation at inference time by conditioning on demonstrations rather than updating model parameters. Although recent time-series foundation models incorporate contextual conditioning, retrieval, or example-based prompting, they typically rely on implicit positional structure or task-specific objectives rather than explicit instruction-conditioned input-output demonstrations. We introduce iAmTime, a time-series foundation model trained with instruction-conditioned amortized meta-learning to infer tasks directly from example demonstrations. iAmTime represents each episode as a structured prompt over historical context and future-known variables using specialized semantic tokens that attend to designated time-series regions, exchange information across demonstrations, and inject task information into the query representation. The model combines a Hierarchical Multi-Scope Transformer Encoder, which captures temporal and covariate dynamics while inferring latent task structure from demonstrated input-output mappings, with a Task-Conditioned Patch Decoder, which adapts decoding through expert-based routing. We train iAmTime on large-scale real and synthetic corpora using supervised and self-supervised instruction-conditioned tasks, including forecasting, imputation, reconstruction, classification, anomaly detection, and source de-mixing. Across diverse domains, frequencies, and horizons, iAmTime improves zero-shot adaptation over strong time-series foundation baselines on probabilistic and point forecasting benchmarks, while achieving competitive performance on non-forecasting tasks such as classification.", "authors": ["Anish Saha", "Konstantin Shmakov"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2026-03-23", "url": "https://arxiv.org/abs/2603.22586", "pdf_url": "https://arxiv.org/pdf/2603.22586v3", "arxiv_id": "2603.22586", "doi": "10.48550/arXiv.2603.22586", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.5179}
{"id": "f4a397fc6e2b9a9ae84cb89bd66c0ec5435b02a721ce2c065d30feea8940b4a4", "sources": ["arxiv", "semantic_scholar"], "title": "FISformer: Replacing Self-Attention with a Fuzzy Inference System in Transformer Models for Time Series Forecasting", "abstract": "Transformers have achieved remarkable progress in time series forecasting, yet their reliance on deterministic dot-product attention limits their capacity to model uncertainty and nonlinear dependencies across multivariate temporal dimensions. To address this limitation, we propose FISFormer, a Fuzzy Inference System-driven Transformer that replaces conventional attention with a FIS Interaction mechanism. In this framework, each query-key pair undergoes a fuzzy inference process for every feature dimension, where learnable membership functions and rule-based reasoning estimate token-wise relational strengths. These FIS-derived interaction weights capture uncertainty and provide interpretable, continuous mappings between tokens. A softmax operation is applied along the token axis to normalize these weights, which are then combined with the corresponding value features through element-wise multiplication to yield the final context-enhanced token representations. This design fuses the interpretability and uncertainty modeling of fuzzy logic with the representational power of Transformers. Extensive experiments on multiple benchmark datasets demonstrate that FISFormer achieves superior forecasting accuracy, noise robustness, and interpretability compared to state-of-the-art Transformer variants, establishing fuzzy inference as an effective alternative to conventional attention mechanisms.", "authors": ["Bulent Haznedar", "Levent Karacan"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2026-03-23", "url": "https://arxiv.org/abs/2603.21724", "pdf_url": "https://arxiv.org/pdf/2603.21724v1", "arxiv_id": "2603.21724", "doi": "10.1109/tfuzz.2026.3690012", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "IEEE transactions on fuzzy systems", "quality_score": 0.5179}
{"id": "ff57824304777f9714ec230d3f8b234041a2234d5466e58ce161aab76fbee0bf", "sources": ["arxiv", "semantic_scholar"], "title": "Integrating Inductive Biases in Transformers via Distillation for Financial Time Series Forecasting", "abstract": "Transformer-based models have been widely adopted for time-series forecasting due to their high representational capacity and architectural flexibility. However, many Transformer variants implicitly assume stationarity and stable temporal dynamics -- assumptions routinely violated in financial markets characterized by regime shifts and non-stationarity. Empirically, state-of-the-art time-series Transformers often underperform even vanilla Transformers on financial tasks, while simpler architectures with distinct inductive biases, such as CNNs and RNNs, can achieve stronger performance with substantially lower complexity. At the same time, no single inductive bias dominates across markets or regimes, suggesting that robust financial forecasting requires integrating complementary temporal priors. We propose TIPS (Transformer with Inductive Prior Synthesis), a knowledge distillation framework that synthesizes diverse inductive biases -- causality, locality, and periodicity -- within a unified Transformer. TIPS trains bias-specialized Transformer teachers via attention masking, then distills their knowledge into a single student model with regime-dependent alignment across inductive biases. Across four major equity markets, TIPS achieves state-of-the-art performance, outperforming strong ensemble baselines by 55%, 9%, and 16% in annual return, Sharpe ratio, and Calmar ratio, while requiring only 38% of the inference-time computation. Further analyses show that TIPS generates statistically significant excess returns beyond both vanilla Transformers and its teacher ensembles, and exhibits regime-dependent behavioral alignment with classical architectures during their profitable periods. These results highlight the importance of regime-dependent inductive bias utilization for robust generalization in non-stationary financial time series.", "authors": ["Yu-Chen Den", "Kuan-Yu Chen", "Kendro Vincent", "Darby Tien-Hao Chang"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2026-03-17", "url": "https://arxiv.org/abs/2603.16985", "pdf_url": "https://arxiv.org/pdf/2603.16985v2", "arxiv_id": "2603.16985", "doi": "10.48550/arXiv.2603.16985", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.511}
{"id": "1b056a184b59695aa9b66441e253390dcb0b00522c7178533f96303fa69f9574", "sources": ["arxiv", "semantic_scholar"], "title": "Not All Retrievals are Useful: Cross-Attention for Input-Aware RAG in Time Series Forecasting", "abstract": "Retrieval-augmented generation (RAG) enhances zero-shot time series (TS) forecasting by leveraging external knowledge bases, yet existing approaches overlook input-level relevance when fusing retrieved samples with the query. We argue that not all retrievals are equally useful, and irrelevant ones can degrade performance. To this end, we propose Cross-RAG, a zero-shot RAG-based forecasting framework that selectively attends to query-relevant retrieved samples via query--retrieval cross-attention. By modeling input-level relevance between the query and retrieved samples, Cross-RAG jointly incorporates three sources of information: 1) the query itself, 2) the retrieved samples, and 3) their relational interactions. In particular, this input-aware design enables Cross-RAG to remain stable as the number of retrieved samples $k$ grows, whereas prior methods without cross-attention require careful $k$ tuning to avoid degradation from irrelevant retrievals. Extensive experiments demonstrate that Cross-RAG consistently improves zero-shot forecasting performance across multiple TSFM backbones and various RAG methods, with additional analyses confirming its effectiveness across various retrieval scenarios. Code is available at https://github.com/seunghan96/cross-rag/.", "authors": ["Seunghan Lee", "Jaehoon Lee", "Jun Seo", "Sungdong Yoo", "Minjae Kim", "Tae Yoon Lim", "Dongwan Kang", "Hwanil Choi", "SoonYoung Lee", "Wonbin Ahn"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2026-03-16", "url": "https://arxiv.org/abs/2603.14709", "pdf_url": "https://arxiv.org/pdf/2603.14709v2", "arxiv_id": "2603.14709", "doi": null, "citation_count": 1, "influential_citation_count": 1, "has_code": true, "code_url": "https://github.com/seunghan96/cross-rag/", "venue": null, "quality_score": 0.6026}
{"id": "bc258b366c67b84e16aae670e16334fa8aa92cfbb94d5caac33fe94bd82f4d69", "sources": ["arxiv", "semantic_scholar"], "title": "Interventional Time Series Priors for Causal Foundation Models", "abstract": "Prior-data fitted networks (PFNs) have emerged as powerful foundation models for tabular causal inference, yet their extension to time series remains limited by the absence of synthetic data generators that provide interventional targets. Existing time series benchmarks generate observational data with ground-truth causal graphs but lack the interventional data required for training causal foundation models. To address this, we propose \\textbf{CausalTimePrior}, a principled framework for generating synthetic temporal structural causal models (TSCMs) with paired observational and interventional time series. Our prior supports configurable causal graph structures, nonlinear autoregressive mechanisms, regime-switching dynamics, and multiple intervention types (hard, soft, time-varying). We demonstrate that PFNs trained on CausalTimePrior can perform in-context causal effect estimation on held-out TSCMs, establishing a pathway toward foundation models for time series causal inference.", "authors": ["Dennis Thumm", "Ying Chen"], "categories": ["cs.LG", "stat.ME"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2026-03-11", "url": "https://arxiv.org/abs/2603.11090", "pdf_url": "https://arxiv.org/pdf/2603.11090v2", "arxiv_id": "2603.11090", "doi": "10.48550/arXiv.2603.11090", "citation_count": 2, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.5042}
{"id": "2550e7f22cdcf4fe727f72bf13decd0373bbb1910ad4c840fdb88bd9534e1138", "sources": ["arxiv", "semantic_scholar"], "title": "Dissecting Chronos: Sparse Autoencoders Reveal Causal Feature Hierarchies in Time Series Foundation Models", "abstract": "Time series foundation models (TSFMs) are increasingly deployed in high-stakes domains, yet their internal representations remain opaque. We present the first application of sparse autoencoders (SAEs) to a TSFM, training TopK SAEs on activations of Chronos-T5-Large (710M parameters) across six layers. Through 392 single-feature ablation experiments, we establish that every ablated feature produces a positive CRPS degradation, confirming causal relevance. Our analysis reveals a depth-dependent hierarchy: early encoder layers encode low-level frequency features, the mid-encoder concentrates causally critical change-detection features, and the final encoder compresses a rich but less causally important taxonomy of temporal concepts. The most critical features reside in the mid-encoder (max single-feature Delta CRPS = 38.61), not in the semantically richest final encoder layer, where progressive ablation paradoxically improves forecast quality. These findings demonstrate that mechanistic interpretability transfers effectively to TSFMs and that Chronos-T5 relies on abrupt-dynamics detection rather than periodic pattern recognition.", "authors": ["Anurag Mishra"], "categories": ["cs.LG", "cs.AI", "cs.CL"], "fields_of_study": ["Computer Science"], "published_date": "2026-03-10", "url": "https://arxiv.org/abs/2603.10071", "pdf_url": "https://arxiv.org/pdf/2603.10071v1", "arxiv_id": "2603.10071", "doi": "10.48550/arXiv.2603.10071", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.503}
{"id": "e693f36328b5b2db3ce9e224bcef260977c1a4375a1ee36756be1cd112d8ec2a", "sources": ["arxiv", "semantic_scholar"], "title": "UTICA: Multi-Objective Self-Distllation Foundation Model Pretraining for Time Series Classification", "abstract": "Self-supervised foundation models have achieved remarkable success across domains, including time series. However, the potential of non-contrastive methods, a paradigm that has driven significant advances in computer vision, remains underexplored for time series. In this work, we adapt DINOv2-style self-distillation to pretrain a time series foundation model, building on the Mantis tokenizer and transformer encoder architecture as our backbone. Through a student-teacher framework, our method Utica learns representations that capture both temporal invariance via augmented crops and fine-grained local structure via patch masking. Our approach achieves state-of-the-art classification performance on both UCR and UEA benchmarks. These results suggest that non-contrastive methods are a promising and complementary pretraining strategy for time series foundation models.", "authors": ["Yessin Moakher", "Youssef Attia El Hili", "Vasilii Feofanov"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2026-03-02", "url": "https://arxiv.org/abs/2603.01348", "pdf_url": "https://arxiv.org/pdf/2603.01348v1", "arxiv_id": "2603.01348", "doi": "10.48550/arXiv.2603.01348", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.4939}
{"id": "0c737fdb767a4f7d6bd12d5d3a319506b95227b0bf8929a988570afb1bcacf00", "sources": ["arxiv", "semantic_scholar"], "title": "Retrodictive Forecasting: A Proof-of-Concept for Exploiting Temporal Asymmetry in Time Series Prediction", "abstract": "We propose a retrodictive forecasting paradigm for time series: instead of predicting the future from the past, we identify the future that best explains the observed present via inverse MAP optimization over a Conditional Variational Autoencoder (CVAE). This conditioning is a statistical modeling choice for Bayesian inversion; it does not assert that future events cause past observations. The approach is theoretically grounded in an information-theoretic arrow-of-time measure: the symmetrized Kullback-Leibler divergence between forward and time-reversed trajectory ensembles provides both the conceptual rationale and an operational GO/NO-GO diagnostic for applicability. We implement the paradigm as MAP inference over an inverse CVAE with a learned RealNVP normalizing-flow prior and evaluate it on six time series cases: four synthetic processes with controlled temporal asymmetry and two ERA5 reanalysis datasets (wind speed and solar irradiance). The work makes four contributions: (i) a formal retrodictive inference formulation; (ii) an inverse CVAE architecture; (iii) a model-free irreversibility diagnostic; and (iv) a falsifiable validation protocol with four pre-specified predictions. All pre-specified predictions are empirically supported: the diagnostic correctly classifies all six cases; the learned flow prior improves over an isotropic Gaussian baseline on GO cases; the inverse MAP yields no spurious advantage on time-reversible dynamics; and on irreversible GO cases, it achieves competitive or superior RMSE relative to forward baselines, with a statistically significant 17.7% reduction over a forward MLP on ERA5 solar irradiance. These results provide a structured proof-of-concept that retrodictive forecasting can constitute a viable alternative to conventional forward prediction when statistical time-irreversibility is present and exploitable.", "authors": ["Cedric Damour"], "categories": ["cs.LG", "physics.ao-ph", "stat.ML"], "fields_of_study": ["Computer Science", "Physics", "Mathematics"], "published_date": "2026-02-28", "url": "https://arxiv.org/abs/2603.00636", "pdf_url": "https://arxiv.org/pdf/2603.00636v1", "arxiv_id": "2603.00636", "doi": "10.48550/arXiv.2603.00636", "citation_count": 0, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/cdamour/retrodictive-forecasting", "venue": "arXiv.org", "quality_score": 0.7597}
{"id": "5e6b55a3de03c005894ad5e864197de0f2a85193d45c2607e9519a2f4ad32517", "sources": ["arxiv", "semantic_scholar"], "title": "Time Series Foundation Models as Strong Baselines in Transportation Forecasting: A Large-Scale Benchmark Analysis", "abstract": "Accurate forecasting of transportation dynamics is essential for urban mobility and infrastructure planning. Although recent work has achieved strong performance with deep learning models, these methods typically require dataset-specific training, architecture design and hyper-parameter tuning. This paper evaluates whether general-purpose time-series foundation models can serve as forecasters for transportation tasks by benchmarking the zero-shot performance of the state-of-the-art model, Chronos-2, across ten real-world datasets covering highway traffic volume and flow, urban traffic speed, bike-sharing demand, and electric vehicle charging station data. Under a consistent evaluation protocol, we find that, even without any task-specific fine-tuning, Chronos-2 delivers state-of-the-art or competitive accuracy across most datasets, frequently outperforming classical statistical baselines and specialized deep learning architectures, particularly at longer horizons. Beyond point forecasting, we evaluate its native probabilistic outputs using prediction-interval coverage and sharpness, demonstrating that Chronos-2 also provides useful uncertainty quantification without dataset-specific training. In general, this study supports the adoption of time-series foundation models as a key baseline for transportation forecasting research.", "authors": ["Javier Yanes-Pulido", "Filipe Rodrigues"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2026-02-27", "url": "https://arxiv.org/abs/2602.24238", "pdf_url": "https://arxiv.org/pdf/2602.24238v2", "arxiv_id": "2602.24238", "doi": "10.48550/arXiv.2602.24238", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.4904}
{"id": "51f7d85bf74586edc7e24b1e01f7412331bece504ebf0f48047ee242ec473c4d", "sources": ["arxiv", "semantic_scholar"], "title": "DualWeaver: Synergistic Feature Weaving Surrogates for Multivariate Forecasting with Univariate Time Series Foundation Models", "abstract": "Time-series foundation models (TSFMs) have achieved strong univariate forecasting through large-scale pre-training, yet effectively extending this success to multivariate forecasting remains challenging. To address this, we propose DualWeaver, a novel framework that adapts univariate TSFMs (Uni-TSFMs) for multivariate forecasting by using a pair of learnable, structurally symmetric surrogate series. Generated by a shared auxiliary feature-fusion module that captures cross-variable dependencies, these surrogates are mapped to TSFM-compatible series via the forecasting objective. The symmetric structure enables parameter-free reconstruction of final predictions directly from the surrogates, without additional parametric decoding. A theoretically grounded regularization term is further introduced to enhance robustness against adaptation collapse. Extensive experiments on diverse real-world datasets show that DualWeaver outperforms state-of-the-art multivariate forecasters in both accuracy and stability. We release the code at https://github.com/li-jinpeng/DualWeaver.", "authors": ["Jinpeng Li", "Zhongyi Pei", "Huaze Xue", "Bojian Zheng", "Chen Wang", "Jianmin Wang"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2026-02-25", "url": "https://arxiv.org/abs/2602.22066", "pdf_url": "https://arxiv.org/pdf/2602.22066v1", "arxiv_id": "2602.22066", "doi": "10.48550/arXiv.2602.22066", "citation_count": 0, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/li-jinpeng/DualWeaver", "venue": "arXiv.org", "quality_score": 0.7544}
{"id": "66985c4508b91a1723f5b2359d6e6222afe87deda0a0175fcc39f1730020f928", "sources": ["arxiv", "semantic_scholar"], "title": "Reverso: Efficient Time Series Foundation Models for Zero-shot Forecasting", "abstract": "Learning time series foundation models has been shown to be a promising approach for zero-shot time series forecasting across diverse time series domains. Insofar as scaling has been a critical driver of performance of foundation models in other modalities such as language and vision, much recent work on time series foundation modeling has focused on scaling. This has resulted in time series foundation models with hundreds of millions of parameters that are, while performant, inefficient and expensive to use in practice. This paper describes a simple recipe for learning efficient foundation models for zero-shot time series forecasting that are orders of magnitude smaller. We show that large-scale transformers are not necessary: small hybrid models that interleave long convolution and linear RNN layers (in particular DeltaNet layers) can match the performance of larger transformer-based models while being more than a hundred times smaller. We also describe several data augmentation and inference strategies that further improve performance. This recipe results in Reverso, a family of efficient time series foundation models for zero-shot forecasting that significantly push the performance-efficiency Pareto frontier.", "authors": ["Xinghong Fu", "Yanhong Li", "Georgios Papaioannou", "Yoon Kim"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2026-02-19", "url": "https://arxiv.org/abs/2602.17634", "pdf_url": "https://arxiv.org/pdf/2602.17634v1", "arxiv_id": "2602.17634", "doi": "10.48550/arXiv.2602.17634", "citation_count": 3, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.4813}
{"id": "80275cfa12ac0fd461284c2ab694729311d10e60598387b2751c5eee1076d5cd", "sources": ["arxiv", "semantic_scholar"], "title": "Structure-Aware Set Transformers: Temporal and Variable-Type Attention Biases for Asynchronous Clinical Time Series", "abstract": "Electronic health records (EHR) are irregular, asynchronous multivariate time series. As time-series foundation models increasingly tokenize events rather than discretizing time, the input layout becomes a key design choice. Grids expose time$\\times$variable structure but require imputation or missingness masks, risking error or sampling-policy shortcuts. Point-set tokenization avoids discretization but loses within-variable trajectories and time-local cross-variable context (Fig.1). We restore these priors in STructure-AwaRe (STAR) Set Transformer by adding parameter-efficient soft attention biases: a temporal locality penalty $-|Δt|/τ$ with learnable timescales and a variable-type affinity $B_{s_i,s_j}$ from a learned feature-compatibility matrix. We benchmark 10 depth-wise fusion schedules (Fig.2). On three ICU prediction tasks, STAR-Set achieves AUC/APR of 0.7158/0.0026 (CPR), 0.9164/0.2033 (mortality), and 0.8373/0.1258 (vasopressor use), outperforming regular-grid, event-time grid, and prior set baselines. Learned $τ$ and $B$ provide interpretable summaries of temporal context and variable interactions, offering a practical plug-in for context-informed time-series models.", "authors": ["Joohyung Lee", "Kwanhyung Lee", "Changhun Kim", "Eunho Yang"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2026-02-18", "url": "https://arxiv.org/abs/2603.06605", "pdf_url": "https://arxiv.org/pdf/2603.06605v2", "arxiv_id": "2603.06605", "doi": "10.48550/arXiv.2603.06605", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.4801}
{"id": "849683763fdb23eed49854d8333921e61adcbe6b41c423a4a47108228172da98", "sources": ["arxiv", "semantic_scholar"], "title": "EIDOS: Latent-Space Predictive Learning for Time Series Foundation Models", "abstract": "Most time series foundation models are pretrained by directly predicting future observations, which often yields weakly structured latent representations that capture surface noise rather than coherent and predictable temporal dynamics. In this work, we introduce EIDOS, a foundation model family that shifts pretraining from future value prediction to latent-space predictive learning. We train a causal Transformer to predict the evolution of latent representations, encouraging the emergence of structured and temporally coherent latent states. To ensure stable targets for latent-space learning, we design a lightweight aggregation branch to construct target representations. EIDOS is optimized via a joint objective that integrates latent-space alignment, observational grounding to anchor representations to the input signal, and direct forecasting supervision. On the GIFT-Eval benchmark, EIDOS mitigates structural fragmentation in the representation space and achieves state-of-the-art performance. These results demonstrate that constraining models to learn predictable latent dynamics is a principled step toward more robust and reliable time series foundation models.", "authors": ["Xinxing Zhou", "Qingren Yao", "Yiji Zhao", "Chenghao Liu", "Flora Salim", "Xiaojie Yuan", "Yanlong Wen", "Ming Jin"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2026-02-15", "url": "https://arxiv.org/abs/2602.14024", "pdf_url": "https://arxiv.org/pdf/2602.14024v1", "arxiv_id": "2602.14024", "doi": "10.48550/arXiv.2602.14024", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.4767}
{"id": "c0afc7849df6d894effc0b8f4c76ddc1c729ab13d24779ba9775458303262d23", "sources": ["arxiv", "semantic_scholar"], "title": "Forecasting Commencing Enrolments Under Data Sparsity: A Zero-Shot Time Series Foundation Models Framework for Higher Education Planning", "abstract": "Effective resource allocation in higher education depends on reliable enrolment forecasts, yet institutional planners frequently face data series disrupted by structural shifts. This paper investigates whether zero-shot Time Series Foundation Models (TSFMs) can provide rigorous decision support for annual enrolment forecasting under severe data sparsity. We benchmark multiple TSFMs against classical operational baselines using an expanding-window backtest that mirrors decision-time constraints. To capture environmental shifts without look-ahead bias, we introduce a leakage-safe covariate protocol that integrates feature-engineered Google Trends with the Institutional Operating Conditions Index (IOCI), a transferable regime measure extracted from historical narrative evidence. Our evaluation demonstrates that covariate-conditioned TSFMs are competitive with classical methods and can improve accuracy without requiring bespoke institutional training. However, the operational benefits depend on cohort characteristics and covariate design. This study provides an auditable and transferable forecasting protocol for operational researchers and university administrators, helping institutions determine when context-aware forecasting adds practical value under limited data and structural instability.", "authors": ["Jittarin Jetwiriyanon", "Teo Susnjak", "Surangika Ranathunga"], "categories": ["cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2026-02-12", "url": "https://arxiv.org/abs/2602.12120", "pdf_url": "https://arxiv.org/pdf/2602.12120v3", "arxiv_id": "2602.12120", "doi": null, "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.3011}
{"id": "daed730a2d8e299725fbe5052cb0b7608e1f468cdc721b9cb60c6e9519d57ec9", "sources": ["arxiv", "semantic_scholar"], "title": "DiTS: Multimodal Diffusion Transformers Are Time Series Forecasters", "abstract": "While generative modeling on time series facilitates more capable and flexible probabilistic forecasting, existing generative time series models do not address the multi-dimensional properties of time series data well. The prevalent architecture of Diffusion Transformers (DiT), which relies on simplistic conditioning controls and a single-stream Transformer backbone, tends to underutilize cross-variate dependencies in covariate-aware forecasting. Inspired by Multimodal Diffusion Transformers that integrate textual guidance into video generation, we propose Diffusion Transformers for Time Series (DiTS), a general-purpose architecture that frames endogenous and exogenous variates as distinct modalities. To better capture both inter-variate and intra-variate dependencies, we design a dual-stream Transformer block tailored for time-series data, comprising a Time Attention module for autoregressive modeling along the temporal dimension and a Variate Attention module for cross-variate modeling. Unlike the common approach for images, which flattens 2D token grids into 1D sequences, our design leverages the low-rank property inherent in multivariate dependencies, thereby reducing computational costs. Experiments show that DiTS achieves state-of-the-art performance across benchmarks, regardless of the presence of future exogenous variate observations, demonstrating unique generative forecasting strengths over traditional deterministic deep forecasting models.", "authors": ["Haoran Zhang", "Haixuan Liu", "Yong Liu", "Yunzhong Qiu", "Yuxuan Wang", "Jianmin Wang", "Mingsheng Long"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2026-02-06", "url": "https://arxiv.org/abs/2602.06597", "pdf_url": "https://arxiv.org/pdf/2602.06597v1", "arxiv_id": "2602.06597", "doi": "10.48550/arXiv.2602.06597", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.4664}
{"id": "b9610e1d14ba1a6dc3f64b169e68b1bd492c03d26bbb4470aecb41e0f5e3c44c", "sources": ["arxiv", "semantic_scholar"], "title": "Revisiting the Generic Transformer: Deconstructing a Strong Baseline for Time Series Foundation Models", "abstract": "The recent surge in Time Series Foundation Models has rapidly advanced the field, yet the heterogeneous training setups across studies make it difficult to attribute improvements to architectural innovations versus data engineering. In this work, we investigate the potential of a standard patch Transformer, demonstrating that this generic architecture achieves state-of-the-art zero-shot forecasting performance using a straightforward training protocol. We conduct a comprehensive ablation study that covers model scaling, data composition, and training techniques to isolate the essential ingredients for high performance. Our findings identify the key drivers of performance, while confirming that the generic architecture itself demonstrates excellent scalability. By strictly controlling these variables, we provide comprehensive empirical results on model scaling across multiple dimensions. We release our open-source model and detailed findings to establish a transparent, reproducible baseline for future research.", "authors": ["Yunshi Wen", "Wesley M. Gifford", "Chandra Reddy", "Lam M. Nguyen", "Jayant Kalagnanam", "Anak Agung Julius"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2026-02-06", "url": "https://arxiv.org/abs/2602.06909", "pdf_url": "https://arxiv.org/pdf/2602.06909v1", "arxiv_id": "2602.06909", "doi": "10.48550/arXiv.2602.06909", "citation_count": 2, "influential_citation_count": 1, "has_code": true, "code_url": null, "venue": "arXiv.org", "quality_score": 0.7207}
{"id": "858ee827475c15707ff29aa79b5cf892eaecb84ce81bc1e396fc05176a51d215", "sources": ["arxiv", "semantic_scholar"], "title": "Assessing Electricity Demand Forecasting with Exogenous Data in Time Series Foundation Models", "abstract": "Time-series foundation models have emerged as a new paradigm for forecasting, yet their ability to effectively leverage exogenous features -- critical for electricity demand forecasting -- remains unclear. This paper empirically evaluates foundation models capable of modeling cross-channel correlations against a baseline LSTM with reversible instance normalization across Singaporean and Australian electricity markets at hourly and daily granularities. We systematically assess MOIRAI, MOMENT, TinyTimeMixers, ChronosX, and Chronos-2 under three feature configurations: all features, selected features, and target-only. Our findings reveal highly variable effectiveness: while Chronos-2 achieves the best performance among foundation models (in zero-shot settings), the simple baseline frequently outperforms all foundation models in Singapore's stable climate, particularly for short-term horizons. Model architecture proves critical, with synergistic architectural implementations (TTM's channel-mixing, Chronos-2's grouped attention) consistently leveraging exogenous features, while other approaches show inconsistent benefits. Geographic context emerges as equally important, with foundation models demonstrating advantages primarily in variable climates. These results challenge assumptions about universal foundation model superiority and highlight the need for domain-specific models, specifically in the energy domain.", "authors": ["Wei Soon Cheong", "Lian Lian Jiang", "Jamie Ng Suat Ling"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2026-02-05", "url": "https://arxiv.org/abs/2602.05390", "pdf_url": "https://arxiv.org/pdf/2602.05390v1", "arxiv_id": "2602.05390", "doi": "10.48550/arXiv.2602.05390", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.4652}
{"id": "dd9b72e790ea93834e75a536993cbf70560cdd5923a8a00c3ab34791e1020644", "sources": ["arxiv", "semantic_scholar"], "title": "Position: Beyond Model-Centric Prediction -- Agentic Time Series Forecasting", "abstract": "Time series forecasting has traditionally been formulated as a model-centric, static, and single-pass prediction problem that maps historical observations to future values. While this paradigm has driven substantial progress, it proves insufficient in adaptive and multi-turn settings where forecasting requires informative feature extraction, reasoning-driven inference, iterative refinement, and continual adaptation over time. In this paper, we argue for agentic time series forecasting (ATSF), which reframes forecasting as an agentic process composed of perception, planning, action, reflection, and memory. Rather than focusing solely on predictive models, ATSF emphasizes organizing forecasting as an agentic workflow that can interact with tools, incorporate feedback from outcomes, and evolve through experience accumulation. We outline three representative implementation paradigms -- workflow-based design, agentic reinforcement learning, and a hybrid agentic workflow paradigm -- and discuss the opportunities and challenges that arise when shifting from model-centric prediction to agentic forecasting. Together, this position aims to establish agentic forecasting as a foundation for future research at the intersection of time series forecasting.", "authors": ["Mingyue Cheng", "Xiaoyu Tao", "Qi Liu", "Ze Guo", "Enhong Chen"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2026-02-02", "url": "https://arxiv.org/abs/2602.01776", "pdf_url": "https://arxiv.org/pdf/2602.01776v4", "arxiv_id": "2602.01776", "doi": "10.48550/arXiv.2602.01776", "citation_count": 7, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.4618}
{"id": "dc664c75d9bc2170a8a48db9d32cc44c65f700e137c51a47a094ab5d996933be", "sources": ["arxiv", "semantic_scholar"], "title": "SEDformer: Event-Synchronous Spiking Transformers for Irregular Telemetry Time Series Forecasting", "abstract": "Telemetry streams from large-scale Internet-connected systems (e.g., IoT deployments and online platforms) naturally form an irregular multivariate time series (IMTS) whose accurate forecasting is operationally vital. A closer examination reveals a defining Sparsity-Event Duality (SED) property of IMTS, i.e., long stretches with sparse or no observations are punctuated by short, dense bursts where most semantic events (observations) occur. However, existing Graph- and Transformer-based forecasters ignore SED: pre-alignment to uniform grids with heavy padding violates sparsity by inflating sequences and forcing computation at non-informative steps, while relational recasting weakens event semantics by disrupting local temporal continuity. These limitations motivate a more faithful and natural modeling paradigm for IMTS that aligns with its SED property. We find that Spiking Neural Networks meet this requirement, as they communicate via sparse binary spikes and update in an event-driven manner, aligning naturally with the SED nature of IMTS. Therefore, we present SEDformer, an SED-enhanced Spiking Transformer for telemetry IMTS forecasting that couples: (1) a SED-based Spike Encoder converts raw observations into event synchronous spikes using an Event-Aligned LIF neuron, (2) an Event-Preserving Temporal Downsampling module compresses long gaps while retaining salient firings and (3) a stack of SED-based Spike Transformer blocks enable intra-series dependency modeling with a membrane-based linear attention driven by EA-LIF spiking features. Experiments on public telemetry IMTS datasets show that SEDformer attains state-of-the-art forecasting accuracy while reducing energy and memory usage, providing a natural and efficient path for modeling IMTS.", "authors": ["Ziyu Zhou", "Yuchen Fang", "Weilin Ruan", "Shiyu Wang", "James Kwok", "Yuxuan Liang"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2026-02-02", "url": "https://arxiv.org/abs/2602.02230", "pdf_url": "https://arxiv.org/pdf/2602.02230v2", "arxiv_id": "2602.02230", "doi": "10.48550/arXiv.2602.02230", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.4618}
{"id": "0f8d0c64456ade09e8b431c207c62401c24df950d886db75485cb262cefa66c8", "sources": ["arxiv", "semantic_scholar"], "title": "Seg-MoE: Multi-Resolution Segment-wise Mixture-of-Experts for Time Series Forecasting Transformers", "abstract": "Transformer-based models have recently made significant advances in accurate time-series forecasting, but even these architectures struggle to scale efficiently while capturing long-term temporal dynamics. Mixture-of-Experts (MoE) layers are a proven solution to scaling problems in natural language processing. However, existing MoE approaches for time-series forecasting rely on token-wise routing mechanisms, which may fail to exploit the natural locality and continuity of temporal data. In this work, we introduce Seg-MoE, a sparse MoE design that routes and processes contiguous time-step segments rather than making independent expert decisions. Token segments allow each expert to model intra-segment interactions directly, naturally aligning with inherent temporal patterns. We integrate Seg-MoE layers into a time-series Transformer and evaluate it on multiple multivariate long-term forecasting benchmarks. Seg-MoE consistently achieves state-of-the-art forecasting accuracy across almost all prediction horizons, outperforming both dense Transformers and prior token-wise MoE models. Comprehensive ablation studies confirm that segment-level routing is the key factor driving these gains. Our results show that aligning the MoE routing granularity with the inherent structure of time series provides a powerful, yet previously underexplored, inductive bias, opening new avenues for conditionally sparse architectures in sequential data modeling.", "authors": ["Evandro S. Ortigossa", "Eran Segal"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2026-01-29", "url": "https://arxiv.org/abs/2601.21641", "pdf_url": "https://arxiv.org/pdf/2601.21641v2", "arxiv_id": "2601.21641", "doi": "10.48550/arXiv.2601.21641", "citation_count": 2, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.4572}
{"id": "a195cffd2595d091b07b596ec82840e106441496ffe1e05e8a00dc8e6c78242e", "sources": ["arxiv", "semantic_scholar"], "title": "TimeCatcher: A Variational Framework for Volatility-Aware Forecasting of Non-Stationary Time Series", "abstract": "Recent lightweight MLP-based models have achieved strong performance in time series forecasting by capturing stable trends and seasonal patterns. However, their effectiveness hinges on an implicit assumption of local stationarity assumption, making them prone to errors in long-term forecasting of highly non-stationary series, especially when abrupt fluctuations occur, a common challenge in domains like web traffic monitoring. To overcome this limitation, we propose TimeCatcher, a novel Volatility-Aware Variational Forecasting framework. TimeCatcher extends linear architectures with a variational encoder to capture latent dynamic patterns hidden in historical data and a volatility-aware enhancement mechanism to detect and amplify significant local variations. Experiments on nine real-world datasets from traffic, financial, energy, and weather domains show that TimeCatcher consistently outperforms state-of-the-art baselines, with particularly large improvements in long-term forecasting scenarios characterized by high volatility and sudden fluctuations. Our code is available at https://github.com/ColaPrinceCHEN/TimeCatcher.", "authors": ["Zhiyu Chen", "Minhao Liu", "Yanru Zhang"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2026-01-28", "url": "https://arxiv.org/abs/2601.20448", "pdf_url": "https://arxiv.org/pdf/2601.20448v1", "arxiv_id": "2601.20448", "doi": "10.48550/arXiv.2601.20448", "citation_count": 0, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/ColaPrinceCHEN/TimeCatcher", "venue": "arXiv.org", "quality_score": 0.7048}
{"id": "4e0967db5e44b07db6bf9d4f84c322a4b2e3cc68b982e9ba922783899b7f7267", "sources": ["arxiv", "semantic_scholar"], "title": "PatchFormer: A Patch-Based Time Series Foundation Model with Hierarchical Masked Reconstruction and Cross-Domain Transfer Learning for Zero-Shot Multi-Horizon Forecasting", "abstract": "Time series forecasting is a fundamental problem with applications in climate, energy, healthcare, and finance. Many existing approaches require domain-specific feature engineering and substantial labeled data for each task. We introduce PatchFormer, a patch-based time series foundation model that uses hierarchical masked reconstruction for self-supervised pretraining and lightweight adapters for efficient transfer. PatchFormer segments time series into patches and learns multiscale temporal representations with learnable aggregation across temporal scales. Pretraining uses masked patch reconstruction with dynamic masking and objectives that encourage both local accuracy and global consistency, followed by cross-domain knowledge distillation. Experiments on 24 benchmark datasets spanning weather, energy, traffic, finance, and healthcare demonstrate state-of-the-art zero-shot multi-horizon forecasting, reducing mean squared error by 27.3 percent relative to strong baselines while requiring 94 percent less task-specific training data. The model exhibits near log-linear scaling with more pretraining data up to 100 billion points and processes length-512 sequences 3.8x faster than full-sequence transformers.", "authors": ["Olaf Yunus Laitinen Imanov", "Derya Umut Kulali", "Taner Yilmaz"], "categories": ["cs.LG", "eess.SP"], "fields_of_study": ["Computer Science", "Engineering"], "published_date": "2026-01-28", "url": "https://arxiv.org/abs/2601.20845", "pdf_url": "https://arxiv.org/pdf/2601.20845v1", "arxiv_id": "2601.20845", "doi": "10.48550/arXiv.2601.20845", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.456}
{"id": "4245936eb2f615a7fa1aaf99235e0a5783c9154617e8808bca7f098ea5201145", "sources": ["arxiv", "semantic_scholar"], "title": "ScatterFusion: A Hierarchical Scattering Transform Framework for Enhanced Time Series Forecasting", "abstract": "Time series forecasting presents significant challenges due to the complex temporal dependencies at multiple time scales. This paper introduces ScatterFusion, a novel framework that synergistically integrates scattering transforms with hierarchical attention mechanisms for robust time series forecasting. Our approach comprises four key components: (1) a Hierarchical Scattering Transform Module (HSTM) that extracts multi-scale invariant features capturing both local and global patterns; (2) a Scale-Adaptive Feature Enhancement (SAFE) module that dynamically adjusts feature importance across different scales; (3) a Multi-Resolution Temporal Attention (MRTA) mechanism that learns dependencies at varying time horizons; and (4) a Trend-Seasonal-Residual (TSR) decomposition-guided structure-aware loss function. Extensive experiments on seven benchmark datasets demonstrate that ScatterFusion outperforms other common methods, achieving significant reductions in error metrics across various prediction horizons.", "authors": ["Wei Li"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2026-01-28", "url": "https://arxiv.org/abs/2601.20401", "pdf_url": "https://arxiv.org/pdf/2601.20401v1", "arxiv_id": "2601.20401", "doi": "10.48550/arXiv.2601.20401", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "IEEE International Conference on Acoustics, Speech, and Signal Processing", "quality_score": 0.456}
{"id": "121f985d57a8d66848605143e79d8a90c3d0d3d29b5a8cd83cfab07144a1b2d4", "sources": ["arxiv", "semantic_scholar"], "title": "Intermittent time series forecasting: local vs global models", "abstract": "Forecasting intermittent time series, which contain zeros, is a crucial challenge in supply chains as inventory policies require probabilistic forecasts to establish safety levels. Intermittent time series are commonly forecast using local models, trained individually on each time series. In the last years global models, trained on a large collection of time series, have become popular for time series forecasting. Global models are often based on neural networks or gradient boosted trees. We carry out the first study comparing state-of-the-art probabilistic local and global models on intermittent time series. For global models we consider three different distribution heads suitable for intermittent time series: negative binomial, hurdle-shifted negative binomial and Tweedie. To the best of our knowledge, this is the first use of the latter two with neural networks. We perform experiments on five datasets comprising overall more than 40'000 real-world time series. Among global models, TiDE, a simple neural network architecture, achieves the best accuracy; it also consistently outperforms local models and has lower computational requirements. Large global models are instead much more computationally demanding and less accurate. Among the distribution heads, the Tweedie provides the best estimates of the highest quantiles.", "authors": ["Stefano Damato", "Nicolò Rubattu", "Dario Azzimonti", "Giorgio Corani"], "categories": ["stat.ML", "cs.LG"], "fields_of_study": ["Mathematics", "Computer Science"], "published_date": "2026-01-20", "url": "https://arxiv.org/abs/2601.14031", "pdf_url": "https://arxiv.org/pdf/2601.14031v2", "arxiv_id": "2601.14031", "doi": "10.48550/arXiv.2601.14031", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.4469}
{"id": "998bea0a41dd850b2327810d99662b63ec79b12db0cc016ecf29b3c41ec73e84", "sources": ["arxiv", "semantic_scholar"], "title": "Distilling Time Series Foundation Models for Efficient Forecasting", "abstract": "Time Series foundation models (TSFMs) deliver strong forecasting performance through large-scale pretraining, but their large parameter sizes make deployment costly. While knowledge distillation offers a natural and effective approach for model compression, techniques developed for general machine learning tasks are not directly applicable to time series forecasting due to the unique characteristics. To address this, we present DistilTS, the first distillation framework specifically designed for TSFMs. DistilTS addresses two key challenges: (1) task difficulty discrepancy, specific to forecasting, where uniform weighting makes optimization dominated by easier short-term horizons, while long-term horizons receive weaker supervision; and (2) architecture discrepancy, a general challenge in distillation, for which we design an alignment mechanism in the time series forecasting. To overcome these issues, DistilTS introduces horizon-weighted objectives to balance learning across horizons, and a temporal alignment strategy that reduces architectural mismatch, enabling compact models. Experiments on multiple benchmarks demonstrate that DistilTS achieves forecasting performance comparable to full-sized TSFMs, while reducing parameters by up to 1/150 and accelerating inference by up to 6000x. Code is available at: https://github.com/itsnotacie/DistilTS-ICASSP2026.", "authors": ["Yuqi Li", "Kuiye Ding", "Chuanguang Yang", "Szu-Yu Chen", "Yingli Tian"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2026-01-19", "url": "https://arxiv.org/abs/2601.12785", "pdf_url": "https://arxiv.org/pdf/2601.12785v1", "arxiv_id": "2601.12785", "doi": "10.48550/arXiv.2601.12785", "citation_count": 4, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/itsnotacie/DistilTS-ICASSP2026", "venue": "IEEE International Conference on Acoustics, Speech, and Signal Processing", "quality_score": 0.6889}
{"id": "60b5ff7a3aedd3fec3979e4a8999ae53329758c331649447e7864244c66a3e39", "sources": ["arxiv", "semantic_scholar"], "title": "Trend-Adjusted Time Series Models with an Application to Gold Price Forecasting", "abstract": "Time series data play a critical role in various fields, including finance, healthcare, marketing, and engineering. A wide range of techniques (from classical statistical models to neural network-based approaches such as Long Short-Term Memory (LSTM)) have been employed to address time series forecasting challenges. In this paper, we reframe time series forecasting as a two-part task: (1) predicting the trend (directional movement) of the time series at the next time step, and (2) forecasting the quantitative value at the next time step. The trend can be predicted using a binary classifier, while quantitative values can be forecasted using models such as LSTM and Bidirectional Long Short-Term Memory (Bi-LSTM). Building on this reframing, we propose the Trend-Adjusted Time Series (TATS) model, which adjusts the forecasted values based on the predicted trend provided by the binary classifier. We validate the proposed approach through both theoretical analysis and empirical evaluation. The TATS model is applied to a volatile financial time series (the daily gold price) with the objective of forecasting the next days price. Experimental results demonstrate that TATS consistently outperforms standard LSTM and Bi-LSTM models by achieving significantly lower forecasting error. In addition, our results indicate that commonly used metrics such as MSE and MAE are insufficient for fully assessing time series model performance. Therefore, we also incorporate trend detection accuracy, which measures how effectively a model captures trends in a time series.", "authors": ["Sina Kazemdehbashi"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2026-01-19", "url": "https://arxiv.org/abs/2601.12706", "pdf_url": "https://arxiv.org/pdf/2601.12706v2", "arxiv_id": "2601.12706", "doi": "10.48550/arXiv.2601.12706", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.4457}
{"id": "9e07c3f3ec17575e9bd9df8a0bb5cad5ad0199c9a96b8993c41c807c7e66b07c", "sources": ["arxiv", "semantic_scholar"], "title": "Patch-Level Tokenization with CNN Encoders and Attention for Improved Transformer Time-Series Forecasting", "abstract": "Transformer-based models have shown strong performance in time-series forecasting by leveraging self-attention to model long-range temporal dependencies. However, their effectiveness depends critically on the quality and structure of input representations derived from raw multivariate time-series data, particularly as sequence length and data scale increase. This paper proposes a two-stage forecasting framework that explicitly separates local temporal representation learning from global dependency modelling. In the proposed approach, a convolutional neural network operates on fixed-length temporal patches to extract short-range temporal dynamics and non-linear feature interactions, producing compact patch-level token embeddings. Token-level self-attention is applied during representation learning to refine these embeddings, after which a Transformer encoder models inter-patch temporal dependencies to generate forecasts. The method is evaluated on a synthetic multivariate time-series dataset with controlled static and dynamic factors, using an extended sequence length and a larger number of samples. Experimental results demonstrate that the proposed framework consistently outperforms a convolutional baseline under increased temporal context and remains competitive with a strong patch-based Transformer model. These findings indicate that structured patch-level tokenization provides a scalable and effective representation for multivariate time-series forecasting, particularly when longer input sequences are considered.", "authors": ["Saurish Nagrath", "Saroj Kumar Panigrahy"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2026-01-18", "url": "https://arxiv.org/abs/2601.12467", "pdf_url": "https://arxiv.org/pdf/2601.12467v3", "arxiv_id": "2601.12467", "doi": "10.48550/arXiv.2601.12467", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.4446}
{"id": "6acac29a7149291f600b5c9a191c10971fc05507a6cc0f79b616906267991d38", "sources": ["arxiv", "semantic_scholar"], "title": "Shapelets-Enriched Selective Forecasting using Time Series Foundation Models", "abstract": "Time series foundation models have recently gained a lot of attention due to their ability to model complex time series data encompassing different domains including traffic, energy, and weather. Although they exhibit strong average zero-shot performance on forecasting tasks, their predictions on certain critical regions of the data are not always reliable, limiting their usability in real-world applications, especially when data exhibits unique trends. In this paper, we propose a selective forecasting framework to identify these critical segments of time series using shapelets. We learn shapelets using shift-invariant dictionary learning on the validation split of the target domain dataset. Utilizing distance-based similarity to these shapelets, we facilitate the user to selectively discard unreliable predictions and be informed of the model's realistic capabilities. Empirical results on diverse benchmark time series datasets demonstrate that our approach leveraging both zero-shot and full-shot fine-tuned models reduces the overall error by an average of 22.17% for zero-shot and 22.62% for full-shot fine-tuned model. Furthermore, our approach using zero-shot and full-shot fine-tuned models, also outperforms its random selection counterparts by up to 21.41% and 21.43% on one of the datasets.", "authors": ["Shivani Tomar", "Seshu Tirupathi", "Elizabeth Daly", "Ivana Dusparic"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2026-01-16", "url": "https://arxiv.org/abs/2601.11821", "pdf_url": "https://arxiv.org/pdf/2601.11821v1", "arxiv_id": "2601.11821", "doi": "10.48550/arXiv.2601.11821", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.4423}
{"id": "53cea33f0362b098d80f9d42407c4edba16a32377dcfd0da6b931148d4f9c349", "sources": ["arxiv", "semantic_scholar"], "title": "ProbFM: Probabilistic Time Series Foundation Model with Uncertainty Decomposition", "abstract": "Time Series Foundation Models (TSFMs) have emerged as a promising approach for zero-shot financial forecasting, demonstrating strong transferability and data efficiency gains. However, their adoption in financial applications is hindered by fundamental limitations in uncertainty quantification: current approaches either rely on restrictive distributional assumptions, conflate different sources of uncertainty, or lack principled calibration mechanisms. While recent TSFMs employ sophisticated techniques such as mixture models, Student's t-distributions, or conformal prediction, they fail to address the core challenge of providing theoretically-grounded uncertainty decomposition. For the very first time, we present a novel transformer-based probabilistic framework, ProbFM (probabilistic foundation model), that leverages Deep Evidential Regression (DER) to provide principled uncertainty quantification with explicit epistemic-aleatoric decomposition. Unlike existing approaches that pre-specify distributional forms or require sampling-based inference, ProbFM learns optimal uncertainty representations through higher-order evidence learning while maintaining single-pass computational efficiency. To rigorously evaluate the core DER uncertainty quantification approach independent of architectural complexity, we conduct an extensive controlled comparison study using a consistent LSTM architecture across five probabilistic methods: DER, Gaussian NLL, Student's-t NLL, Quantile Loss, and Conformal Prediction. Evaluation on cryptocurrency return forecasting demonstrates that DER maintains competitive forecasting accuracy while providing explicit epistemic-aleatoric uncertainty decomposition. This work establishes both an extensible framework for principled uncertainty quantification in foundation models and empirical evidence for DER's effectiveness in financial applications.", "authors": ["Arundeep Chinta", "Lucas Vinh Tran", "Jay Katukuri"], "categories": ["cs.LG", "cs.AI", "q-fin.RM", "q-fin.TR"], "fields_of_study": ["Computer Science", "Economics"], "published_date": "2026-01-15", "url": "https://arxiv.org/abs/2601.10591", "pdf_url": "https://arxiv.org/pdf/2601.10591v1", "arxiv_id": "2601.10591", "doi": "10.48550/arXiv.2601.10591", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.4411}
{"id": "5a77e80b128258237466e59e7ce541e12123271559b3f531e994fa25e0717daa", "sources": ["arxiv", "semantic_scholar"], "title": "The Promise of Time-Series Foundation Models for Agricultural Forecasting: Evidence from Commodity Prices", "abstract": "Forecasting agricultural markets remains challenging due to nonlinear dynamics, structural breaks, and sparse data. A long-standing belief holds that simple time-series methods outperform more advanced alternatives. This paper provides the first systematic evidence that this belief no longer holds with modern time-series foundation models (TSFMs). Using USDA ERS monthly commodity price data from 1997-2025, we evaluate 17 forecasting approaches across four model classes, including traditional time-series, machine learning, deep learning, and five state-of-the-art TSFMs (Chronos, Chronos-2, TimesFM 2.5, Time-MoE, Moirai-2), and construct annual marketing year price predictions to compare with USDA's futures-based season-average price (SAP) forecasts. We show that zero-shot foundation models consistently outperform traditional time-series methods, machine learning, and deep learning architectures trained from scratch in both monthly and annual forecasting. Furthermore, foundation models remarkably outperform USDA's futures-based forecasts on three of four major commodities despite USDA's information advantage from forward-looking futures markets. Time-MoE delivers the largest accuracy gains, achieving 54.9% improvement on wheat and 18.5% improvement on corn relative to USDA ERS benchmarks on recent data (2017-2024 excluding COVID). These results point to a paradigm shift in agricultural forecasting.", "authors": ["Le Wang", "Boyuan Zhang"], "categories": ["econ.EM", "stat.AP"], "fields_of_study": ["Economics", "Mathematics"], "published_date": "2026-01-10", "url": "https://arxiv.org/abs/2601.06371", "pdf_url": "https://arxiv.org/pdf/2601.06371v2", "arxiv_id": "2601.06371", "doi": null, "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.2771}
{"id": "d0f0c435336580cfa09bc3e038d235ed2a3c48baf616e51b2e21e18833030447", "sources": ["arxiv", "semantic_scholar"], "title": "A Unified Shape-Aware Foundation Model for Time Series Classification", "abstract": "Foundation models pre-trained on large-scale source datasets are reshaping the traditional training paradigm for time series classification. However, existing time series foundation models primarily focus on forecasting tasks and often overlook classification-specific challenges, such as modeling interpretable shapelets that capture class-discriminative temporal features. To bridge this gap, we propose UniShape, a unified shape-aware foundation model designed for time series classification. UniShape incorporates a shape-aware adapter that adaptively aggregates multiscale discriminative subsequences (shapes) into class tokens, effectively selecting the most relevant subsequence scales to enhance model interpretability. Meanwhile, a prototype-based pretraining module is introduced to jointly learn instance- and shape-level representations, enabling the capture of transferable shape patterns. Pre-trained on a large-scale multi-domain time series dataset comprising 1.89 million samples, UniShape exhibits superior generalization across diverse target domains. Experiments on 128 UCR datasets and 30 additional time series datasets demonstrate that UniShape achieves state-of-the-art classification performance, with interpretability and ablation analyses further validating its effectiveness.", "authors": ["Zhen Liu", "Yucheng Wang", "Boyuan Li", "Junhao Zheng", "Emadeldeen Eldele", "Min Wu", "Qianli Ma"], "categories": ["cs.LG", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2026-01-10", "url": "https://arxiv.org/abs/2601.06429", "pdf_url": "https://arxiv.org/pdf/2601.06429v1", "arxiv_id": "2601.06429", "doi": "10.48550/arXiv.2601.06429", "citation_count": 2, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "AAAI Conference on Artificial Intelligence", "quality_score": 0.4354}
{"id": "c4f61dc414f5f6126d5947ab11338f2549c13a202888c9d96501920c803d3292", "sources": ["arxiv", "semantic_scholar"], "title": "Explainable time-series forecasting with sampling-free SHAP for Transformers", "abstract": "Time-series forecasts are essential for planning and decision-making in many domains. Explainability is key to building user trust and meeting transparency requirements. Shapley Additive Explanations (SHAP) is a popular explainable AI framework, but it lacks efficient implementations for time series and often assumes feature independence when sampling counterfactuals. We introduce SHAPformer, an accurate, fast and sampling-free explainable time-series forecasting model based on the Transformer architecture. It leverages attention manipulation to make predictions based on feature subsets. SHAPformer generates explanations in under one second, several orders of magnitude faster than the SHAP Permutation Explainer. On synthetic data with ground truth explanations, SHAPformer provides explanations that are true to the data. Applied to real-world electrical load data, it achieves competitive predictive performance and delivers meaningful local and global insights, such as identifying the past load as the key predictor and revealing a distinct model behavior during the Christmas period.", "authors": ["Matthias Hertel", "Sebastian Pütz", "Ralf Mikut", "Veit Hagenmeyer", "Benjamin Schäfer"], "categories": ["cs.LG"], "fields_of_study": ["Medicine", "Computer Science"], "published_date": "2025-12-23", "url": "https://arxiv.org/abs/2512.20514", "pdf_url": "https://arxiv.org/pdf/2512.20514v1", "arxiv_id": "2512.20514", "doi": "10.1038/s41467-026-73243-5", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "Nature Communications", "quality_score": 0.4148}
{"id": "7b9a965f0a9d839558336958ad69064df99f49ce27b0e97a76030e55b75536a6", "sources": ["arxiv", "semantic_scholar"], "title": "Conversational Time Series Foundation Models: Towards Explainable and Effective Forecasting", "abstract": "The proliferation of time series foundation models has created a landscape where no single method achieves consistent superiority, framing the central challenge not as finding the best model, but as orchestrating an optimal ensemble with interpretability. While Large Language Models (LLMs) offer powerful reasoning capabilities, their direct application to time series forecasting has proven ineffective. We address this gap by repositioning the LLM as an intelligent judge that evaluates, explains, and strategically coordinates an ensemble of foundation models. To overcome the LLM's inherent lack of domain-specific knowledge on time series, we introduce an R1-style finetuning process, guided by SHAP-based faithfulness scores, which teaches the model to interpret ensemble weights as meaningful causal statements about temporal dynamics. The trained agent then engages in iterative, multi-turn conversations to perform forward-looking assessments, provide causally-grounded explanations for its weighting decisions, and adaptively refine the optimization strategy. Validated on the GIFT-Eval benchmark on 23 datasets across 97 settings, our approach significantly outperforms leading time series foundation models on both CRPS and MASE metrics, establishing new state-of-the-art results.", "authors": ["Defu Cao", "Michael Gee", "Jinbo Liu", "Hengxuan Wang", "Wei Yang", "Rui Wang", "Yan Liu"], "categories": ["cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-12-17", "url": "https://arxiv.org/abs/2512.16022", "pdf_url": "https://arxiv.org/pdf/2512.16022v1", "arxiv_id": "2512.16022", "doi": "10.48550/arXiv.2512.16022", "citation_count": 7, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.4079}
{"id": "7a4eab820bac7c2c90e33a157628a1c87fef0b3268e376813c159b19f8861e02", "sources": ["arxiv", "semantic_scholar"], "title": "Adaptive Information Routing for Multimodal Time Series Forecasting", "abstract": "Time series forecasting is a critical task for artificial intelligence with numerous real-world applications. Traditional approaches primarily rely on historical time series data to predict the future values. However, in practical scenarios, this is often insufficient for accurate predictions due to the limited information available. To address this challenge, multimodal time series forecasting methods which incorporate additional data modalities, mainly text data, alongside time series data have been explored. In this work, we introduce the Adaptive Information Routing (AIR) framework, a novel approach for multimodal time series forecasting. Unlike existing methods that treat text data on par with time series data as interchangeable auxiliary features for forecasting, AIR leverages text information to dynamically guide the time series model by controlling how and to what extent multivariate time series information should be combined. We also present a text-refinement pipeline that employs a large language model to convert raw text data into a form suitable for multimodal forecasting, and we introduce a benchmark that facilitates multimodal forecasting experiments based on this pipeline. Experiment results with the real world market data such as crude oil price and exchange rates demonstrate that AIR effectively modulates the behavior of the time series model using textual inputs, significantly enhancing forecasting accuracy in various time series forecasting tasks.", "authors": ["Jun Seo", "Hyeokjun Choe", "Seohui Bae", "Soyeon Park", "Wonbin Ahn", "Taeyoon Lim", "Junhyeok Kang", "Sangjun Han", "Jaehoon Lee", "Dongwan Kang", "Minjae Kim", "Sungdong Yoo", "Soonyoung Lee"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-12-11", "url": "https://arxiv.org/abs/2512.10229", "pdf_url": "https://arxiv.org/pdf/2512.10229v3", "arxiv_id": "2512.10229", "doi": "10.48550/arXiv.2512.10229", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.401}
{"id": "453d93ea0bd94c0d8ceb75dc403f6f70f1d37a5dc4a0ab0e421452f89c5e2e8d", "sources": ["arxiv", "semantic_scholar"], "title": "Time Series Foundation Models for Process Model Forecasting", "abstract": "Process Model Forecasting (PMF) aims to predict how the control-flow structure of a process evolves over time by modeling the temporal dynamics of directly-follows (DF) relations, complementing predictive process monitoring that focuses on single-case prefixes. Prior benchmarks show that machine learning and deep learning models provide only modest gains over statistical baselines, mainly due to the sparsity and heterogeneity of the DF time series. We investigate Time Series Foundation Models (TSFMs), large pre-trained models for generic time series, as an alternative for PMF. Using DF time series derived from real-life event logs, we compare zero-shot use of TSFMs, without additional training, with fine-tuned variants adapted on PMF-specific data. TSFMs generally achieve lower forecasting errors (MAE and RMSE) than traditional and specialized models trained from scratch on the same logs, indicating effective transfer of temporal structure from non-process domains. While fine-tuning can further improve accuracy, the gains are often small and may disappear on smaller or more complex datasets, so zero-shot use remains a strong default. Our study highlights the generalization capability and data efficiency of TSFMs for process-related time series and, to the best of our knowledge, provides the first systematic evaluation of temporal foundation models for PMF.", "authors": ["Yongbo Yu", "Jari Peeperkorn", "Johannes De Smedt", "Jochen De Weerdt"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-12-08", "url": "https://arxiv.org/abs/2512.07624", "pdf_url": "https://arxiv.org/pdf/2512.07624v1", "arxiv_id": "2512.07624", "doi": "10.48550/arXiv.2512.07624", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.3976}
{"id": "d1a21263adea07c9f7347c770db584856f3901611d8ef6ec92f102c747e5da04", "sources": ["arxiv", "semantic_scholar"], "title": "In-Context and Few-Shots Learning for Forecasting Time Series Data based on Large Language Models", "abstract": "Existing data-driven approaches in modeling and predicting time series data include ARIMA (Autoregressive Integrated Moving Average), Transformer-based models, LSTM (Long Short-Term Memory) and TCN (Temporal Convolutional Network). These approaches, and in particular deep learning-based models such as LSTM and TCN, have shown great results in predicting time series data. With the advancement of leveraging pre-trained foundation models such as Large Language Models (LLMs) and more notably Google's recent foundation model for time series data, {\\it TimesFM} (Time Series Foundation Model), it is of interest to investigate whether these foundation models have the capability of outperforming existing modeling approaches in analyzing and predicting time series data. This paper investigates the performance of using LLM models for time series data prediction. We investigate the in-context learning methodology in the training of LLM models that are specific to the underlying application domain. More specifically, the paper explores training LLMs through in-context, zero-shot and few-shot learning and forecasting time series data with OpenAI {\\tt o4-mini} and Gemini 2.5 Flash Lite, as well as the recent Google's Transformer-based TimesFM, a time series-specific foundation model, along with two deep learning models, namely TCN and LSTM networks. The findings indicate that TimesFM has the best overall performance with the lowest RMSE value (0.3023) and the competitive inference time (266 seconds). Furthermore, OpenAI's o4-mini also exhibits a good performance based on Zero Shot learning. These findings highlight pre-trained time series foundation models as a promising direction for real-time forecasting, enabling accurate and scalable deployment with minimal model adaptation.", "authors": ["Saroj Gopali", "Bipin Chhetri", "Deepika Giri", "Sima Siami-Namini", "Akbar Siami Namin"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-12-08", "url": "https://arxiv.org/abs/2512.07705", "pdf_url": "https://arxiv.org/pdf/2512.07705v1", "arxiv_id": "2512.07705", "doi": "10.1109/BigData66926.2025.11401073", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "BigData Congress [Services Society]", "quality_score": 0.3976}
{"id": "076eb10c9b44e8d3f571c13281929d29efecb57f25b088538eda6c796a8beae9", "sources": ["arxiv", "semantic_scholar"], "title": "Robust Tabular Foundation Models", "abstract": "The development of tabular foundation models (TFMs) has accelerated in recent years, showing strong potential to outperform traditional ML methods for structured data. A key finding is that TFMs can be pretrained entirely on synthetic datasets, opening opportunities to design data generators that encourage desirable model properties. Prior work has mainly focused on crafting high-quality priors over generators to improve overall pretraining performance. Our insight is that parameterizing the generator distribution enables an adversarial robustness perspective: during training, we can adapt the generator to emphasize datasets that are particularly challenging for the model. We formalize this by introducing an optimality gap measure, given by the difference between TFM performance and the best achievable performance as estimated by strong baselines such as XGBoost, CatBoost, and Random Forests. Building on this idea, we propose Robust Tabular Foundation Models (RTFM), a model-agnostic adversarial training framework. Applied to the TabPFN V2 classifier, RTFM improves benchmark performance, with up to a 6% increase in mean normalized AUC over the original TabPFN and other baseline algorithms, while requiring less than 100k additional synthetic datasets. These results highlight a promising new direction for targeted adversarial training and fine-tuning of TFMs using synthetic data alone.", "authors": ["Matthew Peroni", "Franck Le", "Vadim Sheinin"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-12-02", "url": "https://arxiv.org/abs/2512.03307", "pdf_url": "https://arxiv.org/pdf/2512.03307v1", "arxiv_id": "2512.03307", "doi": "10.48550/arXiv.2512.03307", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.3907}
{"id": "fcd97c78b6d7efea5ff00c1072d6e83ba654fbcc102edeefa4ca5e1ea3af0440", "sources": ["arxiv", "semantic_scholar"], "title": "CLEF: Clinically-Guided Contrastive Learning for Electrocardiogram Foundation Models", "abstract": "The electrocardiogram (ECG) is a key diagnostic tool in cardiovascular health. Single-lead ECG recording is integrated into both clinical-grade and consumer wearables. While self-supervised pretraining of foundation models on unlabeled ECGs improves diagnostic performance, existing approaches do not incorporate domain knowledge from clinical metadata. We introduce a novel contrastive learning approach that utilizes an established clinical risk score to adaptively weight negative pairs: clinically-guided contrastive learning. It aligns the similarities of ECG embeddings with clinically meaningful differences between subjects, with an explicit mechanism to handle missing metadata. On 12-lead ECGs from 161K patients in the MIMIC-IV dataset, we pretrain single-lead ECG foundation models at three scales, collectively called CLEF, using only routinely collected metadata without requiring per-sample ECG annotations. We evaluate CLEF on 18 clinical classification and regression tasks across 7 held-out datasets, and benchmark against 5 foundation model baselines and 3 self-supervised algorithms. When pretrained on 12-lead ECG data and tested on lead-I data, CLEF outperforms self-supervised foundation model baselines: the medium-sized CLEF achieves average AUROC improvements of at least 2.6% in classification and average reductions in MAEs of at least 3.2% in regression. Comparing with existing self-supervised learning algorithms, CLEF improves the average AUROC by at least 1.8%. Moreover, when pretrained only on lead-I data for classification tasks, CLEF performs comparably to the state-of-the-art ECGFounder, which was trained in a supervised manner. Overall, CLEF enables more accurate and scalable single-lead ECG analysis, advancing remote health monitoring. Code and pretrained CLEF models are available at: github.com/Nokia-Bell-Labs/ecg-foundation-model.", "authors": ["Yuxuan Shu", "Peter H. Charlton", "Fahim Kawsar", "Jussi Hernesniemi", "Mohammad Malekzadeh"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-12-01", "url": "https://arxiv.org/abs/2512.02180", "pdf_url": "https://arxiv.org/pdf/2512.02180v1", "arxiv_id": "2512.02180", "doi": "10.48550/arXiv.2512.02180", "citation_count": 3, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/Nokia-Bell-Labs/ecg-foundation-model", "venue": "arXiv.org", "quality_score": 0.6021}
{"id": "0a61099ae12b410148d71c5c49dbb0ae2696cb22f1e8fd2ade673ece04ae701f", "sources": ["arxiv", "semantic_scholar"], "title": "Zero-Shot Transfer Capabilities of the Sundial Foundation Model for Leaf Area Index Forecasting", "abstract": "This work investigates the zero-shot forecasting capability of time series foundation models for Leaf Area Index (LAI) forecasting in agricultural monitoring. Using the HiQ dataset (U.S., 2000-2022), we systematically compare statistical baselines, a fully supervised LSTM, and the Sundial foundation model under multiple evaluation protocols. We find that Sundial, in the zero-shot setting, can outperform a fully trained LSTM provided that the input context window is sufficiently long-specifically, when covering more than one or two full seasonal cycles. We show that a general-purpose foundation model can surpass specialized supervised models on remote-sensing time series prediction without any task-specific tuning. These results highlight the strong potential of pretrained time series foundation models to serve as effective plug-and-play forecasters in agricultural and environmental applications.", "authors": ["Peining Zhang", "Hongchen Qin", "Haochen Zhang", "Ziqi Guo", "Guiling Wang", "Jinbo Bi"], "categories": ["cs.LG", "cs.AI", "cs.CV"], "fields_of_study": ["Computer Science"], "published_date": "2025-11-25", "url": "https://arxiv.org/abs/2511.20004", "pdf_url": "https://arxiv.org/pdf/2511.20004v2", "arxiv_id": "2511.20004", "doi": "10.48550/arXiv.2511.20004", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.3827}
{"id": "20e99198add138e01bd93481d5a1a356f4542377f32e31ced758b9bb74e50c6a", "sources": ["arxiv", "semantic_scholar"], "title": "Tiny-TSM: Efficiently Training a Lightweight SOTA Time Series Foundation Model", "abstract": "We present Tiny-TSM, a time series foundation model characterized by small scale, economical training, and state-of-the-art performance. It comprises 23M total parameters, trained on a single A100 GPU in less than a week using a new synthetic data generation and data augmentation pipeline (SynthTS). Without any neural architecture search, hyperparameter tuning, or scaling up model size, Tiny-TSM achieves state-of-the-art performance on a wide range of time series benchmark datasets, often outperforming much larger models and even matching the performance of much larger, industrial-scale, likely highly tuned foundation models. Specifically, Tiny-TSM outperforms all other time series foundation models we evaluated on medium- and long-term forecasting tasks under MSE loss, while short-term accuracy is still competitive with state-of-the-art models. We also introduce a causal input normalization scheme that enables time series models to be trained with dense next-token prediction loss, significantly accelerating convergence speed and reducing training time. All experiments were conducted on a single A100 GPU, illustrating the practicality of the proposed approach in a resource-constrained setting.", "authors": ["Felix Birkel"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-11-24", "url": "https://arxiv.org/abs/2511.19272", "pdf_url": "https://arxiv.org/pdf/2511.19272v1", "arxiv_id": "2511.19272", "doi": "10.48550/arXiv.2511.19272", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.3816}
{"id": "2327d1e3416aebf7bf6550a5cbd73eff8ce2145014fb2b4363b13d21a7bccb81", "sources": ["arxiv", "semantic_scholar"], "title": "TiCT: A Synthetically Pre-Trained Foundation Model for Time Series Classification", "abstract": "The ubiquity of time series data creates a strong demand for general-purpose foundation models, yet developing them for classification remains a significant challenge, largely due to the high cost of labeled data. Foundation models capable of in-context learning (ICL) offer a powerful solution, adapting to new tasks with minimal examples and reducing the need for extensive retraining. However, prior work on large-scale time series models has predominantly focused on forecasting, leaving a critical gap for versatile, fine-tuning-free classification. To address this, we introduce TiCT (Time-series in-Context Transformer), a transformer-based model pre-trained exclusively on synthetic data to perform in-context classification. We make two primary technical contributions: 1) a novel architecture featuring a scalable bit-based label encoding and a special output attention mechanism to handle an arbitrary number of classes; and 2) a synthetic pre-training framework that combines a Mixup-inspired process with data augmentation to foster generalization and noise invariance. Extensive evaluations on the UCR Archive show that TiCT achieves competitive performance against state-of-the-art supervised methods. Crucially, this is accomplished using only in-context examples at inference time, without updating a single model weight.", "authors": ["Chin-Chia Michael Yeh", "Uday Singh Saini", "Junpeng Wang", "Xin Dai", "Xiran Fan", "Jiarui Sun", "Yujie Fan", "Yan Zheng"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-11-24", "url": "https://arxiv.org/abs/2511.19694", "pdf_url": "https://arxiv.org/pdf/2511.19694v2", "arxiv_id": "2511.19694", "doi": "10.48550/arXiv.2511.19694", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.3816}
{"id": "c5dbfa612630a75a6297579a45583ffcdf98558fe0dc3814b15531ebf1ce181a", "sources": ["arxiv", "semantic_scholar"], "title": "KAN vs LSTM Performance in Time Series Forecasting", "abstract": "This paper compares Kolmogorov-Arnold Networks (KAN) and Long Short-Term Memory networks (LSTM) for forecasting non-deterministic stock price data, evaluating predictive accuracy versus interpretability trade-offs using Root Mean Square Error (RMSE).LSTM demonstrates substantial superiority across all tested prediction horizons, confirming their established effectiveness for sequential data modelling. Standard KAN, while offering theoretical interpretability through the Kolmogorov-Arnold representation theorem, exhibits significantly higher error rates and limited practical applicability for time series forecasting. The results confirm LSTM dominance in accuracy-critical time series applications while identifying computational efficiency as KANs' primary advantage in resource-constrained scenarios where accuracy requirements are less stringent. The findings support LSTM adoption for practical financial forecasting while suggesting that continued research into specialised KAN architectures may yield future improvements.", "authors": ["Tabish Ali Rather", "S M Mahmudul Hasan Joy", "Nadezda Sukhorukova", "Federico Frascoli"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-11-23", "url": "https://arxiv.org/abs/2511.18613", "pdf_url": "https://arxiv.org/pdf/2511.18613v1", "arxiv_id": "2511.18613", "doi": "10.48550/arXiv.2511.18613", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.3804}
{"id": "0669f1c82d8e4cf5424ef9f2ab9e08268c702e0692afc844fa2d2e12c07c7380", "sources": ["arxiv", "semantic_scholar"], "title": "Accelerating Time Series Foundation Models with Speculative Decoding", "abstract": "Modern web applications--from real-time content recommendation and dynamic pricing to CDN optimization--increasingly rely on time-series forecasting to deliver personalized experiences to billions of users. Large-scale Transformer-based models have achieved state-of-the-art performance in time-series forecasting but suffer from high computational costs, limiting their deployment in latency-sensitive web applications. To address this challenge, we propose a general inference acceleration framework that adapts speculative decoding to autoregressive time-series models. Our approach employs a smaller \"draft\" model to propose future time-series patches, which are then verified in parallel by a larger \"target\" model, reducing the number of sequential forward passes required. We address key technical challenges in adapting this technique from discrete language tokens to continuous time-series distributions, including the design of acceptance criteria for multivariate Gaussian patches and practical variants that balance efficiency with accuracy. Through experiments on time series forecasting benchmarks relevant to web applications, we demonstrate significant inference speedups while maintaining competitive accuracy. The framework requires no architectural modifications to existing foundation models, making it immediately applicable to accelerate deployed time-series forecasting systems. Our implementation can be found at https://github.com/PranavSubbaraman/STRIDE", "authors": ["Pranav Subbaraman", "Fang Sun", "Yue Yao", "Huacong Tang", "Xiao Luo", "Yizhou Sun"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-11-22", "url": "https://arxiv.org/abs/2511.18191", "pdf_url": "https://arxiv.org/pdf/2511.18191v1", "arxiv_id": "2511.18191", "doi": "10.48550/arXiv.2511.18191", "citation_count": 1, "influential_citation_count": 1, "has_code": true, "code_url": "https://github.com/PranavSubbaraman/STRIDE", "venue": "arXiv.org", "quality_score": 0.5861}
{"id": "0f323f242f8e3020cd0f4391a666e9bcc730c1c0c40b36118d19af54e304f4a8", "sources": ["arxiv", "semantic_scholar"], "title": "Optimal Look-back Horizon for Time Series Forecasting in Federated Learning", "abstract": "Selecting an appropriate look-back horizon remains a fundamental challenge in time series forecasting (TSF), particularly in the federated learning scenarios where data is decentralized, heterogeneous, and often non-independent. While recent work has explored horizon selection by preserving forecasting-relevant information in an intrinsic space, these approaches are primarily restricted to centralized and independently distributed settings. This paper presents a principled framework for adaptive horizon selection in federated time series forecasting through an intrinsic space formulation. We introduce a synthetic data generator (SDG) that captures essential temporal structures in client data, including autoregressive dependencies, seasonality, and trend, while incorporating client-specific heterogeneity. Building on this model, we define a transformation that maps time series windows into an intrinsic representation space with well-defined geometric and statistical properties. We then derive a decomposition of the forecasting loss into a Bayesian term, which reflects irreducible uncertainty, and an approximation term, which accounts for finite-sample effects and limited model capacity. Our analysis shows that while increasing the look-back horizon improves the identifiability of deterministic patterns, it also increases approximation error due to higher model complexity and reduced sample efficiency. We prove that the total forecasting loss is minimized at the smallest horizon where the irreducible loss starts to saturate, while the approximation loss continues to rise. This work provides a rigorous theoretical foundation for adaptive horizon selection for time series forecasting in federated learning.", "authors": ["Dahao Tang", "Nan Yang", "Yanli Li", "Zhiyu Zhu", "Zhibo Jin", "Dong Yuan"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-11-16", "url": "https://arxiv.org/abs/2511.12791", "pdf_url": "https://arxiv.org/pdf/2511.12791v3", "arxiv_id": "2511.12791", "doi": "10.48550/arXiv.2511.12791", "citation_count": 2, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "AAAI Conference on Artificial Intelligence", "quality_score": 0.3724}
{"id": "07bdf3b969bc00f0581c19bda1480e8fce2b4902ac9729aa192dc901b6dd7134", "sources": ["arxiv", "semantic_scholar"], "title": "EMAformer: Enhancing Transformer through Embedding Armor for Time Series Forecasting", "abstract": "Multivariate time series forecasting is crucial across a wide range of domains. While presenting notable progress for the Transformer architecture, iTransformer still lags behind the latest MLP-based models. We attribute this performance gap to unstable inter-channel relationships. To bridge this gap, we propose EMAformer, a simple yet effective model that enhances the Transformer with an auxiliary embedding suite, akin to armor that reinforces its ability. By introducing three key inductive biases, i.e., \\textit{global stability}, \\textit{phase sensitivity}, and \\textit{cross-axis specificity}, EMAformer unlocks the further potential of the Transformer architecture, achieving state-of-the-art performance on 12 real-world benchmarks and reducing forecasting errors by an average of 2.73\\% in MSE and 5.15\\% in MAE. This significantly advances the practical applicability of Transformer-based approaches for multivariate time series forecasting. The code is available on https://github.com/PlanckChang/EMAformer.", "authors": ["Zhiwei Zhang", "Xinyi Du", "Xuanchi Guo", "Weihao Wang", "Wenjuan Han"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-11-11", "url": "https://arxiv.org/abs/2511.08396", "pdf_url": "https://arxiv.org/pdf/2511.08396v1", "arxiv_id": "2511.08396", "doi": "10.48550/arXiv.2511.08396", "citation_count": 2, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/PlanckChang/EMAformer", "venue": "AAAI Conference on Artificial Intelligence", "quality_score": 0.5667}
{"id": "8e0c141037c289ce220bed64a9fbee72a3523cd5d23dd917a4115b9a8732724b", "sources": ["arxiv", "semantic_scholar"], "title": "Bitcoin Forecasting with Classical Time Series Models on Prices and Volatility", "abstract": "This paper evaluates the performance of classical time series models in forecasting Bitcoin prices, focusing on ARIMA, SARIMA, GARCH, and EGARCH. Daily price data from 2010 to 2020 were analyzed, with models trained on the first 90 percent and tested on the final 10 percent. Forecast accuracy was assessed using MAE, RMSE, AIC, and BIC. The results show that ARIMA provided the strongest forecasts for short-run log-price dynamics, while EGARCH offered the best fit for volatility by capturing asymmetry in responses to shocks. These findings suggest that despite Bitcoin's extreme volatility, classical time series models remain valuable for short-run forecasting. The study contributes to understanding cryptocurrency predictability and sets the stage for future work integrating machine learning and macroeconomic variables.", "authors": ["Anmar Kareem", "Alexander Aue"], "categories": ["q-fin.ST"], "fields_of_study": ["Economics"], "published_date": "2025-11-09", "url": "https://arxiv.org/abs/2511.06224", "pdf_url": "https://arxiv.org/pdf/2511.06224v1", "arxiv_id": "2511.06224", "doi": null, "citation_count": 1, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.2319}
{"id": "192780847161989db5bb5103ad967a90a21d26f5f040561a13cb14bedfa9ed86", "sources": ["arxiv", "semantic_scholar"], "title": "Frequency Matters: When Time Series Foundation Models Fail Under Spectral Shift", "abstract": "Time series foundation models (TSFMs) have shown strong results on public benchmarks, prompting comparisons to a \"BERT moment\" for time series. Their effectiveness in industrial settings, however, remains uncertain. We examine why TSFMs often struggle to generalize and highlight spectral shift (a mismatch between the dominant frequency components in downstream tasks and those represented during pretraining) as a key factor. We present evidence from an industrial-scale player engagement prediction task in mobile gaming, where TSFMs underperform domain-adapted baselines. To isolate the mechanism, we design controlled synthetic experiments contrasting signals with seen versus unseen frequency bands, observing systematic degradation under spectral mismatch. These findings position frequency awareness as critical for robust TSFM deployment and motivate new pretraining and evaluation protocols that explicitly account for spectral diversity.", "authors": ["Tianze Wang", "Sofiane Ennadir", "John Pertoft", "Gabriela Zarzar Gandler", "Lele Cao", "Zineb Senane", "Styliani Katsarou", "Sahar Asadi", "Axel Karlsson", "Oleg Smirnov"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-11-06", "url": "https://arxiv.org/abs/2511.05619", "pdf_url": "https://arxiv.org/pdf/2511.05619v1", "arxiv_id": "2511.05619", "doi": "10.48550/arXiv.2511.05619", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.3609}
{"id": "d4d353dfa58d5b07a5a08cac58ef292a84bf5dc448feaed89b014a8d4bbbba78", "sources": ["arxiv", "semantic_scholar"], "title": "Leveraging Generic Time Series Foundation Models for EEG Classification", "abstract": "Foundation models for time series are emerging as powerful general-purpose backbones, yet their potential for domain-specific biomedical signals such as electroencephalography (EEG) remains rather unexplored. In this work, we investigate the applicability a recently proposed time series classification foundation model, to a different EEG tasks such as motor imagery classification and sleep stage prediction. We test two pretraining regimes: (a) pretraining on heterogeneous real-world time series from multiple domains, and (b) pretraining on purely synthetic data. We find that both variants yield strong performance, consistently outperforming EEGNet, a widely used convolutional baseline, and CBraMod, the most recent EEG-specific foundation model. These results suggest that generalist time series foundation models, even when pretrained on data of non-neural origin or on synthetic signals, can transfer effectively to EEG. Our findings highlight the promise of leveraging cross-domain pretrained models for brain signal analysis, suggesting that EEG may benefit from advances in the broader time series literature.", "authors": ["Théo Gnassounou", "Yessin Moakher", "Shifeng Xie", "Vasilii Feofanov", "Ievgen Redko"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-10-31", "url": "https://arxiv.org/abs/2510.27522", "pdf_url": "https://arxiv.org/pdf/2510.27522v1", "arxiv_id": "2510.27522", "doi": "10.48550/arXiv.2510.27522", "citation_count": 3, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.3541}
{"id": "71b3860e50a965c7742515d0a2e0599281696e10c2ce967baa52f032bc975b92", "sources": ["arxiv", "semantic_scholar"], "title": "Pre-trained Forecasting Models: Strong Zero-Shot Feature Extractors for Time Series Classification", "abstract": "Recent research on time series foundation models has primarily focused on forecasting, leaving it unclear how generalizable their learned representations are. In this study, we examine whether frozen pre-trained forecasting models can provide effective representations for classification. To this end, we compare different representation extraction strategies and introduce two model-agnostic embedding augmentations. Our experiments show that the best forecasting models achieve classification accuracy that matches or even surpasses that of state-of-the-art models pre-trained specifically for classification. Moreover, we observe a positive correlation between forecasting and classification performance. These findings challenge the assumption that task-specific pre-training is necessary, and suggest that learning to forecast may provide a powerful route toward constructing general-purpose time series foundation models.", "authors": ["Andreas Auer", "Daniel Klotz", "Sebastinan Böck", "Sepp Hochreiter"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-10-30", "url": "https://arxiv.org/abs/2510.26777", "pdf_url": "https://arxiv.org/pdf/2510.26777v1", "arxiv_id": "2510.26777", "doi": "10.48550/arXiv.2510.26777", "citation_count": 5, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.3529}
{"id": "2664bf0c855da7ed33fc2b389962347a3166532ebb9576614d42d68b0d4bfd6a", "sources": ["arxiv", "semantic_scholar"], "title": "Time-Series Foundation Models for ISP Traffic Forecasting", "abstract": "Accurate network-traffic forecasting enables proactive capacity planning and anomaly detection in Internet Service Provider (ISP) networks. Recent advances in time-series foundation models (TSFMs) have demonstrated strong zero-shot and few-shot generalization across diverse domains, yet their effectiveness for computer networking remains unexplored. This paper presents a systematic evaluation of a TSFM, IBM's Tiny Time Mixer (TTM), on the CESNET-TimeSeries24 dataset, a 40-week real-world ISP telemetry corpus. We assess TTM under zero-shot and few-shot settings across multiple forecasting horizons (hours to days), aggregation hierarchies (institutions, subnets, IPs), and temporal resolutions (10-minute and hourly). Results show that TTM achieves consistent accuracy (RMSE 0.026-0.057) and stable $R^2$ scores across horizons and context lengths, outperforming or matching fully trained deep learning baselines such as GRU and LSTM. Inference latency remains under 0.05s per 100 points on a single MacBook Pro using CPU-only computation, confirming deployability without dedicated GPU or MPS acceleration. These findings highlight the potential of pretrained TSFMs to enable scalable, efficient, and training-free forecasting for modern network monitoring and management systems.", "authors": ["Fan Liu", "Behrooz Farkiani", "Patrick Crowley"], "categories": ["cs.NI"], "fields_of_study": ["Computer Science"], "published_date": "2025-10-29", "url": "https://arxiv.org/abs/2511.17529", "pdf_url": "https://arxiv.org/pdf/2511.17529v2", "arxiv_id": "2511.17529", "doi": "10.48550/arXiv.2511.17529", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.3518}
{"id": "c566fc0c41ac0c1bdd7f0f21332a111f6299efcaeef14080911aedf1d5d6c3de", "sources": ["arxiv", "semantic_scholar"], "title": "Solar flare forecasting with foundational transformer models across image, video, and time-series modalities", "abstract": "We present a comparative study of transformer-based architectures for solar flare forecasting using heterogeneous data modalities, including images, video sequences, and time-series observations. Our analysis evaluates three recent foundational models - SigLIP2 for image encoding, VideoMAE for spatio-temporal video representation, and Moirai2 for multivariate time-series forecasting - applied to publicly available datasets of solar magnetograms from the SDO/HMI mission and soft X-ray fluxes acquired by GOES satellites. All models are trained and validated under consistent data splits and evaluation criteria, with the goal of assessing the strengths and limitations of transformer backbones across spatial and temporal representations of solar activity. We investigate multiple loss formulations (weighted BCE, focal, and score-oriented) and training balance strategies to mitigate class imbalance typical of flare datasets. Results show that while both SigLIP2 and VideoMAE achieve typical performance on image and video data (True Skill Statistic TSS~0.60-0.65), the time-series model Moirai2 reaches superior forecasting skill (TSS~0.74) using irradiance-based temporal evolution alone. These findings highlight the potential of pretrained transformer architectures and cross-modal learning for advancing operational space weather forecasting, paving the way toward unified multimodal models that integrate visual and temporal information.", "authors": ["S. Riggi", "P. Romano", "A. Pilzer", "U. Becciani"], "categories": ["astro-ph.IM", "astro-ph.SR"], "fields_of_study": ["Physics", "Computer Science"], "published_date": "2025-10-27", "url": "https://arxiv.org/abs/2510.23400", "pdf_url": "https://arxiv.org/pdf/2510.23400v2", "arxiv_id": "2510.23400", "doi": "10.1016/j.ascom.2025.101042", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "Astronomy and Computing", "quality_score": 0.3495}
{"id": "f1f6df6a123f4edfa2256e9b7917fde47d71b78ae08439226b2839654e6dff91", "sources": ["arxiv", "semantic_scholar"], "title": "SEMPO: Lightweight Foundation Models for Time Series Forecasting", "abstract": "The recent boom of large pre-trained models witnesses remarkable success in developing foundation models (FMs) for time series forecasting. Despite impressive performance across diverse downstream forecasting tasks, existing time series FMs possess massive network architectures and require substantial pre-training on large-scale datasets, which significantly hinders their deployment in resource-constrained environments. In response to this growing tension between versatility and affordability, we propose SEMPO, a novel lightweight foundation model that requires pretraining on relatively small-scale data, yet exhibits strong general time series forecasting. Concretely, SEMPO comprises two key modules: 1) energy-aware SpEctral decomposition module, that substantially improves the utilization of pre-training data by modeling not only the high-energy frequency signals but also the low-energy yet informative frequency signals that are ignored in current methods; and 2) Mixture-of-PrOmpts enabled Transformer, that learns heterogeneous temporal patterns through small dataset-specific prompts and adaptively routes time series tokens to prompt-based experts for parameter-efficient model adaptation across different datasets and domains. Equipped with these modules, SEMPO significantly reduces both pre-training data scale and model size, while achieving strong generalization. Extensive experiments on two large-scale benchmarks covering 16 datasets demonstrate the superior performance of SEMPO in both zero-shot and few-shot forecasting scenarios compared with state-of-the-art methods. Code and data are available at https://github.com/mala-lab/SEMPO.", "authors": ["Hui He", "Kun Yi", "Yuanchi Ma", "Qi Zhang", "Zhendong Niu", "Guansong Pang"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-10-22", "url": "https://arxiv.org/abs/2510.19710", "pdf_url": "https://arxiv.org/pdf/2510.19710v1", "arxiv_id": "2510.19710", "doi": "10.48550/arXiv.2510.19710", "citation_count": 3, "influential_citation_count": 1, "has_code": true, "code_url": "https://github.com/mala-lab/SEMPO", "venue": "arXiv.org", "quality_score": 0.5312}
{"id": "4fd2b2f51e1cb8661fe0af5b98f39a650bca9a4d88839ec58b4e461dac3513f5", "sources": ["arxiv", "semantic_scholar"], "title": "QKCV Attention: Enhancing Time Series Forecasting with Static Categorical Embeddings for Both Lightweight and Pre-trained Foundation Models", "abstract": "In real-world time series forecasting tasks, category information plays a pivotal role in capturing inherent data patterns. This paper introduces QKCV (Query-Key-Category-Value) attention, an extension of the traditional QKV framework that incorporates a static categorical embedding C to emphasize category-specific information. As a versatile plug-in module, QKCV enhances the forecasting accuracy of attention-based models (e.g., Vanilla Transformer, Informer, PatchTST, TFT) across diverse real-world datasets. Furthermore, QKCV demonstrates remarkable adaptability in fine-tuning univariate time series foundation model by solely updating the static embedding C while preserving pretrained weights, thereby reducing computational overhead and achieving superior fine-tuning performance.", "authors": ["Hao Wang", "Baojun Ma"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-10-21", "url": "https://arxiv.org/abs/2510.20222", "pdf_url": "https://arxiv.org/pdf/2510.20222v1", "arxiv_id": "2510.20222", "doi": "10.48550/arXiv.2510.20222", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.3426}
{"id": "c0d588e03f4c002eacbc24a8a7df8fc700a128e46ebab3ac85bb5e37fed6a669", "sources": ["arxiv", "semantic_scholar"], "title": "Benchmarking Probabilistic Time Series Forecasting Models on Neural Activity", "abstract": "Neural activity forecasting is central to understanding neural systems and enabling closed-loop control. While deep learning has recently advanced the state-of-the-art in the time series forecasting literature, its application to neural activity forecasting remains limited. To bridge this gap, we systematically evaluated eight probabilistic deep learning models, including two foundation models, that have demonstrated strong performance on general forecasting benchmarks. We compared them against four classical statistical models and two baseline methods on spontaneous neural activity recorded from mouse cortex via widefield imaging. Across prediction horizons, several deep learning models consistently outperformed classical approaches, with the best model producing informative forecasts up to 1.5 seconds into the future. Our findings point toward future control applications and open new avenues for probing the intrinsic temporal structure of neural activity.", "authors": ["Ziyu Lu", "Anna J. Li", "Alexander E. Ladd", "Pascha Matveev", "Aditya Deole", "Eric Shea-Brown", "J. Nathan Kutz", "Nicholas A. Steinmetz"], "categories": ["cs.LG", "q-bio.NC", "stat.ML"], "fields_of_study": ["Computer Science", "Biology", "Mathematics", "Medicine"], "published_date": "2025-10-20", "url": "https://arxiv.org/abs/2510.18037", "pdf_url": "https://arxiv.org/pdf/2510.18037v2", "arxiv_id": "2510.18037", "doi": "10.48550/arXiv.2510.18037", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.3415}
{"id": "34b5b32eb5f705be6a2355cdab227bdd8c3d283da147f08e57e1dbede12943b1", "sources": ["arxiv", "semantic_scholar"], "title": "Beyond Accuracy: Are Time Series Foundation Models Well-Calibrated?", "abstract": "The recent development of foundation models for time series data has generated considerable interest in using such models across a variety of applications. Although foundation models achieve state-of-the-art predictive performance, their calibration properties remain relatively underexplored, despite the fact that calibration can be critical for many practical applications. In this paper, we investigate the calibration-related properties of five recent time series foundation models and two competitive baselines. We perform a series of systematic evaluations assessing model calibration (i.e., over- or under-confidence), effects of varying prediction heads, and calibration under long-term autoregressive forecasting. We find that time series foundation models are consistently better calibrated than baseline models and tend not to be either systematically over- or under-confident, in contrast to the overconfidence often seen in other deep learning models.", "authors": ["Coen Adler", "Yuxin Chang", "Felix Draxler", "Samar Abdi", "Padhraic Smyth"], "categories": ["cs.LG", "cs.AI", "stat.ME", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2025-10-17", "url": "https://arxiv.org/abs/2510.16060", "pdf_url": "https://arxiv.org/pdf/2510.16060v2", "arxiv_id": "2510.16060", "doi": "10.48550/arXiv.2510.16060", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.338}
{"id": "54b8666124ce554d714789b5896ae66dae9f27964084eeabe16b59d1d284e189", "sources": ["arxiv", "semantic_scholar"], "title": "CoRA: Covariate-Aware Adaptation of Time Series Foundation Models", "abstract": "Time Series Foundation Models (TSFMs) have shown significant impact through their model capacity, scalability, and zero-shot generalization. However, due to the heterogeneity of inter-variate dependencies and the backbone scalability on large-scale multivariate datasets, most TSFMs are typically pre-trained on univariate time series. This limitation renders them oblivious to crucial information from diverse covariates in real-world forecasting tasks. To further enhance the performance of TSFMs, we propose a general covariate-aware adaptation (CoRA) framework for TSFMs. It leverages pre-trained backbones of foundation models while effectively incorporating exogenous covariates from various modalities, including time series, language, and images, to improve the quality of predictions. Technically, CoRA maintains the equivalence of initialization and parameter consistency during adaptation. With preserved backbones of foundation models as frozen feature extractors, the outcome embeddings from foundation models are empirically demonstrated more informative than raw data. Further, CoRA employs a novel Granger Causality Embedding (GCE) to automatically evaluate covariates regarding their causal predictability with respect to the target variate. We incorporate these weighted embeddings with a zero-initialized condition-injection mechanism, avoiding catastrophic forgetting of pre-trained foundation models and gradually integrates exogenous information. Extensive experiments show that CoRA of TSFMs surpasses state-of-the-art covariate-aware deep forecasters with full or few-shot training samples, achieving 31.1% MSE reduction on covariate-aware forecasting. Compared to other adaptation methods, CoRA exhibits strong compatibility with various advanced TSFMs and extends the scope of covariates to other modalities, presenting a practical paradigm for the application of TSFMs.", "authors": ["Guo Qin", "Zhi Chen", "Yong Liu", "Zhiyuan Shi", "Haixuan Liu", "Xiangdong Huang", "Jianmin Wang", "Mingsheng Long"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-10-14", "url": "https://arxiv.org/abs/2510.12681", "pdf_url": "https://arxiv.org/pdf/2510.12681v1", "arxiv_id": "2510.12681", "doi": "10.48550/arXiv.2510.12681", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.3346}
{"id": "02ba6795e9f1cea20f675d452c8b5e46a5812967faa6ba2606bfd3bdfa1d7f4b", "sources": ["arxiv", "semantic_scholar"], "title": "Why Do Transformers Fail to Forecast Time Series In-Context?", "abstract": "Time series forecasting (TSF) remains a challenging and largely unsolved problem in machine learning, despite significant recent efforts leveraging Large Language Models (LLMs), which predominantly rely on Transformer architectures. Empirical evidence consistently shows that even powerful Transformers often fail to outperform much simpler models, e.g., linear models, on TSF tasks; however, a rigorous theoretical understanding of this phenomenon remains limited. In this paper, we provide a theoretical analysis of Transformers' limitations for TSF through the lens of In-Context Learning (ICL) theory. Specifically, under AR($p$) data, we establish that: (1) Linear Self-Attention (LSA) models $\\textit{cannot}$ achieve lower expected MSE than classical linear models for in-context forecasting; (2) as the context length approaches to infinity, LSA asymptotically recovers the optimal linear predictor; and (3) under Chain-of-Thought (CoT) style inference, predictions collapse to the mean exponentially. We empirically validate these findings through carefully designed experiments. Our theory not only sheds light on several previously underexplored phenomena but also offers practical insights for designing more effective forecasting architectures. We hope our work encourages the broader research community to revisit the fundamental theoretical limitations of TSF and to critically evaluate the direct application of increasingly sophisticated architectures without deeper scrutiny.", "authors": ["Yufa Zhou", "Yixiao Wang", "Surbhi Goel", "Anru R. Zhang"], "categories": ["cs.LG", "cs.AI", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2025-10-10", "url": "https://arxiv.org/abs/2510.09776", "pdf_url": "https://arxiv.org/pdf/2510.09776v1", "arxiv_id": "2510.09776", "doi": "10.48550/arXiv.2510.09776", "citation_count": 6, "influential_citation_count": 1, "has_code": true, "code_url": "https://github.com/MasterZhou1/ICL-Time-Series", "venue": "arXiv.org", "quality_score": 0.51}
{"id": "da09caa62d16bed8af67efcf97799cd733a9003c96948f2b7e9e1ede882f7fdc", "sources": ["arxiv", "semantic_scholar"], "title": "Synthetic Series-Symbol Data Generation for Time Series Foundation Models", "abstract": "Foundation models for time series analysis (TSA) have attracted significant attention. However, challenges such as training data scarcity and imbalance continue to hinder their development. Inspired by complex dynamic system theories, we design a series-symbol data generation mechanism, enabling the unrestricted creation of high-quality time series data paired with corresponding symbolic expressions. To leverage series-symbol data pairs with strong correlations, we develop SymTime, a pre-trained foundation model for enhancing time series representation using symbolic information. SymTime demonstrates competitive performance across five major TSA tasks when fine-tunes with downstream tasks, rivaling foundation models pre-trained on real-world datasets. This approach underscores the potential of series-symbol data generation and pretraining mechanisms in overcoming data scarcity and enhancing task performance. The code is available at https://github.com/wwhenxuan/SymTime.", "authors": ["Wenxuan Wang", "Kai Wu", "Yujian Betterest Li", "Dan Wang", "Xiaoyu Zhang"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-10-09", "url": "https://arxiv.org/abs/2510.08445", "pdf_url": "https://arxiv.org/pdf/2510.08445v3", "arxiv_id": "2510.08445", "doi": "10.48550/arXiv.2510.08445", "citation_count": 0, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/wwhenxuan/SymTime", "venue": "arXiv.org", "quality_score": 0.5082}
{"id": "09a960cea14f1a7ab2b5705096ea71fa6ae2a2505c4878d42db5bd1b3d2b8d4a", "sources": ["arxiv", "semantic_scholar"], "title": "HTMformer: Hybrid Time and Multivariate Transformer for Time Series Forecasting", "abstract": "Transformer-based methods have achieved impressive results in time series forecasting. However, existing Transformers still exhibit limitations in sequence modeling as they tend to overemphasize temporal dependencies. This incurs additional computational overhead without yielding corresponding performance gains. We find that the performance of Transformers is highly dependent on the embedding method used to learn effective representations. To address this issue, we extract multivariate features to augment the effective information captured in the embedding layer, yielding multidimensional embeddings that convey richer and more meaningful sequence representations. These representations enable Transformer-based forecasters to better understand the series. Specifically, we introduce Hybrid Temporal and Multivariate Embeddings (HTME). The HTME extractor integrates a lightweight temporal feature extraction module with a carefully designed multivariate feature extraction module to provide complementary features, thereby achieving a balance between model complexity and performance. By combining HTME with the Transformer architecture, we present HTMformer, leveraging the enhanced feature extraction capability of the HTME extractor to build a lightweight forecaster. Experiments conducted on eight real-world datasets demonstrate that our approach outperforms existing baselines in both accuracy and efficiency.", "authors": ["Tan Wang", "Yun Wei Dong", "Qi Wang"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-10-08", "url": "https://arxiv.org/abs/2510.07084", "pdf_url": "https://arxiv.org/pdf/2510.07084v3", "arxiv_id": "2510.07084", "doi": "10.48550/arXiv.2510.07084", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.3277}
{"id": "b3622480fabb731be475b5898bc4b8ecb6d0a10cd036a3d5a10666b2941da4ce", "sources": ["arxiv", "semantic_scholar"], "title": "TimeFormer: Transformer with Attention Modulation Empowered by Temporal Characteristics for Time Series Forecasting", "abstract": "Although Transformers excel in natural language processing, their extension to time series forecasting remains challenging due to insufficient consideration of the differences between textual and temporal modalities. In this paper, we develop a novel Transformer architecture designed for time series data, aiming to maximize its representational capacity. We identify two key but often overlooked characteristics of time series: (1) unidirectional influence from the past to the future, and (2) the phenomenon of decaying influence over time. These characteristics are introduced to enhance the attention mechanism of Transformers. We propose TimeFormer, whose core innovation is a self-attention mechanism with two modulation terms (MoSA), designed to capture these temporal priors of time series under the constraints of the Hawkes process and causal masking. Additionally, TimeFormer introduces a framework based on multi-scale and subsequence analysis to capture semantic dependencies at different temporal scales, enriching the temporal dependencies. Extensive experiments conducted on multiple real-world datasets show that TimeFormer significantly outperforms state-of-the-art methods, achieving up to a 7.45% reduction in MSE compared to the best baseline and setting new benchmarks on 94.04\\% of evaluation metrics. Moreover, we demonstrate that the MoSA mechanism can be broadly applied to enhance the performance of other Transformer-based models.", "authors": ["Zhipeng Liu", "Peibo Duan", "Xuan Tang", "Baixin Li", "Yongsheng Huang", "Mingyang Geng", "Changsheng Zhang", "Bin Zhang", "Binwu Wang"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-10-08", "url": "https://arxiv.org/abs/2510.06680", "pdf_url": "https://arxiv.org/pdf/2510.06680v1", "arxiv_id": "2510.06680", "doi": "10.48550/arXiv.2510.06680", "citation_count": 6, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "Expert systems with applications", "quality_score": 0.3277}
{"id": "27ec03256d9e8b025c79991660fa4ba77c89f3f2a5dec455ebbc7fbd321f40bd", "sources": ["arxiv", "semantic_scholar"], "title": "Exploring Accuracy Law for Deep Time Series Forecasters: An Empirical Study", "abstract": "Deep time series forecasting has emerged as a rapidly growing field in recent years. Despite the exponential growth of community interests, progress on standard benchmarks is often limited to marginal improvements. A common consensus of the community is that time series forecasting inherently faces a non-zero error lower bound due to its partially observable and uncertain nature. However, a fundamental question arises: how to estimate the performance upper bound of deep time series forecasters? We delve into univariate time series forecasting, a prevalent forecasting paradigm spanning traditional statistical models to advanced time series foundation models. Going beyond classical series-wise predictability metrics, we realize that the forecasting performance is highly related to window-wise properties due to the sequence-to-sequence forecasting paradigm of deep time series models and introduce a quantitative measurement of window-wise pattern complexity. Through rigorous statistical analyses over more than 4700 newly trained deep forecasting models, we discover a consistent empirical relationship between the minimum attainable forecasting error of deep models and the complexity of window-wise series patterns, which is termed the accuracy law. We further demonstrate that this empirical finding successfully guides us to identify saturated tasks from widely used benchmarks and derive an effective training strategy for time series foundation models, offering valuable insights for future research.", "authors": ["Yuxuan Wang", "Haixu Wu", "Yuezhou Ma", "Yuchen Fang", "Ziyi Zhang", "Yong Liu", "Shiyu Wang", "Zhou Ye", "Yang Xiang", "Jianmin Wang", "Mingsheng Long"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-10-03", "url": "https://arxiv.org/abs/2510.02729", "pdf_url": "https://arxiv.org/pdf/2510.02729v2", "arxiv_id": "2510.02729", "doi": null, "citation_count": 3, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.2049}
{"id": "e88b8cf91018e1316e5f47dc3b21488b6407fa8cb0b97c1f6aa77cf1bc04931d", "sources": ["arxiv", "semantic_scholar"], "title": "RAxSS: Retrieval-Augmented Sparse Sampling for Explainable Variable-Length Medical Time Series Classification", "abstract": "Medical time series analysis is challenging due to data sparsity, noise, and highly variable recording lengths. Prior work has shown that stochastic sparse sampling effectively handles variable-length signals, while retrieval-augmented approaches improve explainability and robustness to noise and weak temporal correlations. In this study, we generalize the stochastic sparse sampling framework for retrieval-informed classification. Specifically, we weight window predictions by within-channel similarity and aggregate them in probability space, yielding convex series-level scores and an explicit evidence trail for explainability. Our method achieves competitive iEEG classification performance and provides practitioners with greater transparency and explainability. We evaluate our method in iEEG recordings collected in four medical centers, demonstrating its potential for reliable and explainable clinical variable-length time series classification.", "authors": ["Aydin Javadov", "Samir Garibov", "Tobias Hoesli", "Qiyang Sun", "Florian von Wangenheim", "Joseph Ollier", "Björn W. Schuller"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-10-03", "url": "https://arxiv.org/abs/2510.02936", "pdf_url": "https://arxiv.org/pdf/2510.02936v1", "arxiv_id": "2510.02936", "doi": "10.48550/arXiv.2510.02936", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.322}
{"id": "5ce982ca3a8e33cea6554043e172661e13a8c7505a2c8606e7fadfe1c4817465", "sources": ["arxiv", "semantic_scholar"], "title": "AI Foundation Model for Time Series with Innovations Representation", "abstract": "This paper introduces an Artificial Intelligence (AI) foundation model for time series in engineering applications, where causal operations are required for real-time monitoring and control. Since engineering time series are governed by physical, rather than linguistic, laws, large-language-model-based AI foundation models may be ineffective or inefficient. Building on the classical innovations representation theory of Wiener, Kallianpur, and Rosenblatt, we propose Time Series GPT (TS-GPT) -- an innovations-representation-based Generative Pre-trained Transformer for engineering monitoring and control. As an example of foundation model adaptation, we consider Probabilistic Generative Forecasting, which produces future time series samples from conditional probability distributions given past realizations. We demonstrate the effectiveness of TS-GPT in forecasting real-time locational marginal prices using historical data from U.S. independent system operators.", "authors": ["Lang Tong", "Xinyi Wang"], "categories": ["stat.ML", "cs.LG"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2025-10-02", "url": "https://arxiv.org/abs/2510.01560", "pdf_url": "https://arxiv.org/pdf/2510.01560v1", "arxiv_id": "2510.01560", "doi": "10.48550/arXiv.2510.01560", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.3208}
{"id": "93c4dba0ca963ff1f7edc2575ed240d450382ac0bc02721644a4bcafa5e2d358", "sources": ["arxiv", "semantic_scholar"], "title": "KAIROS: Unified Training for Universal Non-Autoregressive Time Series Forecasting", "abstract": "In the World Wide Web, reliable time series forecasts provide the forward-looking signals that drive resource planning, cache placement, and anomaly response, enabling platforms to operate efficiently as user behavior and content distributions evolve. Compared with other domains, time series forecasting for Web applications requires much faster responsiveness to support real-time decision making. We present KAIROS, a non-autoregressive time series forecasting framework that directly models segment-level multi-peak distributions. Unlike autoregressive approaches, KAIROS avoids error accumulation and achieves just-in-time inference, while improving over existing non-autoregressive models that collapse to over-smoothed predictions. Trained on the large-scale corpus, KAIROS demonstrates strong zero-shot generalization on six widely used benchmarks, delivering forecasting performance comparable to state-of-the-art foundation models with similar scale, at a fraction of their inference cost. Beyond empirical results, KAIROS highlights the importance of non-autoregressive design as a scalable paradigm for foundation models in time series.", "authors": ["Kuiye Ding", "Fanda Fan", "Zheya Wang", "Hongxiao Li", "Yifan Wang", "Lei Wang", "Chunjie Luo", "Jianfeng Zhan"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-10-02", "url": "https://arxiv.org/abs/2510.02084", "pdf_url": "https://arxiv.org/pdf/2510.02084v2", "arxiv_id": "2510.02084", "doi": "10.48550/arXiv.2510.02084", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.3208}
{"id": "92f8db258356b24811ff1ae46d539cd5a119ec8d37618dc9100d8d6d4c11a3e5", "sources": ["arxiv", "semantic_scholar"], "title": "How Foundational are Foundation Models for Time Series Forecasting?", "abstract": "Foundation Models are designed to serve as versatile embedding machines, with strong zero shot capabilities and superior generalization performance when fine-tuned on diverse downstream tasks. While this is largely true for language and vision foundation models, we argue that the inherent diversity of time series data makes them less suited for building effective foundation models. We demonstrate this using forecasting as our downstream task. We show that the zero-shot capabilities of a time series foundation model are significantly influenced and tied to the specific domains it has been pretrained on. Furthermore, when applied to unseen real-world time series data, fine-tuned foundation models do not consistently yield substantially better results, relative to their increased parameter count and memory footprint, than smaller, dedicated models tailored to the specific forecasting task at hand.", "authors": ["Nouha Karaouli", "Denis Coquenet", "Elisa Fromont", "Martial Mermillod", "Marina Reyboz"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-10-01", "url": "https://arxiv.org/abs/2510.00742", "pdf_url": "https://arxiv.org/pdf/2510.00742v3", "arxiv_id": "2510.00742", "doi": "10.48550/arXiv.2510.00742", "citation_count": 2, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.3197}
{"id": "3d5a4ad590594eb05dd9bbb193fdfb4c3416666478b4b37f9ded978dabe6ab88", "sources": ["arxiv", "semantic_scholar"], "title": "Uncovering Zero-Shot Generalization Gaps in Time-Series Foundation Models Using Real-World Videos", "abstract": "Recent research on time-series foundation models (TSFMs) has underscored the scarcity of real-world data, often supplemented with synthetic sources in existing datasets, whose generalizability remains however debated. As such, in this work, we propose a novel benchmarking approach: in particular, we aim at building a curated dataset reflecting real world physical temporal dynamics, extracting temporal signals from real-world videos using optical flow. As such, we introduce REAL-V-TSFM, a novel dataset designed to capture rich and diverse time series derived from real-world videos. Experimental results on state-of-the-art TSFMs under zero-shot forecasting show that, despite strong performance on conventional benchmarks, these models exhibit performance degradation on the proposed dataset, suggesting limited generalizability to novel datasets. These findings underscore the need for novel approaches to acquiring time series data and highlight the lack of universality in recent TSFMs, while further validating the effectiveness of our video-based time series data extraction pipeline.", "authors": ["Lujun Li", "Lama Sleem", "Yiqun Wang", "Yangjie Xu", "Niccolò Gentile", "Radu State"], "categories": ["cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-09-30", "url": "https://arxiv.org/abs/2509.26347", "pdf_url": "https://arxiv.org/pdf/2509.26347v2", "arxiv_id": "2509.26347", "doi": null, "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.2027}
{"id": "f037d9037a7f70d363b348817bd237e6776b4491d3c2c47c46154a545c6713c0", "sources": ["arxiv", "semantic_scholar"], "title": "DSAT-HD: Dual-Stream Adaptive Transformer with Hybrid Decomposition for Multivariate Time Series Forecasting", "abstract": "Time series forecasting is crucial for various applications, such as weather, traffic, electricity, and energy predictions. Currently, common time series forecasting methods are based on Transformers. However, existing approaches primarily model limited time series or fixed scales, making it more challenging to capture diverse features cross different ranges. Additionally, traditional methods like STL for complex seasonality-trend decomposition require pre-specified seasonal periods and typically handle only single, fixed seasonality. We propose the Hybrid Decomposition Dual-Stream Adaptive Transformer (DSAT-HD), which integrates three key innovations to address the limitations of existing methods: 1) A hybrid decomposition mechanism combining EMA and Fourier decomposition with RevIN normalization, dynamically balancing seasonal and trend components through noise Top-k gating; 2) A multi-scale adaptive pathway leveraging a sparse allocator to route features to four parallel Transformer layers, followed by feature merging via a sparse combiner, enhanced by hybrid attention combining local CNNs and global interactions; 3) A dual-stream residual learning framework where CNN and MLP branches separately process seasonal and trend components, coordinated by a balanced loss function minimizing expert collaboration variance. Extensive experiments on nine datasets demonstrate that DSAT-HD outperforms existing methods overall and achieves state-of-the-art performance on some datasets. Notably, it also exhibits stronger generalization capabilities across various transfer scenarios.", "authors": ["Zixu Wang", "Hongbin Dong", "Xiaoping Zhang"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-09-29", "url": "https://arxiv.org/abs/2509.24800", "pdf_url": "https://arxiv.org/pdf/2509.24800v1", "arxiv_id": "2509.24800", "doi": "10.48550/arXiv.2509.24800", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.3174}
{"id": "c9b07565dc13dd210b4236810b315051482102074ed2d2785f40b7fcde5e1862", "sources": ["arxiv", "semantic_scholar"], "title": "Numerion: A Multi-Hypercomplex Model for Time Series Forecasting", "abstract": "Many methods aim to enhance time series forecasting by decomposing the series through intricate model structures and prior knowledge, yet they are inevitably limited by computational complexity and the robustness of the assumptions. Our research uncovers that in the complex domain and higher-order hypercomplex spaces, the characteristic frequencies of time series naturally decrease. Leveraging this insight, we propose Numerion, a time series forecasting model based on multiple hypercomplex spaces. Specifically, grounded in theoretical support, we generalize linear layers and activation functions to hypercomplex spaces of arbitrary power-of-two dimensions and introduce a novel Real-Hypercomplex-Real Domain Multi-Layer Perceptron (RHR-MLP) architecture. Numerion utilizes multiple RHR-MLPs to map time series into hypercomplex spaces of varying dimensions, naturally decomposing and independently modeling the series, and adaptively fuses the latent patterns exhibited in different spaces through a dynamic fusion mechanism. Experiments validate the model`s performance, achieving state-of-the-art results on multiple public datasets. Visualizations and quantitative analyses comprehensively demonstrate the ability of multi-dimensional RHR-MLPs to naturally decompose time series and reveal the tendency of higher dimensional hypercomplex spaces to capture lower frequency features.", "authors": ["Hanzhong Cao", "Wenbo Yan", "Ying Tan"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-09-26", "url": "https://arxiv.org/abs/2510.03251", "pdf_url": "https://arxiv.org/pdf/2510.03251v1", "arxiv_id": "2510.03251", "doi": "10.48550/arXiv.2510.03251", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.314}
{"id": "b71a5f7eb82ea1637a8114758635ed6cee1a2f3ab10838f49470e162e29d37d4", "sources": ["arxiv", "semantic_scholar"], "title": "WDformer: A Wavelet-based Differential Transformer Model for Time Series Forecasting", "abstract": "Time series forecasting has various applications, such as meteorological rainfall prediction, traffic flow analysis, financial forecasting, and operational load monitoring for various systems. Due to the sparsity of time series data, relying solely on time-domain or frequency-domain modeling limits the model's ability to fully leverage multi-domain information. Moreover, when applied to time series forecasting tasks, traditional attention mechanisms tend to over-focus on irrelevant historical information, which may introduce noise into the prediction process, leading to biased results. We proposed WDformer, a wavelet-based differential Transformer model. This study employs the wavelet transform to conduct a multi-resolution analysis of time series data. By leveraging the advantages of joint representation in the time-frequency domain, it accurately extracts the key information components that reflect the essential characteristics of the data. Furthermore, we apply attention mechanisms on inverted dimensions, allowing the attention mechanism to capture relationships between multiple variables. When performing attention calculations, we introduced the differential attention mechanism, which computes the attention score by taking the difference between two separate softmax attention matrices. This approach enables the model to focus more on important information and reduce noise. WDformer has achieved state-of-the-art (SOTA) results on multiple challenging real-world datasets, demonstrating its accuracy and effectiveness. Code is available at https://github.com/xiaowangbc/WDformer.", "authors": ["Xiaojian Wang", "Chaoli Zhang", "Zhonglong Zheng", "Yunliang Jiang"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-09-25", "url": "https://arxiv.org/abs/2509.25231", "pdf_url": "https://arxiv.org/pdf/2509.25231v1", "arxiv_id": "2509.25231", "doi": "10.1145/3746252.3761348", "citation_count": 1, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/xiaowangbc/WDformer", "venue": "International Conference on Information and Knowledge Management", "quality_score": 0.4834}
{"id": "fe55feabfed0fe277396d5d5de6640572abc52d5345193852cff0e7d4cda39a5", "sources": ["arxiv", "semantic_scholar"], "title": "Why Attention Fails: The Degeneration of Transformers into MLPs in Time Series Forecasting", "abstract": "Transformer-based architectures achieved high performance in natural language processing and computer vision, yet many studies have shown that they have not demonstrated a clear advantage in time series forecasting and even underperform simple linear baselines in some cases. However, most of these studies have not thoroughly explored the reasons behind the failure of transformers. To better understand time-series transformers(TST), we designed a series of experiments, progressively modifying transformers into MLPs to investigate the impact of the attention mechanism. Surprisingly, transformer blocks often degenerate into simple MLPs in existing time-series transformers. We designed a interpretable dataset to investigate the reasons behind the failure of the attention mechanism and revealed that the attention mechanism is not working in the expected way. We theoretically analyzed the reasons behind this phenomenon, demonstrating that the current embedding methods fail to allow transformers to function in a well-structured latent space, and further analyzed the deeper underlying causes of the failure of embedding.", "authors": ["Zida Liang", "Jiayi Zhu", "Weiqiang Sun"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-09-25", "url": "https://arxiv.org/abs/2509.20942", "pdf_url": "https://arxiv.org/pdf/2509.20942v1", "arxiv_id": "2509.20942", "doi": "10.48550/arXiv.2509.20942", "citation_count": 3, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.3128}
{"id": "80da69210d538121d2e150cce1e700554b5b11f40e3c8fbc8033afe9a75ad266", "sources": ["arxiv", "semantic_scholar"], "title": "Towards Self-Supervised Foundation Models for Critical Care Time Series", "abstract": "Domain-specific foundation models for healthcare have expanded rapidly in recent years, yet foundation models for critical care time series remain relatively underexplored due to the limited size and availability of datasets. In this work, we introduce an early-stage pre-trained foundation model for critical care time-series based on the Bi-Axial Transformer (BAT), trained on pooled electronic health record datasets. We demonstrate effective transfer learning by fine-tuning the model on a dataset distinct from the training sources for mortality prediction, where it outperforms supervised baselines, particularly for small datasets ($<5,000$). These contributions highlight the potential of self-supervised foundation models for critical care times series to support generalizable and robust clinical applications in resource-limited settings.", "authors": ["Katja Naasunnguaq Jagd", "Rachael DeVries", "Ole Winther"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-09-24", "url": "https://arxiv.org/abs/2509.19885", "pdf_url": "https://arxiv.org/pdf/2509.19885v1", "arxiv_id": "2509.19885", "doi": "10.48550/arXiv.2509.19885", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.3117}
{"id": "767c9c90346a7948580392701abc0de251cfe7abb0058201c3b9e3513f2c7981", "sources": ["arxiv", "semantic_scholar"], "title": "A More Realistic Evaluation of Cross-Frequency Transfer Learning and Foundation Forecasting Models", "abstract": "Cross-frequency transfer learning (CFTL) has emerged as a popular framework for curating large-scale time series datasets to pre-train foundation forecasting models (FFMs). Although CFTL has shown promise, current benchmarking practices fall short of accurately assessing its performance. This shortcoming stems from many factors: an over-reliance on small-scale evaluation datasets; inadequate treatment of sample size when computing summary statistics; reporting of suboptimal statistical models; and failing to account for non-negligible risks of overlap between pre-training and test datasets. To address these limitations, we introduce a unified reimplementation of widely-adopted neural forecasting networks, adapting them for the CFTL setup; we pre-train only on proprietary and synthetic data, being careful to prevent test leakage; and we evaluate on 15 large, diverse public forecast competition datasets. Our empirical analysis reveals that statistical models' accuracy is frequently underreported. Notably, we confirm that statistical models and their ensembles consistently outperform existing FFMs by more than 8.2% in sCRPS, and by more than 20% MASE, across datasets. However, we also find that synthetic dataset pre-training does improve the accuracy of a FFM by 7% percent.", "authors": ["Kin G. Olivares", "Malcolm Wolff", "Tatiana Konstantinova", "Shankar Ramasubramanian", "Boris Oreshkin", "Andrew Gordon Wilson", "Andres Potapczynski", "Willa Potosnak", "Michael W. Mahoney", "Mengfei Cao", "Dmitry Efimov"], "categories": ["cs.LG", "cs.AI", "stat.AP"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2025-09-23", "url": "https://arxiv.org/abs/2509.19465", "pdf_url": "https://arxiv.org/pdf/2509.19465v3", "arxiv_id": "2509.19465", "doi": null, "citation_count": 2, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.1976}
{"id": "aaf3778852fa4801b9c8e42a1f220a2fd0d8b8fd9a4b16198be9e1571536e312", "sources": ["arxiv", "semantic_scholar"], "title": "Time Series Forecasting Using a Hybrid Deep Learning Method: A Bi-LSTM Embedding Denoising Auto Encoder Transformer", "abstract": "Time series data is a prevalent form of data found in various fields. It consists of a series of measurements taken over time. Forecasting is a crucial application of time series models, where future values are predicted based on historical data. Accurate forecasting is essential for making well-informed decisions across industries. When it comes to electric vehicles (EVs), precise predictions play a key role in planning infrastructure development, load balancing, and energy management. This study introduces a BI-LSTM embedding denoising autoencoder model (BDM) designed to address time series problems, focusing on short-term EV charging load prediction. The performance of the proposed model is evaluated by comparing it with benchmark models like Transformer, CNN, RNN, LSTM, and GRU. Based on the results of the study, the proposed model outperforms the benchmark models in four of the five-time steps, demonstrating its effectiveness for time series forecasting. This research makes a significant contribution to enhancing time series forecasting, thereby improving decision-making processes.", "authors": ["Sahar Koohfar", "Wubeshet Woldemariam"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-09-21", "url": "https://arxiv.org/abs/2509.17165", "pdf_url": "https://arxiv.org/pdf/2509.17165v1", "arxiv_id": "2509.17165", "doi": "10.48550/arXiv.2509.17165", "citation_count": 2, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.3082}
{"id": "a4ab77d95d0aa8f2c5e98296d58a4b696ff8dd1ad7c3c1dafd7539626f2f979c", "sources": ["arxiv", "semantic_scholar"], "title": "TFMAdapter: Lightweight Instance-Level Adaptation of Foundation Models for Forecasting with Covariates", "abstract": "Time Series Foundation Models (TSFMs) have recently achieved state-of-the-art performance in univariate forecasting on new time series simply by conditioned on a brief history of past values. Their success demonstrates that large-scale pretraining across diverse domains can acquire the inductive bias to generalize from temporal patterns in a brief history. However, most TSFMs are unable to leverage covariates -- future-available exogenous variables critical for accurate forecasting in many applications -- due to their domain-specific nature and the lack of associated inductive bias. We propose TFMAdapter, a lightweight, instance-level adapter that augments TSFMs with covariate information without fine-tuning. Instead of retraining, TFMAdapter operates on the limited history provided during a single model call, learning a non-parametric cascade that combines covariates with univariate TSFM forecasts. However, such learning would require univariate forecasts at all steps in the history, requiring too many calls to the TSFM. To enable training on the full historical context while limiting TSFM invocations, TFMAdapter uses a two-stage method: (1) generating pseudo-forecasts with a simple regression model, and (2) training a Gaussian Process regressor to refine predictions using both pseudo- and TSFM forecasts alongside covariates. Extensive experiments on real-world datasets demonstrate that TFMAdapter consistently outperforms both foundation models and supervised baselines, achieving a 24-27\\% improvement over base foundation models with minimal data and computational overhead. Our results highlight the potential of lightweight adapters to bridge the gap between generic foundation models and domain-specific forecasting needs.", "authors": ["Afrin Dange", "Sunita Sarawagi"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-09-17", "url": "https://arxiv.org/abs/2509.13906", "pdf_url": "https://arxiv.org/pdf/2509.13906v1", "arxiv_id": "2509.13906", "doi": "10.1145/3746252.3761272", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "International Conference on Information and Knowledge Management", "quality_score": 0.3036}
{"id": "90736b77f40b3ea25382bb465fe79186bbe341855a03d66ecaa67c6bdc871b0f", "sources": ["arxiv", "semantic_scholar"], "title": "Prediction and Causality of functional MRI and synthetic signal using a Zero-Shot Time-Series Foundation Model", "abstract": "Time-series forecasting and causal discovery are central in neuroscience, as predicting brain activity and identifying causal relationships between neural populations and circuits can shed light on the mechanisms underlying cognition and disease. With the rise of foundation models, an open question is how they compare to traditional methods for brain signal forecasting and causality analysis, and whether they can be applied in a zero-shot setting. In this work, we evaluate a foundation model against classical methods for inferring directional interactions from spontaneous brain activity measured with functional magnetic resonance imaging (fMRI) in humans. Traditional approaches often rely on Wiener-Granger causality. We tested the forecasting ability of the foundation model in both zero-shot and fine-tuned settings, and assessed causality by comparing Granger-like estimates from the model with standard Granger causality. We validated the approach using synthetic time series generated from ground-truth causal models, including logistic map coupling and Ornstein-Uhlenbeck processes. The foundation model achieved competitive zero-shot forecasting fMRI time series (mean absolute percentage error of 0.55 in controls and 0.27 in patients). Although standard Granger causality did not show clear quantitative differences between models, the foundation model provided a more precise detection of causal interactions. Overall, these findings suggest that foundation models offer versatility, strong zero-shot performance, and potential utility for forecasting and causal discovery in time-series data.", "authors": ["Alessandro Crimi", "Andrea Brovelli"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-09-15", "url": "https://arxiv.org/abs/2509.12497", "pdf_url": "https://arxiv.org/pdf/2509.12497v2", "arxiv_id": "2509.12497", "doi": "10.48550/arXiv.2509.12497", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.3014}
{"id": "4bb25c635f1e685d7791ab02db8b63bed5560e7394bf66d084b84c754b11d9d4", "sources": ["arxiv", "semantic_scholar"], "title": "FinZero: Launching Multi-modal Financial Time Series Forecast with Large Reasoning Model", "abstract": "Financial time series forecasting is both highly significant and challenging. Previous approaches typically standardized time series data before feeding it into forecasting models, but this encoding process inherently leads to a loss of important information. Moreover, past time series models generally require fixed numbers of variables or lookback window lengths, which further limits the scalability of time series forecasting. Besides, the interpretability and the uncertainty in forecasting remain areas requiring further research, as these factors directly impact the reliability and practical value of predictions. To address these issues, we first construct a diverse financial image-text dataset (FVLDB) and develop the Uncertainty-adjusted Group Relative Policy Optimization (UARPO) method to enable the model not only output predictions but also analyze the uncertainty of those predictions. We then proposed FinZero, a multimodal pre-trained model finetuned by UARPO to perform reasoning, prediction, and analytical understanding on the FVLDB financial time series. Extensive experiments validate that FinZero exhibits strong adaptability and scalability. After fine-tuning with UARPO, FinZero achieves an approximate 13.48\\% improvement in prediction accuracy over GPT-4o in the high-confidence group, demonstrating the effectiveness of reinforcement learning fine-tuning in multimodal large model, including in financial time series forecasting tasks.", "authors": ["Yanlong Wang", "Jian Xu", "Fei Ma", "Hongkang Zhang", "Hang Yu", "Tiantian Gao", "Yu Wang", "Haochen You", "Shao-Lun Huang", "Danny Dongning Sun", "Xiao-Ping Zhang"], "categories": ["q-fin.CP", "cs.AI"], "fields_of_study": ["Computer Science", "Economics"], "published_date": "2025-09-10", "url": "https://arxiv.org/abs/2509.08742", "pdf_url": "https://arxiv.org/pdf/2509.08742v1", "arxiv_id": "2509.08742", "doi": "10.48550/arXiv.2509.08742", "citation_count": 4, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.2956}
{"id": "7e5bb10b50da46a59e4667ce096d38092886d59e824bb6b558044d614ebc3991", "sources": ["arxiv", "semantic_scholar"], "title": "ARIES: Relation Assessment and Model Recommendation for Deep Time Series Forecasting", "abstract": "Recent advancements in deep learning models for time series forecasting have been significant. These models often leverage fundamental time series properties such as seasonality and non-stationarity, which may suggest an intrinsic link between model performance and data properties. However, existing benchmark datasets fail to offer diverse and well-defined temporal patterns, restricting the systematic evaluation of such connections. Additionally, there is no effective model recommendation approach, leading to high time and cost expenditures when testing different architectures across different downstream applications. For those reasons, we propose ARIES, a framework for assessing relation between time series properties and modeling strategies, and for recommending deep forcasting models for realistic time series. First, we construct a synthetic dataset with multiple distinct patterns, and design a comprehensive system to compute the properties of time series. Next, we conduct an extensive benchmarking of over 50 forecasting models, and establish the relationship between time series properties and modeling strategies. Our experimental results reveal a clear correlation. Based on these findings, we propose the first deep forecasting model recommender, capable of providing interpretable suggestions for real-world time series. In summary, ARIES is the first study to establish the relations between the properties of time series data and modeling strategies, while also implementing a model recommendation system. The code is available at: https://github.com/blisky-li/ARIES.", "authors": ["Fei Wang", "Yujie Li", "Zezhi Shao", "Chengqing Yu", "Yisong Fu", "Zhulin An", "Yongjun Xu", "Xueqi Cheng"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-09-07", "url": "https://arxiv.org/abs/2509.06060", "pdf_url": "https://arxiv.org/pdf/2509.06060v1", "arxiv_id": "2509.06060", "doi": "10.48550/arXiv.2509.06060", "citation_count": 0, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/blisky-li/ARIES", "venue": "arXiv.org", "quality_score": 0.4516}
{"id": "da971cd457c6f5f0b555968f7972d4652def25cb2c6a2074fdbb8d63c75827fb", "sources": ["arxiv", "semantic_scholar"], "title": "time2time: Causal Intervention in Hidden States to Simulate Rare Events in Time Series Foundation Models", "abstract": "While transformer-based foundation models excel at forecasting routine patterns, two questions remain: do they internalize semantic concepts such as market regimes, or merely fit curves? And can their internal representations be leveraged to simulate rare, high-stakes events such as market crashes? To investigate this, we introduce activation transplantation, a causal intervention that manipulates hidden states by imposing the statistical moments of one event (e.g., a historical crash) onto another (e.g., a calm period) during the forward pass. This procedure deterministically steers forecasts: injecting crash semantics induces downturn predictions, while injecting calm semantics suppresses crashes and restores stability. Beyond binary control, we find that models encode a graded notion of event severity, with the latent vector norm directly correlating with the magnitude of systemic shocks. Validated across two architecturally distinct TSFMs, Toto (decoder only) and Chronos (encoder-decoder), our results demonstrate that steerable, semantically grounded representations are a robust property of large time series transformers. Our findings provide evidence for a latent concept space that governs model predictions, shifting interpretability from post-hoc attribution to direct causal intervention, and enabling semantic \"what-if\" analysis for strategic stress-testing.", "authors": ["Debdeep Sanyal", "Aaryan Nagpal", "Dhruv Kumar", "Murari Mandal", "Saurabh Deshpande"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-09-06", "url": "https://arxiv.org/abs/2509.05801", "pdf_url": "https://arxiv.org/pdf/2509.05801v2", "arxiv_id": "2509.05801", "doi": "10.48550/arXiv.2509.05801", "citation_count": 2, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.291}
{"id": "22c106237b62171fab2b4f75ed6c40ad5229c6dc166fd391a865e4f5cebd59ba", "sources": ["arxiv", "semantic_scholar"], "title": "VARMA-Enhanced Transformer for Time Series Forecasting", "abstract": "Transformer-based models have significantly advanced time series forecasting. Recent work, like the Cross-Attention-only Time Series transformer (CATS), shows that removing self-attention can make the model more accurate and efficient. However, these streamlined architectures may overlook the fine-grained, local temporal dependencies effectively captured by classical statistical models like Vector AutoRegressive Moving Average model (VARMA). To address this gap, we propose VARMAformer, a novel architecture that synergizes the efficiency of a cross-attention-only framework with the principles of classical time series analysis. Our model introduces two key innovations: (1) a dedicated VARMA-inspired Feature Extractor (VFE) that explicitly models autoregressive (AR) and moving-average (MA) patterns at the patch level, and (2) a VARMA-Enhanced Attention (VE-atten) mechanism that employs a temporal gate to make queries more context-aware. By fusing these classical insights into a modern backbone, VARMAformer captures both global, long-range dependencies and local, statistical structures. Through extensive experiments on widely-used benchmark datasets, we demonstrate that our model consistently outperforms existing state-of-the-art methods. Our work validates the significant benefit of integrating classical statistical insights into modern deep learning frameworks for time series forecasting.", "authors": ["Jiajun Song", "Xiaoou Liu"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-09-05", "url": "https://arxiv.org/abs/2509.04782", "pdf_url": "https://arxiv.org/pdf/2509.04782v1", "arxiv_id": "2509.04782", "doi": "10.48550/arXiv.2509.04782", "citation_count": 1, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "Pacific Rim International Conference on Artificial Intelligence", "quality_score": 0.2899}
{"id": "1116db55cf6ee03a47c19b2d02aff3c58d0f129e6e20fa9993319f94074c925a", "sources": ["arxiv", "semantic_scholar"], "title": "ChronoGraph: A Real-World Graph-Based Multivariate Time Series Dataset", "abstract": "We present ChronoGraph, a graph-structured multivariate time series forecasting dataset built from real-world production microservices. Each node is a service that emits a multivariate stream of system-level performance metrics, capturing CPU, memory, and network usage patterns, while directed edges encode dependencies between services. The primary task is forecasting future values of these signals at the service level. In addition, ChronoGraph provides expert-annotated incident windows as anomaly labels, enabling evaluation of anomaly detection methods and assessment of forecast robustness during operational disruptions. Compared to existing benchmarks from industrial control systems or traffic and air-quality domains, ChronoGraph uniquely combines (i) multivariate time series, (ii) an explicit, machine-readable dependency graph, and (iii) anomaly labels aligned with real incidents. We report baseline results spanning forecasting models, pretrained time-series foundation models, and standard anomaly detectors. ChronoGraph offers a realistic benchmark for studying structure-aware forecasting and incident-aware evaluation in microservice systems.", "authors": ["Adrian Catalin Lutu", "Ioana Pintilie", "Elena Burceanu", "Andrei Manolache"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-09-04", "url": "https://arxiv.org/abs/2509.04449", "pdf_url": "https://arxiv.org/pdf/2509.04449v3", "arxiv_id": "2509.04449", "doi": "10.48550/arXiv.2509.04449", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.2888}
{"id": "4eb744ab06e7ba402231dc571f646ffaed4f092697c59cd8ad8b2f9767ac0450", "sources": ["arxiv", "semantic_scholar"], "title": "On Identifying Why and When Foundation Models Perform Well on Time-Series Forecasting Using Automated Explanations and Rating", "abstract": "Time-series forecasting models (TSFM) have evolved from classical statistical methods to sophisticated foundation models, yet understanding why and when these models succeed or fail remains challenging. Despite this known limitation, time series forecasting models are increasingly used to generate information that informs real-world actions with equally real consequences. Understanding the complexity, performance variability, and opaque nature of these models then becomes a valuable endeavor to combat serious concerns about how users should interact with and rely on these models' outputs. This work addresses these concerns by combining traditional explainable AI (XAI) methods with Rating Driven Explanations (RDE) to assess TSFM performance and interpretability across diverse domains and use cases. We evaluate four distinct model architectures: ARIMA, Gradient Boosting, Chronos (time-series specific foundation model), Llama (general-purpose; both fine-tuned and base models) on four heterogeneous datasets spanning finance, energy, transportation, and automotive sales domains. In doing so, we demonstrate that feature-engineered models (e.g., Gradient Boosting) consistently outperform foundation models (e.g., Chronos) in volatile or sparse domains (e.g., power, car parts) while providing more interpretable explanations, whereas foundation models excel only in stable or trend-driven contexts (e.g., finance).", "authors": ["Michael Widener", "Kausik Lakkaraju", "John Aydin", "Biplav Srivastava"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-08-28", "url": "https://arxiv.org/abs/2508.20437", "pdf_url": "https://arxiv.org/pdf/2508.20437v1", "arxiv_id": "2508.20437", "doi": "10.48550/arXiv.2508.20437", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.1786}
{"id": "3f645c256e1dd430d7d60cf6a0db7c55ac5ebd399e324f6f7b2c0c076ce717fb", "sources": ["arxiv", "semantic_scholar"], "title": "FinCast: A Foundation Model for Financial Time-Series Forecasting", "abstract": "Financial time-series forecasting is critical for maintaining economic stability, guiding informed policymaking, and promoting sustainable investment practices. However, it remains challenging due to various underlying pattern shifts. These shifts arise primarily from three sources: temporal non-stationarity (distribution changes over time), multi-domain diversity (distinct patterns across financial domains such as stocks, commodities, and futures), and varying temporal resolutions (patterns differing across per-second, hourly, daily, or weekly indicators). While recent deep learning methods attempt to address these complexities, they frequently suffer from overfitting and typically require extensive domain-specific fine-tuning. To overcome these limitations, we introduce FinCast, the first foundation model specifically designed for financial time-series forecasting, trained on large-scale financial datasets. Remarkably, FinCast exhibits robust zero-shot performance, effectively capturing diverse patterns without domain-specific fine-tuning. Comprehensive empirical and qualitative evaluations demonstrate that FinCast surpasses existing state-of-the-art methods, highlighting its strong generalization capabilities.", "authors": ["Zhuohang Zhu", "Haodong Chen", "Qiang Qu", "Vera Chung"], "categories": ["cs.LG", "cs.AI", "q-fin.CP"], "fields_of_study": ["Computer Science", "Economics"], "published_date": "2025-08-27", "url": "https://arxiv.org/abs/2508.19609", "pdf_url": "https://arxiv.org/pdf/2508.19609v1", "arxiv_id": "2508.19609", "doi": "10.1145/3746252.3761261", "citation_count": 12, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "International Conference on Information and Knowledge Management", "quality_score": 0.2796}
{"id": "a640d8666b25c77939bd972b83c6b759ac3db07c2421dc38e723ce2772ce4371", "sources": ["arxiv", "semantic_scholar"], "title": "PAX-TS: Model-agnostic multi-granular explanations for time series forecasting via localized perturbations", "abstract": "Time series forecasting has seen considerable improvement during the last years, with transformer models and large language models driving advancements of the state of the art. Modern forecasting models are generally opaque and do not provide explanations for their forecasts, while well-known post-hoc explainability methods like LIME are not suitable for the forecasting context. We propose PAX-TS, a model-agnostic post-hoc algorithm to explain time series forecasting models and their forecasts. Our method is based on localized input perturbations and results in multi-granular explanations. Further, it is able to characterize cross-channel correlations for multivariate time series forecasts. We clearly outline the algorithmic procedure behind PAX-TS, demonstrate it on a benchmark with 7 algorithms and 10 diverse datasets, compare it with two other state-of-the-art explanation algorithms, and present the different explanation types of the method. We found that the explanations of high-performing and low-performing algorithms differ on the same datasets, highlighting that the explanations of PAX-TS effectively capture a model's behavior. Based on time step correlation matrices resulting from the benchmark, we identify 6 classes of patterns that repeatedly occur across different datasets and algorithms. We found that the patterns are indicators of performance, with noticeable differences in forecasting error between the classes. Lastly, we outline a multivariate example where PAX-TS demonstrates how the forecasting model takes cross-channel correlations into account. With PAX-TS, time series forecasting models' mechanisms can be illustrated in different levels of detail, and its explanations can be used to answer practical questions on forecasts.", "authors": ["Tim Kreuzer", "Jelena Zdravkovic", "Panagiotis Papapetrou"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-08-26", "url": "https://arxiv.org/abs/2508.18982", "pdf_url": "https://arxiv.org/pdf/2508.18982v2", "arxiv_id": "2508.18982", "doi": "10.48550/arXiv.2508.18982", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.2784}
{"id": "5c5765d32adda37c25e8b85705bdbbea7b58ae5e1370e2695e487c72b945a5c2", "sources": ["arxiv", "semantic_scholar"], "title": "Cross-device Zero-shot Label Transfer via Alignment of Time Series Foundation Model Embeddings", "abstract": "High-quality, medically validated labels exist for clinical actigraphy data but not for ubiquitous consumer wearables like the Apple Watch. Manually labeling wearables data is expensive and doesn't scale. This paper offers a novel framework that transfers valuable labels from a source domain (e.g., actigraphy) to a target domain (e.g., Apple Watch) without requiring paired data. Instead of working with raw time-series signals, we project both domains into a shared latent embedding space using time-series foundation models (TSFMs) and develop a new framework to align the cross-device representations. Our method, Adversarial Alignment of TSFM Embeddings forces the distributions of source and target embeddings to align within this space, facilitating label transfer across device type.", "authors": ["Neal G. Ravindra", "Arijit Sehanobish"], "categories": ["eess.SP", "cs.AI", "cs.LG"], "fields_of_study": ["Engineering", "Computer Science"], "published_date": "2025-08-22", "url": "https://arxiv.org/abs/2509.06966", "pdf_url": "https://arxiv.org/pdf/2509.06966v1", "arxiv_id": "2509.06966", "doi": "10.48550/arXiv.2509.06966", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.2739}
{"id": "d53c210ba008a2310463fbea6872d44e2cbfc3e08e99048317400278012133ee", "sources": ["arxiv", "semantic_scholar"], "title": "PENGUIN: Enhancing Transformer with Periodic-Nested Group Attention for Long-term Time Series Forecasting", "abstract": "Despite advances in the Transformer architecture, their effectiveness for long-term time series forecasting (LTSF) remains controversial. In this paper, we investigate the potential of integrating explicit periodicity modeling into the self-attention mechanism to enhance the performance of Transformer-based architectures for LTSF. Specifically, we propose PENGUIN, a simple yet effective periodic-nested group attention mechanism. Our approach introduces a periodic-aware relative attention bias to directly capture periodic structures and a grouped multi-query attention mechanism to handle multiple coexisting periodicities (e.g., daily and weekly cycles) within time series data. Extensive experiments across diverse benchmarks demonstrate that PENGUIN consistently outperforms both MLP-based and Transformer-based models. Code is available at https://github.com/ysygMhdxw/AISTATS2026_PENGUIN.", "authors": ["Tian Sun", "Yuqi Chen", "Weiwei Sun"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-08-19", "url": "https://arxiv.org/abs/2508.13773", "pdf_url": "https://arxiv.org/pdf/2508.13773v3", "arxiv_id": "2508.13773", "doi": "10.48550/arXiv.2508.13773", "citation_count": 3, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/ysygMhdxw/AISTATS2026_PENGUIN", "venue": "arXiv.org", "quality_score": 0.4179}
{"id": "0a38c18b03533447828d2a62dcf2848415016e2c17b295d8b89e4e62540aea30", "sources": ["arxiv", "semantic_scholar"], "title": "Enhancing Transformer-Based Foundation Models for Time Series Forecasting via Bagging, Boosting and Statistical Ensembles", "abstract": "Time series foundation models (TSFMs) such as Lag-Llama, TimeGPT, Chronos, MOMENT, UniTS, and TimesFM have shown strong generalization and zero-shot capabilities for time series forecasting, anomaly detection, classification, and imputation. Despite these advantages, their predictions still suffer from variance, domain-specific bias, and limited uncertainty quantification when deployed on real operational data. This paper investigates a suite of statistical and ensemble-based enhancement techniques, including bootstrap-based bagging, regression-based stacking, prediction interval construction, statistical residual modeling, and iterative error feedback, to improve robustness and accuracy. Using the Belgium Electricity Short-Term Load Forecasting dataset as a case study, we demonstrate that the proposed hybrids consistently outperform standalone foundation models across multiple horizons. Regression-based ensembles achieve the lowest mean squared error; bootstrap aggregation markedly reduces long-context errors; residual modeling corrects systematic bias; and the resulting prediction intervals achieve near nominal coverage with widths shrinking as context length increases. The results indicate that integrating statistical reasoning with modern foundation models yields measurable gains in accuracy, reliability, and interpretability for real-world time series applications.", "authors": ["Dhruv D. Modi", "Rong Pan"], "categories": ["cs.LG", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2025-08-18", "url": "https://arxiv.org/abs/2508.16641", "pdf_url": "https://arxiv.org/pdf/2508.16641v1", "arxiv_id": "2508.16641", "doi": "10.48550/arXiv.2508.16641", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.2693}
{"id": "2bb19850d8571322feb48723b3338bdb63fcb550ec67c2c09b831a24f9aaaee5", "sources": ["arxiv", "semantic_scholar"], "title": "CC-Time: Cross-Model and Cross-Modality Time Series Forecasting", "abstract": "With the success of pre-trained language models (PLMs) in various application fields beyond natural language processing, language models have raised emerging attention in the field of time series forecasting (TSF) and have shown great prospects. However, current PLM-based TSF methods still fail to achieve satisfactory prediction accuracy matching the strong sequential modeling power of language models. To address this issue, we propose Cross-Model and Cross-Modality Learning with PLMs for time series forecasting (CC-Time). We explore the potential of PLMs for time series forecasting from two aspects: 1) what time series features could be modeled by PLMs, and 2) whether relying solely on PLMs is sufficient for building time series models. In the first aspect, CC-Time incorporates cross-modality learning to model temporal dependency and channel correlations in the language model from both time series sequences and their corresponding text descriptions. In the second aspect, CC-Time further proposes the cross-model fusion block to adaptively integrate knowledge from the PLMs and time series model to form a more comprehensive modeling of time series patterns. Extensive experiments on nine real-world datasets demonstrate that CC-Time achieves state-of-the-art prediction accuracy in both full-data training and few-shot learning situations.", "authors": ["Peng Chen", "Yihang Wang", "Yang Shu", "Yunyao Cheng", "Kai Zhao", "Zhongwen Rao", "Lujia Pan", "Bin Yang", "Chenjuan Guo"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-08-17", "url": "https://arxiv.org/abs/2508.12235", "pdf_url": "https://arxiv.org/pdf/2508.12235v3", "arxiv_id": "2508.12235", "doi": "10.48550/arXiv.2508.12235", "citation_count": 7, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.2681}
{"id": "73aba175d859d9bc303e5750dc25e63ffc6866d1679bc44c1aec8720adbb6b38", "sources": ["arxiv", "semantic_scholar"], "title": "Measuring Time Series Forecast Stability for Demand Planning", "abstract": "Time series forecasting is a critical first step in generating demand plans for supply chains. Experiments on time series models typically focus on demonstrating improvements in forecast accuracy over existing/baseline solutions, quantified according to some accuracy metric. There is no doubt that forecast accuracy is important; however in production systems, demand planners often value consistency and stability over incremental accuracy improvements. Assuming that the inputs have not changed significantly, forecasts that vary drastically from one planning cycle to the next require high amounts of human intervention, which frustrates demand planners and can even cause them to lose trust in ML forecasting models. We study model-induced stochasticity, which quantifies the variance of a set of forecasts produced by a single model when the set of inputs is fixed. Models with lower variance are more stable. Recently the forecasting community has seen significant advances in forecast accuracy through the development of deep machine learning models for time series forecasting. We perform a case study measuring the stability and accuracy of state-of-the-art forecasting models (Chronos, DeepAR, PatchTST, Temporal Fusion Transformer, TiDE, and the AutoGluon best quality ensemble) on public data sets from the M5 competition and Favorita grocery sales. We show that ensemble models improve stability without significantly deteriorating (or even improving) forecast accuracy. While these results may not be surprising, the main point of this paper is to propose the need for further study of forecast stability for models that are being deployed in production systems.", "authors": ["Steven Klee", "Yuntian Xia"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-08-13", "url": "https://arxiv.org/abs/2508.10063", "pdf_url": "https://arxiv.org/pdf/2508.10063v1", "arxiv_id": "2508.10063", "doi": "10.48550/arXiv.2508.10063", "citation_count": 2, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.2635}
{"id": "6c33a8e9ca12c83275fd796a28b830682e4ecf1e0de99f8161c9f492fcc70e26", "sources": ["arxiv", "semantic_scholar"], "title": "Wavelet Mixture of Experts for Time Series Forecasting", "abstract": "The field of time series forecasting is rapidly advancing, with recent large-scale Transformers and lightweight Multilayer Perceptron (MLP) models showing strong predictive performance. However, conventional Transformer models are often hindered by their large number of parameters and their limited ability to capture non-stationary features in data through smoothing. Similarly, MLP models struggle to manage multi-channel dependencies effectively. To address these limitations, we propose a novel, lightweight time series prediction model, WaveTS-B. This model combines wavelet transforms with MLP to capture both periodic and non-stationary characteristics of data in the wavelet domain. Building on this foundation, we propose a channel clustering strategy that incorporates a Mixture of Experts (MoE) framework, utilizing a gating mechanism and expert network to handle multi-channel dependencies efficiently. We propose WaveTS-M, an advanced model tailored for multi-channel time series prediction. Empirical evaluation across eight real-world time series datasets demonstrates that our WaveTS series models achieve state-of-the-art (SOTA) performance with significantly fewer parameters. Notably, WaveTS-M shows substantial improvements on multi-channel datasets, highlighting its effectiveness.", "authors": ["Zheng Zhou", "Yu-Jie Xiong", "Jia-Chen Zhang", "Chun-Ming Xia", "Xi-Jiong Xie"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-08-12", "url": "https://arxiv.org/abs/2508.08825", "pdf_url": "https://arxiv.org/pdf/2508.08825v1", "arxiv_id": "2508.08825", "doi": "10.48550/arXiv.2508.08825", "citation_count": 3, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.2624}
{"id": "8e21eca0320cef4ea34edba0d35e0b5abcfc556cee38b038aed3bcc091e84b7d", "sources": ["arxiv", "semantic_scholar"], "title": "Comparative Analysis of Time Series Foundation Models for Demographic Forecasting: Enhancing Predictive Accuracy in US Population Dynamics", "abstract": "Demographic shifts, influenced by globalization, economic conditions, geopolitical events, and environmental factors, pose significant challenges for policymakers and researchers. Accurate demographic forecasting is essential for informed decision-making in areas such as urban planning, healthcare, and economic policy. This study explores the application of time series foundation models to predict demographic changes in the United States using datasets from the U.S. Census Bureau and Federal Reserve Economic Data (FRED). We evaluate the performance of the Time Series Foundation Model (TimesFM) against traditional baselines including Long Short-Term Memory (LSTM) networks, Autoregressive Integrated Moving Average (ARIMA), and Linear Regression. Our experiments across six demographically diverse states demonstrate that TimesFM achieves the lowest Mean Squared Error (MSE) in 86.67% of test cases, with particularly strong performance on minority populations with sparse historical data. These findings highlight the potential of pre-trained foundation models to enhance demographic analysis and inform proactive policy interventions without requiring extensive task-specific fine-tuning.", "authors": ["Aditya Akella", "Jonathan Farah"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-08-09", "url": "https://arxiv.org/abs/2508.11680", "pdf_url": "https://arxiv.org/pdf/2508.11680v2", "arxiv_id": "2508.11680", "doi": "10.48550/arXiv.2508.11680", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.259}
{"id": "a17b60710c2d93f7e3119fdfd89dcd504bcbea273f22181039f806203368930d", "sources": ["arxiv", "semantic_scholar"], "title": "EnergyPatchTST: Multi-scale Time Series Transformers with Uncertainty Estimation for Energy Forecasting", "abstract": "Accurate and reliable energy time series prediction is of great significance for power generation planning and allocation. At present, deep learning time series prediction has become the mainstream method. However, the multi-scale time dynamics and the irregularity of real data lead to the limitations of the existing methods. Therefore, we propose EnergyPatchTST, which is an extension of the Patch Time Series Transformer specially designed for energy forecasting. The main innovations of our method are as follows: (1) multi-scale feature extraction mechanism to capture patterns with different time resolutions; (2) probability prediction framework to estimate uncertainty through Monte Carlo elimination; (3) integration path of future known variables (such as temperature and wind conditions); And (4) Pre-training and Fine-tuning examples to enhance the performance of limited energy data sets. A series of experiments on common energy data sets show that EnergyPatchTST is superior to other commonly used methods, the prediction error is reduced by 7-12%, and reliable uncertainty estimation is provided, which provides an important reference for time series prediction in the energy field.", "authors": ["Wei Li", "Zixin Wang", "Qizheng Sun", "Qixiang Gao", "Fenglei Yang"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-08-07", "url": "https://arxiv.org/abs/2508.05454", "pdf_url": "https://arxiv.org/pdf/2508.05454v1", "arxiv_id": "2508.05454", "doi": "10.1007/978-981-96-9815-8_27", "citation_count": 4, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "International Conference on Intelligent Computing", "quality_score": 0.2567}
{"id": "e9276069ce77b3f8316bd671a773f29d218150c59d6e46e7ca92d61106424742", "sources": ["arxiv", "semantic_scholar"], "title": "Empowering Time Series Forecasting with LLM-Agents", "abstract": "Large Language Model (LLM) powered agents have emerged as effective planners for Automated Machine Learning (AutoML) systems. While most existing AutoML approaches focus on automating feature engineering and model architecture search, recent studies in time series forecasting suggest that lightweight models can often achieve state-of-the-art performance. This observation led us to explore improving data quality, rather than model architecture, as a potentially fruitful direction for AutoML on time series data. We propose DCATS, a Data-Centric Agent for Time Series. DCATS leverages metadata accompanying time series to clean data while optimizing forecasting performance. We evaluated DCATS using four time series forecasting models on a large-scale traffic volume forecasting dataset. Results demonstrate that DCATS achieves an average 6% error reduction across all tested models and time horizons, highlighting the potential of data-centric approaches in AutoML for time series forecasting.", "authors": ["Chin-Chia Michael Yeh", "Vivian Lai", "Uday Singh Saini", "Xiran Fan", "Yujie Fan", "Junpeng Wang", "Xin Dai", "Yan Zheng"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-08-06", "url": "https://arxiv.org/abs/2508.04231", "pdf_url": "https://arxiv.org/pdf/2508.04231v2", "arxiv_id": "2508.04231", "doi": "10.1109/BigData66926.2025.11402044", "citation_count": 10, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "BigData Congress [Services Society]", "quality_score": 0.2603}
{"id": "09c4a35de5d70dc3b4e103a8a888f0bb2006e7cc545d7413cf5db615ae8baf33", "sources": ["arxiv", "semantic_scholar"], "title": "DeepKoopFormer: A Koopman Enhanced Transformer Based Architecture for Time Series Forecasting", "abstract": "Time series forecasting plays a vital role across scientific, industrial, and environmental domains, especially when dealing with high-dimensional and nonlinear systems. While Transformer-based models have recently achieved state-of-the-art performance in long-range forecasting, they often suffer from interpretability issues and instability in the presence of noise or dynamical uncertainty. In this work, we propose DeepKoopFormer, a principled forecasting framework that combines the representational power of Transformers with the theoretical rigor of Koopman operator theory. Our model features a modular encoder-propagator-decoder structure, where temporal dynamics are learned via a spectrally constrained, linear Koopman operator in a latent space. We impose structural guarantees-such as bounded spectral radius, Lyapunov based energy regularization, and orthogonal parameterization to ensure stability and interpretability. Comprehensive evaluations are conducted on both synthetic dynamical systems, real-world climate dataset (wind speed and surface pressure), financial time series (cryptocurrency), and electricity generation dataset using the Python package that is prepared for this purpose. Across all experiments, DeepKoopFormer consistently outperforms standard LSTM and baseline Transformer models in terms of accuracy, robustness to noise, and long-term forecasting stability. These results establish DeepKoopFormer as a flexible, interpretable, and robust framework for forecasting in high dimensional and dynamical settings.", "authors": ["Ali Forootani", "Mohammad Khosravi", "Masoud Barati"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-08-04", "url": "https://arxiv.org/abs/2508.02616", "pdf_url": "https://arxiv.org/pdf/2508.02616v1", "arxiv_id": "2508.02616", "doi": "10.48550/arXiv.2508.02616", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.1611}
{"id": "68dd32a03deba58079529aaa1c1ae0e809de3e7bfbd0c1cb95420c0acca4ab2c", "sources": ["arxiv", "semantic_scholar"], "title": "L-GTA: Latent Generative Modeling for Time Series Augmentation", "abstract": "Data augmentation is gaining importance across various aspects of time series analysis, from forecasting to classification and anomaly detection tasks. We introduce the Latent Generative Transformer Augmentation (L-GTA) model, a generative approach using a transformer-based variational recurrent autoencoder. This model uses controlled transformations within the latent space of the model to generate new time series that preserve the intrinsic properties of the original dataset. L-GTA enables the application of diverse transformations, ranging from simple jittering to magnitude warping, and combining these basic transformations to generate more complex synthetic time series datasets. Our evaluation of several real-world datasets demonstrates the ability of L-GTA to produce more reliable, consistent, and controllable augmented data. This translates into significant improvements in predictive accuracy and similarity measures compared to direct transformation methods.", "authors": ["Luis Roque", "Carlos Soares", "Vitor Cerqueira", "Luis Torgo"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-07-31", "url": "https://arxiv.org/abs/2507.23615", "pdf_url": "https://arxiv.org/pdf/2507.23615v1", "arxiv_id": "2507.23615", "doi": "10.48550/arXiv.2507.23615", "citation_count": 2, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.2486}
{"id": "2f434967e097ab20eaee3a483b99b15f40c0ad22ba10989294b2afa7b4218705", "sources": ["arxiv", "semantic_scholar"], "title": "Watermarking Large Language Model-based Time Series Forecasting", "abstract": "Large Language Model-based Time Series Forecasting (LLMTS) has shown remarkable promise in handling complex and diverse temporal data, representing a significant step toward foundation models for time series analysis. However, this emerging paradigm introduces two critical challenges. First, the substantial commercial potential and resource-intensive development raise urgent concerns about intellectual property (IP) protection. Second, their powerful time series forecasting capabilities may be misused to produce misleading or fabricated deepfake time series data. To address these concerns, we explore watermarking the outputs of LLMTS models, that is, embedding imperceptible signals into the generated time series data that remain detectable by specialized algorithms. We propose a novel post-hoc watermarking framework, Waltz, which is broadly compatible with existing LLMTS models. Waltz is inspired by the empirical observation that time series patch embeddings are rarely aligned with a specific set of LLM tokens, which we term ``cold tokens''. Leveraging this insight, Waltz embeds watermarks by rewiring the similarity statistics between patch embeddings and cold token embeddings, and detects watermarks using similarity z-scores. To minimize potential side effects, we introduce a similarity-based embedding position identification strategy and employ projected gradient descent to constrain the watermark noise within a defined boundary. Extensive experiments using two popular LLMTS models across seven benchmark datasets demonstrate that Waltz achieves high watermark detection accuracy with minimal impact on the quality of the generated time series.", "authors": ["Wei Yuan", "Chaoqun Yang", "Yu Xing", "Tong Chen", "Nguyen Quoc Viet Hung", "Hongzhi Yin"], "categories": ["cs.IR"], "fields_of_study": ["Computer Science"], "published_date": "2025-07-28", "url": "https://arxiv.org/abs/2507.20762", "pdf_url": "https://arxiv.org/pdf/2507.20762v1", "arxiv_id": "2507.20762", "doi": "10.48550/arXiv.2507.20762", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.2452}
{"id": "226f1fab29d7fe40ceb4b2dea088c023323c6c6e7573ec2a12f70f5521e36147", "sources": ["arxiv", "semantic_scholar"], "title": "U-Cast: Learning Hierarchical Structures for High-Dimensional Time Series Forecasting", "abstract": "Time series forecasting (TSF) is a central problem in time series analysis. However, as the number of channels in time series datasets scales to the thousands or more, a scenario we define as High-Dimensional Time Series Forecasting (HDTSF), it introduces significant new modeling challenges that are often not the primary focus of traditional TSF research. HDTSF is challenging because the channel correlation often forms complex and hierarchical patterns. Existing TSF models either ignore these interactions or fail to scale as dimensionality grows. To address this issue, we propose U-Cast, a channel-dependent forecasting architecture that learns latent hierarchical channel structures with an innovative query-based attention. To disentangle highly correlated channel representation, U-Cast adds a full-rank regularization during training. We also release Time-HD, the first benchmark of large, diverse, high-dimensional datasets. Our theory shows that exploiting cross-channel information lowers forecasting risk, and experiments on Time-HD demonstrate that U-Cast surpasses strong baselines in both accuracy and efficiency. Together, U-Cast and Time-HD provide a solid basis for future HDTSF research.", "authors": ["Juntong Ni", "Shiyu Wang", "Zewen Liu", "Xiaoming Shi", "Xinyue Zhong", "Zhou Ye", "Wei Jin"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-07-20", "url": "https://arxiv.org/abs/2507.15119", "pdf_url": "https://arxiv.org/pdf/2507.15119v2", "arxiv_id": "2507.15119", "doi": null, "citation_count": 6, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/UnifiedTSAI/Time-HD-Lib", "venue": null, "quality_score": 0.279}
{"id": "8993d91e09dd0f68038d3f71ee7974a0801bac90cea0088fae612395a5546ab1", "sources": ["arxiv", "semantic_scholar"], "title": "The Power of Architecture: Deep Dive into Transformer Architectures for Long-Term Time Series Forecasting", "abstract": "Transformer-based models have recently become dominant in Long-term Time Series Forecasting (LTSF), yet the variations in their architecture, such as encoder-only, encoder-decoder, and decoder-only designs, raise a crucial question: What Transformer architecture works best for LTSF tasks? However, existing models are often tightly coupled with various time-series-specific designs, making it difficult to isolate the impact of the architecture itself. To address this, we propose a novel taxonomy that disentangles these designs, enabling clearer and more unified comparisons of Transformer architectures. Our taxonomy considers key aspects such as attention mechanisms, forecasting aggregations, forecasting paradigms, and normalization layers. Through extensive experiments, we uncover several key insights: bi-directional attention with joint-attention is most effective; more complete forecasting aggregation improves performance; and the direct-mapping paradigm outperforms autoregressive approaches. Furthermore, our combined model, utilizing optimal architectural choices, consistently outperforms several existing models, reinforcing the validity of our conclusions. We hope these findings offer valuable guidance for future research on Transformer architectural designs in LTSF. Our code is available at https://github.com/HALF111/TSF_architecture.", "authors": ["Lefei Shen", "Mouxiang Chen", "Han Fu", "Xiaoxue Ren", "Xiaoyun Joy Wang", "Jianling Sun", "Zhuo Li", "Chenghao Liu"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-07-17", "url": "https://arxiv.org/abs/2507.13043", "pdf_url": "https://arxiv.org/pdf/2507.13043v1", "arxiv_id": "2507.13043", "doi": "10.48550/arXiv.2507.13043", "citation_count": 0, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/HALF111/TSF_architecture", "venue": "arXiv.org", "quality_score": 0.3595}
{"id": "521168c649bc0734c684bc5a5c49044aaff943a7678994a17a49ef2f34b63b58", "sources": ["arxiv", "semantic_scholar"], "title": "MoTM: Towards a Foundation Model for Time Series Imputation based on Continuous Modeling", "abstract": "Recent years have witnessed a growing interest for time series foundation models, with a strong emphasis on the forecasting task. Yet, the crucial task of out-of-domain imputation of missing values remains largely underexplored. We propose a first step to fill this gap by leveraging implicit neural representations (INRs). INRs model time series as continuous functions and naturally handle various missing data scenarios and sampling rates. While they have shown strong performance within specific distributions, they struggle under distribution shifts. To address this, we introduce MoTM (Mixture of Timeflow Models), a step toward a foundation model for time series imputation. Building on the idea that a new time series is a mixture of previously seen patterns, MoTM combines a basis of INRs, each trained independently on a distinct family of time series, with a ridge regressor that adapts to the observed context at inference. We demonstrate robust in-domain and out-of-domain generalization across diverse imputation scenarios (e.g., block and pointwise missingness, variable sampling rates), paving the way for adaptable foundation imputation models.", "authors": ["Etienne Le Naour", "Tahar Nabil", "Ghislain Agoua"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-07-17", "url": "https://arxiv.org/abs/2507.13207", "pdf_url": "https://arxiv.org/pdf/2507.13207v3", "arxiv_id": "2507.13207", "doi": "10.48550/arXiv.2507.13207", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.148}
{"id": "0b68cc65fcb4e113d4e823009d0ff4c23b4e9b22346b338888372810900407f1", "sources": ["arxiv", "semantic_scholar"], "title": "Towards Interpretable Time Series Foundation Models", "abstract": "In this paper, we investigate the distillation of time series reasoning capabilities into small, instruction-tuned language models as a step toward building interpretable time series foundation models. Leveraging a synthetic dataset of mean-reverting time series with systematically varied trends and noise levels, we generate natural language annotations using a large multimodal model and use these to supervise the fine-tuning of compact Qwen models. We introduce evaluation metrics that assess the quality of the distilled reasoning - focusing on trend direction, noise intensity, and extremum localization - and show that the post-trained models acquire meaningful interpretive capabilities. Our results highlight the feasibility of compressing time series understanding into lightweight, language-capable models suitable for on-device or privacy-sensitive deployment. This work contributes a concrete foundation toward developing small, interpretable models that explain temporal patterns in natural language.", "authors": ["Matthieu Boileau", "Philippe Helluy", "Jeremy Pawlus", "Svitlana Vyetrenko"], "categories": ["cs.CL", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-07-10", "url": "https://arxiv.org/abs/2507.07439", "pdf_url": "https://arxiv.org/pdf/2507.07439v1", "arxiv_id": "2507.07439", "doi": "10.48550/arXiv.2507.07439", "citation_count": 2, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.2246}
{"id": "e11e4e4ced755b2607aa214ddc29e7017b3d71ed131a1cad0d673739671303a2", "sources": ["arxiv", "semantic_scholar"], "title": "Time Series Foundation Models for Multivariate Financial Time Series Forecasting", "abstract": "Financial time series forecasting presents significant challenges due to complex nonlinear relationships, temporal dependencies, variable interdependencies and limited data availability, particularly for tasks involving low-frequency data, newly listed instruments, or emerging market assets. Time Series Foundation Models (TSFMs) offer a promising solution through pretraining on diverse time series corpora followed by task-specific adaptation. This study evaluates two TSFMs (Tiny Time Mixers (TTM) and Chronos) across three financial forecasting tasks: US 10-year Treasury yield changes, EUR/USD volatility, and equity spread prediction. Results demonstrate that TTM exhibits strong transferability. When fine-tuning both the pretrained version of TTM and an untrained model with the same architecture, the pretrained version achieved 25-50% better performance when fine-tuned on limited data and 15-30% improvements even when fine-tuned on lengthier datasets. Notably, TTM's zero-shot performance outperformed naive benchmarks in volatility forecasting and equity spread prediction, with the latter demonstrating that TSFMs can surpass traditional benchmark models without fine-tuning. The pretrained model consistently required 3-10 fewer years of data to achieve comparable performance levels compared to the untrained model, demonstrating significant sample-efficiency gains. However, while TTM outperformed naive baselines, traditional specialised models matched or exceeded its performance in two of three tasks, suggesting TSFMs prioritise breadth over task-specific optimisation. These findings indicate that TSFMs, though still nascent, offer substantial promise for financial forecasting-particularly in noisy, data-constrained tasks-but achieving competitive performance likely requires domain-specific pretraining and architectural refinements tailored to financial time series characteristics.", "authors": ["Ben A. Marconi"], "categories": ["q-fin.GN", "cs.LG"], "fields_of_study": ["Economics", "Computer Science"], "published_date": "2025-07-09", "url": "https://arxiv.org/abs/2507.07296", "pdf_url": "https://arxiv.org/pdf/2507.07296v1", "arxiv_id": "2507.07296", "doi": "10.48550/arXiv.2507.07296", "citation_count": 4, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.2234}
{"id": "3cbed2298b8acdb4aead158ec98d2fb003de64b14bc090349b1cb28f8cc75055", "sources": ["arxiv", "semantic_scholar"], "title": "Foundation models for time series forecasting: Application in conformal prediction", "abstract": "The zero-shot capabilities of foundation models (FMs) for time series forecasting offer promising potentials in conformal prediction, as most of the available data can be allocated to calibration. This study compares the performance of Time Series Foundation Models (TSFMs) with traditional methods, including statistical models and gradient boosting, within a conformal prediction setting. Our findings highlight two key advantages of TSFMs. First, when the volume of data is limited, TSFMs provide more reliable conformalized prediction intervals than classic models, thanks to their superior predictive accuracy. Second, the calibration process is more stable because more data are used for calibration. Morever, the fewer data available, the more pronounced these benefits become, as classic models require a substantial amount of data for effective training. These results underscore the potential of foundation models in improving conformal prediction reliability in time series applications, particularly in data-constrained cases. All the code to reproduce the experiments is available.", "authors": ["Sami Achour", "Yassine Bouher", "Duong Nguyen", "Nicolas Chesneau"], "categories": ["cs.LG", "cs.AI", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2025-07-09", "url": "https://arxiv.org/abs/2507.08858", "pdf_url": "https://arxiv.org/pdf/2507.08858v1", "arxiv_id": "2507.08858", "doi": "10.48550/arXiv.2507.08858", "citation_count": 7, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.2258}
{"id": "d5ff59c843c3bdbf8805570bbb824474fd815a3f4d87cc72ffd15e27350dea3d", "sources": ["arxiv", "semantic_scholar"], "title": "Accurate Parameter-Efficient Test-Time Adaptation for Time Series Forecasting", "abstract": "Real-world time series often exhibit a non-stationary nature, degrading the performance of pre-trained forecasting models. Test-Time Adaptation (TTA) addresses this by adjusting models during inference, but existing methods typically update the full model, increasing memory and compute costs. We propose PETSA, a parameter-efficient method that adapts forecasters at test time by only updating small calibration modules on the input and output. PETSA uses low-rank adapters and dynamic gating to adjust representations without retraining. To maintain accuracy despite limited adaptation capacity, we introduce a specialized loss combining three components: (1) a robust term, (2) a frequency-domain term to preserve periodicity, and (3) a patch-wise structural term for structural alignment. PETSA improves the adaptability of various forecasting backbones while requiring fewer parameters than baselines. Experimental results on benchmark datasets show that PETSA achieves competitive or better performance across all horizons. Our code is available at: https://github.com/BorealisAI/PETSA", "authors": ["Heitor R. Medeiros", "Hossein Sharifi-Noghabi", "Gabriel L. Oliveira", "Saghar Irandoust"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-06-29", "url": "https://arxiv.org/abs/2506.23424", "pdf_url": "https://arxiv.org/pdf/2506.23424v1", "arxiv_id": "2506.23424", "doi": "10.48550/arXiv.2506.23424", "citation_count": 8, "influential_citation_count": 2, "has_code": true, "code_url": "https://github.com/BorealisAI/PETSA", "venue": "arXiv.org", "quality_score": 0.3276}
{"id": "4961e4a8e8e5f340626727b26931469ab44c4708a5fe50e4560e17f26b69fdba", "sources": ["arxiv", "semantic_scholar"], "title": "The language of time: a language model perspective on time-series foundation models", "abstract": "With the rise of large language models, the paradigm of training foundation models with massive parameter counts on vast datasets has been adopted in multiple domains to achieve remarkable success. Time series foundation models represent a significant extension of this paradigm, demonstrating exceptional expressive power, generalization, and cross-domain transferability. However, this gives rise to a fundamental paradox: time series data reflect distinct dynamical systems, making cross-domain transfer intuitively implausible, yet this is contradicted by the models' empirical success. To resolve this paradox, this paper investigates, from both theoretical and experimental perspectives, the representation learning mechanisms and generalization capabilities of patch-based time series foundation models. We argue that such models are not merely applying a new architecture but are fundamentally generalizing the representation paradigm of language models by extending deterministic vector-based representations to latent probabilistic distributional forms. Our theoretical analysis supports this framework by demonstrating that continuous time-series patches can be faithfully quantized into a discrete vocabulary whose key statistical properties are highly consistent with those of natural language. This generalization allows time series models to inherit the robust representation and transfer abilities of large language models, thereby explaining their superior performance in temporal tasks. Ultimately, our work provides a rigorous theoretical cornerstone for understanding, evaluating, and improving the safety and reliability of large-scale time series foundation models.", "authors": ["Yi Xie", "Yun Xiong", "Zejian Shi", "Hao Niu", "Zhengfu Liu"], "categories": ["cs.LG", "cs.AI", "cs.CL"], "fields_of_study": ["Computer Science"], "published_date": "2025-06-29", "url": "https://arxiv.org/abs/2507.00078", "pdf_url": "https://arxiv.org/pdf/2507.00078v1", "arxiv_id": "2507.00078", "doi": "10.48550/arXiv.2507.00078", "citation_count": 2, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.212}
{"id": "b1a7acb56a2cc8bf06d1474432404de83b38e2912dc40bfae3796f8befe57eeb", "sources": ["arxiv", "semantic_scholar"], "title": "UniCA: Unified Covariate Adaptation for Time Series Foundation Model", "abstract": "Time Series Foundation Models (TSFMs) have achieved remarkable success through large-scale pretraining. However, their design primarily targets real-valued series, limiting their ability to handle general forecasting tasks involving diverse and often heterogeneous covariates -- such as categorical variables and multimodal data (e.g., images, text) -- which are typically task-specific and difficult to leverage during pretraining. To address this gap, we propose Unified Covariate Adaptation (UniCA), a framework to bridge TSFMs with general covariate-aware forecasting. UniCA first performs covariate homogenization to transform heterogeneous covariates into high-level homogeneous series representations and then fuses them via a unified attention-based fusion mechanism. UniCA is compatible and universal for adaptation with both homogeneous and heterogeneous covariates, incorporating extra covariate information while preserving the generalization ability of TSFMs.Extensive experiments on multiple unimodal and multimodal covariate-aware forecasting benchmarks demonstrate the superiority of UniCA, highlighting the promise of covariate-aware TSFM adaptation in real-world forecasting scenarios.Code: https://github.com/hanlu-nju/UniCA.", "authors": ["Lu Han", "Yu Liu", "Lan Li", "Qiwen Deng", "Jian Jiang", "Yinbo Sun", "Zhe Yu", "Binfeng Wang", "Xingyu Lu", "Lintao Ma", "Han-Jia Ye", "De-Chuan Zhan"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-06-27", "url": "https://arxiv.org/abs/2506.22039", "pdf_url": "https://arxiv.org/pdf/2506.22039v2", "arxiv_id": "2506.22039", "doi": null, "citation_count": 2, "influential_citation_count": 2, "has_code": true, "code_url": "https://github.com/hanlu-nju/UniCA", "venue": null, "quality_score": 0.2478}
{"id": "1579d17e85df5bcd808c7141af3667fdd9a3debfe0d803dce961b2530b2956c2", "sources": ["arxiv", "semantic_scholar"], "title": "Scaling Transformers for Time Series Forecasting: Do Pretrained Large Models Outperform Small-Scale Alternatives?", "abstract": "Large pre-trained models have demonstrated remarkable capabilities across domains, but their effectiveness in time series forecasting remains understudied. This work empirically examines whether pre-trained large-scale time series models (LSTSMs) trained on diverse datasets can outperform traditional non-pretrained small-scale transformers in forecasting tasks. We analyze state-of-the-art (SOTA) pre-trained universal time series models (e.g., Moirai, TimeGPT) alongside conventional transformers, evaluating accuracy, computational efficiency, and interpretability across multiple benchmarks. Our findings reveal the strengths and limitations of pre-trained LSTSMs, providing insights into their suitability for time series tasks compared to task-specific small-scale architectures. The results highlight scenarios where pretraining offers advantages and where simpler models remain competitive.", "authors": ["Sanjay Chakraborty", "Ibrahim Delibasoglu", "Fredrik Heintz"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-06-24", "url": "https://arxiv.org/abs/2507.02907", "pdf_url": "https://arxiv.org/pdf/2507.02907v1", "arxiv_id": "2507.02907", "doi": "10.1007/s10462-025-11481-7", "citation_count": 2, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "Artificial Intelligence Review", "quality_score": 0.2063}
{"id": "30b16377e77e24440b1f7632c173e4e773d4cc3b0f77cd1c7311ec2c94632f48", "sources": ["arxiv", "semantic_scholar"], "title": "A Review of the Long Horizon Forecasting Problem in Time Series Analysis", "abstract": "The long horizon forecasting (LHF) problem has come up in the time series literature for over the last 35 years or so. This review covers aspects of LHF in this period and how deep learning has incorporated variants of trend, seasonality, fourier and wavelet transforms, misspecification bias reduction and bandpass filters while contributing using convolutions, residual connections, sparsity reduction, strided convolutions, attention masks, SSMs, normalization methods, low-rank approximations and gating mechanisms. We highlight time series decomposition techniques, input data preprocessing and dataset windowing schemes that improve performance. Multi-layer perceptron models, recurrent neural network hybrids, self-attention models that improve and/or address the performances of the LHF problem are described, with an emphasis on the feature space construction. Ablation studies are conducted over the ETTm2 dataset in the multivariate and univariate high useful load (HUFL) forecasting contexts, evaluated over the last 4 months of the dataset. The heatmaps of MSE averages per time step over test set series in the horizon show that there is a steady increase in the error proportionate to its length except with xLSTM and Triformer models and motivate LHF as an error propagation problem. The trained models are available here: https://bit.ly/LHFModelZoo", "authors": ["Hans Krupakar", "Kandappan V A"], "categories": ["cs.LG", "cs.ET", "cs.PF", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2025-06-15", "url": "https://arxiv.org/abs/2506.12809", "pdf_url": "https://arxiv.org/pdf/2506.12809v1", "arxiv_id": "2506.12809", "doi": "10.48550/arXiv.2506.12809", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1959}
{"id": "a111bfbe391bfb9ee9b6815db7dac383f7f24d509ac66dfc0eed4b859bf8799d", "sources": ["arxiv", "semantic_scholar"], "title": "DeepSeq: High-Throughput Single-Cell RNA Sequencing Data Labeling via Web Search-Augmented Agentic Generative AI Foundation Models", "abstract": "Generative AI foundation models offer transformative potential for processing structured biological data, particularly in single-cell RNA sequencing, where datasets are rapidly scaling toward billions of cells. We propose the use of agentic foundation models with real-time web search to automate the labeling of experimental data, achieving up to 82.5% accuracy. This addresses a key bottleneck in supervised learning for structured omics data by increasing annotation throughput without manual curation and human error. Our approach enables the development of virtual cell foundation models capable of downstream tasks such as cell-typing and perturbation prediction. As data volume grows, these models may surpass human performance in labeling, paving the way for reliable inference in large-scale perturbation screens. This application demonstrates domain-specific innovation in health monitoring and diagnostics, aligned with efforts like the Human Cell Atlas and Human Tumor Atlas Network.", "authors": ["Saleem A. Al Dajani", "Abel Sanchez", "John R. Williams"], "categories": ["q-bio.GN", "cs.AI", "cs.LG", "cs.SE", "q-bio.QM"], "fields_of_study": ["Biology", "Computer Science"], "published_date": "2025-06-14", "url": "https://arxiv.org/abs/2506.13817", "pdf_url": "https://arxiv.org/pdf/2506.13817v1", "arxiv_id": "2506.13817", "doi": "10.1101/2025.06.17.660107", "citation_count": 3, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "bioRxiv", "quality_score": 0.1948}
{"id": "372bc99f416ce916fd0f6851ddf02543bef3446e24f25c10e115db7f7c1b220b", "sources": ["arxiv", "semantic_scholar"], "title": "ST-MTM: Masked Time Series Modeling with Seasonal-Trend Decomposition for Time Series Forecasting", "abstract": "Forecasting complex time series is an important yet challenging problem that involves various industrial applications. Recently, masked time-series modeling has been proposed to effectively model temporal dependencies for forecasting by reconstructing masked segments from unmasked ones. However, since the semantic information in time series is involved in intricate temporal variations generated by multiple time series components, simply masking a raw time series ignores the inherent semantic structure, which may cause MTM to learn spurious temporal patterns present in the raw data. To capture distinct temporal semantics, we show that masked modeling techniques should address entangled patterns through a decomposition approach. Specifically, we propose ST-MTM, a masked time-series modeling framework with seasonal-trend decomposition, which includes a novel masking method for the seasonal-trend components that incorporates different temporal variations from each component. ST-MTM uses a period masking strategy for seasonal components to produce multiple masked seasonal series based on inherent multi-periodicity and a sub-series masking strategy for trend components to mask temporal regions that share similar variations. The proposed masking method presents an effective pre-training task for learning intricate temporal variations and dependencies. Additionally, ST-MTM introduces a contrastive learning task to support masked modeling by enhancing contextual consistency among multiple masked seasonal representations. Experimental results show that our proposed ST-MTM achieves consistently superior forecasting performance compared to existing masked modeling, contrastive learning, and supervised forecasting methods.", "authors": ["Hyunwoo Seo", "Chiehyeon Lim"], "categories": ["cs.LG", "cs.AI", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2025-06-13", "url": "https://arxiv.org/abs/2507.00013", "pdf_url": "https://arxiv.org/pdf/2507.00013v1", "arxiv_id": "2507.00013", "doi": "10.1145/3690624.3709254", "citation_count": 5, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "Knowledge Discovery and Data Mining", "quality_score": 0.1945}
{"id": "e91a8ff7af33324f2778cd627b99677928505c52397c093275d539edb0e8b933", "sources": ["arxiv", "semantic_scholar"], "title": "Tailored Architectures for Time Series Forecasting: Evaluating Deep Learning Models on Gaussian Process-Generated Data", "abstract": "Developments in Deep Learning have significantly improved time series forecasting by enabling more accurate modeling of complex temporal dependencies inherent in sequential data. The effectiveness of such models is often demonstrated on limited sets of specific real-world data. Although this allows for comparative analysis, it still does not demonstrate how specific data characteristics align with the architectural strengths of individual models. Our research aims at uncovering clear connections between time series characteristics and particular models. We introduce a novel dataset generated using Gaussian Processes, specifically designed to display distinct, known characteristics for targeted evaluations of model adaptability to them. Furthermore, we present TimeFlex, a new model that incorporates a modular architecture tailored to handle diverse temporal dynamics, including trends and periodic patterns. This model is compared to current state-of-the-art models, offering a deeper understanding of how models perform under varied time series conditions.", "authors": ["Victoria Hankemeier", "Malte Schilling"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-06-10", "url": "https://arxiv.org/abs/2506.08977", "pdf_url": "https://arxiv.org/pdf/2506.08977v1", "arxiv_id": "2506.08977", "doi": "10.1109/IJCNN64981.2025.11228652", "citation_count": 0, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/vicky-hnk/time-flex", "venue": "IEEE International Joint Conference on Neural Network", "quality_score": 0.294}
{"id": "0f181b93656b7c0bd710fa2421ef6b84b83caf483806e674f8f79ce651f2fdbd", "sources": ["arxiv", "semantic_scholar"], "title": "MIRA: Medical Time Series Foundation Model for Real-World Health Data", "abstract": "A unified foundation model for medical time series -- pretrained on open access and ethics board-approved medical corpora -- offers the potential to reduce annotation burdens, minimize model customization, and enable robust transfer across clinical institutions, modalities, and tasks, particularly in data-scarce or privacy-constrained environments. However, existing generalist time series foundation models struggle to handle medical time series data due to their inherent challenges, including irregular intervals, heterogeneous sampling rates, and frequent missing values. To address these challenges, we introduce MIRA, a unified foundation model specifically designed for medical time series forecasting. MIRA incorporates a Continuous-Time Rotary Positional Encoding that enables fine-grained modeling of variable time intervals, a frequency-specific mixture-of-experts layer that routes computation across latent frequency regimes to further promote temporal specialization, and a Continuous Dynamics Extrapolation Block based on Neural ODE that models the continuous trajectory of latent states, enabling accurate forecasting at arbitrary target timestamps. Pretrained on a large-scale and diverse medical corpus comprising over 454 billion time points collect from publicly available datasets, MIRA achieves reductions in forecasting errors by an average of 10% and 7% in out-of-distribution and in-distribution scenarios, respectively, when compared to other zero-shot and fine-tuned baselines. We also introduce a comprehensive benchmark spanning multiple downstream clinical tasks, establishing a foundation for future research in medical time series modeling.", "authors": ["Hao Li", "Bowen Deng", "Chang Xu", "Zhiyuan Feng", "Viktor Schlegel", "Yu-Hao Huang", "Yizheng Sun", "Jingyuan Sun", "Kailai Yang", "Yiyao Yu", "Jiang Bian"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-06-09", "url": "https://arxiv.org/abs/2506.07584", "pdf_url": "https://arxiv.org/pdf/2506.07584v7", "arxiv_id": "2506.07584", "doi": "10.48550/arXiv.2506.07584", "citation_count": 14, "influential_citation_count": 2, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.294}
{"id": "9973ef9542e02f57c8079911b40541963ab0918a6c7cb62772632ce60f35b6e0", "sources": ["arxiv", "semantic_scholar"], "title": "Benchmarking Pre-Trained Time Series Models for Electricity Price Forecasting", "abstract": "Accurate electricity price forecasting (EPF) is crucial for effective decision-making in power trading on the spot market. While recent advances in generative artificial intelligence (GenAI) and pre-trained large language models (LLMs) have inspired the development of numerous time series foundation models (TSFMs) for time series forecasting, their effectiveness in EPF remains uncertain. To address this gap, we benchmark several state-of-the-art pretrained models--Chronos-Bolt, Chronos-T5, TimesFM, Moirai, Time-MoE, and TimeGPT--against established statistical and machine learning (ML) methods for EPF. Using 2024 day-ahead auction (DAA) electricity prices from Germany, France, the Netherlands, Austria, and Belgium, we generate daily forecasts with a one-day horizon. Chronos-Bolt and Time-MoE emerge as the strongest among the TSFMs, performing on par with traditional models. However, the biseasonal MSTL model, which captures daily and weekly seasonality, stands out for its consistent performance across countries and evaluation metrics, with no TSFM statistically outperforming it.", "authors": ["Timothée Hornek Amir Sartipi", "Igor Tchappi", "Gilbert Fridgen"], "categories": ["cs.LG", "cs.AI", "q-fin.ST"], "fields_of_study": ["Computer Science", "Economics"], "published_date": "2025-06-09", "url": "https://arxiv.org/abs/2506.08113", "pdf_url": "https://arxiv.org/pdf/2506.08113v2", "arxiv_id": "2506.08113", "doi": "10.1109/EEM64765.2025.11050326", "citation_count": 5, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.1945}
{"id": "cfdf13f2a99445f592a4e996740e1473d4a8ac4a27dbc1aabb4996390d0a0cf0", "sources": ["arxiv", "semantic_scholar"], "title": "LightGTS: A Lightweight General Time Series Forecasting Model", "abstract": "Existing works on general time series forecasting build foundation models with heavy model parameters through large-scale multi-source pre-training. These models achieve superior generalization ability across various datasets at the cost of significant computational burdens and limitations in resource-constrained scenarios. This paper introduces LightGTS, a lightweight general time series forecasting model designed from the perspective of consistent periodical modeling. To handle diverse scales and intrinsic periods in multi-source pre-training, we introduce Periodical Tokenization, which extracts consistent periodic patterns across different datasets with varying scales. To better utilize the periodicity in the decoding process, we further introduce Periodical Parallel Decoding, which leverages historical tokens to improve forecasting. Based on the two techniques above which fully leverage the inductive bias of periods inherent in time series, LightGTS uses a lightweight model to achieve outstanding performance on general time series forecasting. It achieves state-of-the-art forecasting performance on 9 real-world benchmarks in both zero-shot and full-shot settings with much better efficiency compared with existing time series foundation models.", "authors": ["Yihang Wang", "Yuying Qiu", "Peng Chen", "Yang Shu", "Zhongwen Rao", "Lujia Pan", "Bin Yang", "Chenjuan Guo"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-06-06", "url": "https://arxiv.org/abs/2506.06005", "pdf_url": "https://arxiv.org/pdf/2506.06005v1", "arxiv_id": "2506.06005", "doi": "10.48550/arXiv.2506.06005", "citation_count": 27, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "International Conference on Machine Learning", "quality_score": 0.3618}
{"id": "62f6ea69a0a074d176967485ba9119c7d51915ce3f654056f0f49b3df10b489b", "sources": ["arxiv", "semantic_scholar"], "title": "FaCTR: Factorized Channel-Temporal Representation Transformers for Efficient Time Series Forecasting", "abstract": "While Transformers excel in language and vision-where inputs are semantically rich and exhibit univariate dependency structures-their architectural complexity leads to diminishing returns in time series forecasting. Time series data is characterized by low per-timestep information density and complex dependencies across channels and covariates, requiring conditioning on structured variable interactions. To address this mismatch and overparameterization, we propose FaCTR, a lightweight spatiotemporal Transformer with an explicitly structural design. FaCTR injects dynamic, symmetric cross-channel interactions-modeled via a low-rank Factorization Machine into temporally contextualized patch embeddings through a learnable gating mechanism. It further encodes static and dynamic covariates for multivariate conditioning. Despite its compact design, FaCTR achieves state-of-the-art performance on eleven public forecasting benchmarks spanning both short-term and long-term horizons, with its largest variant using close to only 400K parameters-on average 50x smaller than competitive spatiotemporal transformer baselines. In addition, its structured design enables interpretability through cross-channel influence scores-an essential requirement for real-world decision-making. Finally, FaCTR supports self-supervised pretraining, positioning it as a compact yet versatile foundation for downstream time series tasks.", "authors": ["Yash Vijay", "Harini Subramanyan"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-06-05", "url": "https://arxiv.org/abs/2506.05597", "pdf_url": "https://arxiv.org/pdf/2506.05597v1", "arxiv_id": "2506.05597", "doi": "10.48550/arXiv.2506.05597", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1845}
{"id": "65618e4ce5d551d29a92e59cab31a476dee7b485049e9a11ea7f8551415c8c56", "sources": ["arxiv", "semantic_scholar"], "title": "RATFM: Retrieval-augmented Time Series Foundation Model for Anomaly Detection", "abstract": "Inspired by the success of large language models (LLMs) in natural language processing, recent research has explored the building of time series foundation models and applied them to tasks such as forecasting, classification, and anomaly detection. However, their performances vary between different domains and tasks. In LLM-based approaches, test-time adaptation using example-based prompting has become common, owing to the high cost of retraining. In the context of anomaly detection, which is the focus of this study, providing normal examples from the target domain can also be effective. However, time series foundation models do not naturally acquire the ability to interpret or utilize examples or instructions, because the nature of time series data used during training does not encourage such capabilities. To address this limitation, we propose a retrieval augmented time series foundation model (RATFM), which enables pretrained time series foundation models to incorporate examples of test-time adaptation. We show that RATFM achieves a performance comparable to that of in-domain fine-tuning while avoiding domain-dependent fine-tuning. Experiments on the UCR Anomaly Archive, a multi-domain dataset including nine domains, confirms the effectiveness of the proposed approach.", "authors": ["Chihiro Maru", "Shoetsu Sato"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-06-02", "url": "https://arxiv.org/abs/2506.02081", "pdf_url": "https://arxiv.org/pdf/2506.02081v1", "arxiv_id": "2506.02081", "doi": "10.48550/arXiv.2506.02081", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.181}
{"id": "655776f04b7e7f1f577de432dd3678f4bdbc41a2be66ba60172c07bef54ce7a1", "sources": ["arxiv", "semantic_scholar"], "title": "Probabilistic Forecasting for Building Energy Systems using Time-Series Foundation Models", "abstract": "Decision-making in building energy systems critically depends on the predictive accuracy of relevant time-series models. In scenarios lacking extensive data from a target building, foundation models (FMs) represent a promising technology that can leverage prior knowledge from vast and diverse pre-training datasets to construct accurate probabilistic predictors for use in decision-making tools. This paper investigates the applicability and fine-tuning strategies of time-series foundation models (TSFMs) in building energy forecasting. We analyze both full fine-tuning and parameter-efficient fine-tuning approaches, particularly low-rank adaptation (LoRA), by using real-world data from a commercial net-zero energy building to capture signals such as room occupancy, carbon emissions, plug loads, and HVAC energy consumption. Our analysis reveals that the zero-shot predictive performance of TSFMs is generally suboptimal. To address this shortcoming, we demonstrate that employing either full fine-tuning or parameter-efficient fine-tuning significantly enhances forecasting accuracy, even with limited historical data. Notably, fine-tuning with low-rank adaptation (LoRA) substantially reduces computational costs without sacrificing accuracy. Furthermore, fine-tuned TSFMs consistently outperform state-of-the-art deep forecasting models (e.g., temporal fusion transformers) in accuracy, robustness, and generalization across varying building zones and seasonal conditions. These results underline the efficacy of TSFMs for practical, data-constrained building energy management systems, enabling improved decision-making in pursuit of energy efficiency and sustainability.", "authors": ["Young Jin Park", "Francois Germain", "Jing Liu", "Ye Wang", "Toshiaki Koike-Akino", "Gordon Wichern", "Navid Azizan", "Christopher R. Laughman", "Ankush Chakrabarty"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-05-31", "url": "https://arxiv.org/abs/2506.00630", "pdf_url": "https://arxiv.org/pdf/2506.00630v1", "arxiv_id": "2506.00630", "doi": "10.48550/arXiv.2506.00630", "citation_count": 5, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "Energy and Buildings", "quality_score": 0.1945}
{"id": "61459610a9fa242c982e12a44e940fd8f7e4096618568576d69a5d36a73d7795", "sources": ["arxiv", "semantic_scholar"], "title": "Generalisation Bounds of Zero-Shot Economic Forecasting using Time Series Foundation Models", "abstract": "This study investigates zero-shot forecasting capabilities of Time Series Foundation Models (TSFMs) for macroeconomic indicators. We apply TSFMs to forecasting economic indicators under univariate conditions, bypassing the need for train bespoke econometric models using and extensive training datasets. Our experiments were conducted on a case study dataset, without additional customisation. We rigorously back-tested three state-of-the-art TSFMs (Chronos, TimeGPT and Moirai) under data-scarce conditions and structural breaks. Our results demonstrate that appropriately engineered TSFMs can internalise rich economic dynamics, accommodate regime shifts, and deliver well-behaved uncertainty estimates out of the box, while matching state-of-the-art multivariate models on this domain. Our findings suggest that, without any fine-tuning, TSFMs can match or exceed classical models during stable economic conditions. However, they are vulnerable to degradation in performances during periods of rapid shocks. The findings offer guidance to practitioners on when zero-shot deployments are viable for macroeconomic monitoring and strategic planning.", "authors": ["Jittarin Jetwiriyanon", "Teo Susnjak", "Surangika Ranathunga"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-05-30", "url": "https://arxiv.org/abs/2506.15705", "pdf_url": "https://arxiv.org/pdf/2506.15705v2", "arxiv_id": "2506.15705", "doi": "10.3390/make7040135", "citation_count": 5, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "Machine Learning and Knowledge Extraction", "quality_score": 0.1945}
{"id": "386aa658128b1dcc8fc61b1aa1e6700c530b07e248c486af406306cc40e4420d", "sources": ["arxiv", "semantic_scholar"], "title": "Synthetic Time Series Forecasting with Transformer Architectures: Extensive Simulation Benchmarks", "abstract": "Time series forecasting plays a critical role in domains such as energy, finance, and healthcare, where accurate predictions inform decision-making under uncertainty. Although Transformer-based models have demonstrated success in sequential modeling, their adoption for time series remains limited by challenges such as noise sensitivity, long-range dependencies, and a lack of inductive bias for temporal structure. In this work, we present a unified and principled framework for benchmarking three prominent Transformer forecasting architectures-Autoformer, Informer, and Patchtst-each evaluated through three architectural variants: Minimal, Standard, and Full, representing increasing levels of complexity and modeling capacity. We conduct over 1500 controlled experiments on a suite of ten synthetic signals, spanning five patch lengths and five forecast horizons under both clean and noisy conditions. Our analysis reveals consistent patterns across model families. To advance this landscape further, we introduce the Koopman-enhanced Transformer framework, Deep Koopformer, which integrates operator-theoretic latent state modeling to improve stability and interpretability. We demonstrate its efficacy on nonlinear and chaotic dynamical systems. Our results highlight Koopman based Transformer as a promising hybrid approach for robust, interpretable, and theoretically grounded time series forecasting in noisy and complex real-world conditions.", "authors": ["Ali Forootani", "Mohammad Khosravi"], "categories": ["cs.LG", "eess.SY"], "fields_of_study": ["Computer Science", "Engineering"], "published_date": "2025-05-26", "url": "https://arxiv.org/abs/2505.20048", "pdf_url": "https://arxiv.org/pdf/2505.20048v1", "arxiv_id": "2505.20048", "doi": "10.48550/arXiv.2505.20048", "citation_count": 3, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.173}
{"id": "808b93db496b17d2a84aa52f1b483e4628564e183f6f42961964e95d696da59d", "sources": ["arxiv", "semantic_scholar"], "title": "Time-o1: Time-Series Forecasting Needs Transformed Label Alignment", "abstract": "Training time-series forecast models presents unique challenges in designing effective learning objectives. Existing methods predominantly utilize the temporal mean squared error, which faces two critical challenges: (1) label autocorrelation, which leads to bias from the label sequence likelihood; (2) excessive amount of tasks, which increases with the forecast horizon and complicates optimization. To address these challenges, we propose Time-o1, a transformation-augmented learning objective tailored for time-series forecasting. The central idea is to transform the label sequence into decorrelated components with discriminated significance. Models are then trained to align the most significant components, thereby effectively mitigating label autocorrelation and reducing task amount. Extensive experiments demonstrate that Time-o1 achieves state-of-the-art performance and is compatible with various forecast models. Code is available at https://github.com/Master-PLC/Time-o1.", "authors": ["Hao Wang", "Licheng Pan", "Zhichao Chen", "Xu Chen", "Qingyang Dai", "Lei Wang", "Haoxuan Li", "Zhouchen Lin"], "categories": ["cs.LG", "cs.AI", "eess.SY"], "fields_of_study": ["Computer Science", "Engineering"], "published_date": "2025-05-23", "url": "https://arxiv.org/abs/2505.17847", "pdf_url": "https://arxiv.org/pdf/2505.17847v2", "arxiv_id": "2505.17847", "doi": null, "citation_count": 21, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/Master-PLC/Time-o1", "venue": "NeurIPS 2025", "quality_score": 0.3356}
{"id": "cd1634771c9bda6b9c9fdc14e19431657d165f4085bb33bf92e25dd3f22d791b", "sources": ["arxiv", "semantic_scholar"], "title": "Time to Embed: Unlocking Foundation Models for Time Series with Channel Descriptions", "abstract": "Traditional time series models are task-specific and often depend on dataset-specific training and extensive feature engineering. While Transformer-based architectures have improved scalability, foundation models, commonplace in text, vision, and audio, remain under-explored for time series and are largely restricted to forecasting. We introduce $\\textbf{CHARM}$, a foundation embedding model for multivariate time series that learns shared, transferable, and domain-aware representations. To address the unique difficulties of time series foundation learning, $\\textbf{CHARM}$ incorporates architectural innovations that integrate channel-level textual descriptions while remaining invariant to channel order. The model is trained using a Joint Embedding Predictive Architecture (JEPA), with novel augmentation schemes and a loss function designed to improve interpretability and training stability. Our $7$M-parameter model achieves state-of-the-art performance across diverse downstream tasks, setting a new benchmark for time series representation learning.", "authors": ["Utsav Dutta", "Sina Khoshfetrat Pakazad", "Henrik Ohlsson"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-05-20", "url": "https://arxiv.org/abs/2505.14543", "pdf_url": "https://arxiv.org/pdf/2505.14543v1", "arxiv_id": "2505.14543", "doi": "10.48550/arXiv.2505.14543", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1661}
{"id": "5e17cd2c315b8b00435cefdc6c4202de5a87dee24650b7e409a55566447aba18", "sources": ["arxiv", "semantic_scholar"], "title": "This Time is Different: An Observability Perspective on Time Series Foundation Models", "abstract": "We introduce Toto, a time series forecasting foundation model with 151 million parameters. Toto uses a modern decoder-only architecture coupled with architectural innovations designed to account for specific challenges found in multivariate observability time series data. Toto's pre-training corpus is a mixture of observability data, open datasets, and synthetic data, and is 4-10$\\times$ larger than those of leading time series foundation models. Additionally, we introduce BOOM, a large-scale benchmark consisting of 350 million observations across 2,807 real-world time series. For both Toto and BOOM, we source observability data exclusively from Datadog's own telemetry and internal observability metrics. Extensive evaluations demonstrate that Toto achieves state-of-the-art performance on both BOOM and on established general purpose time series forecasting benchmarks. Toto's model weights, inference code, and evaluation scripts, as well as BOOM's data and evaluation code, are all available as open source under the Apache 2.0 License available at https://huggingface.co/Datadog/Toto-Open-Base-1.0 and https://github.com/DataDog/toto.", "authors": ["Ben Cohen", "Emaad Khwaja", "Youssef Doubli", "Salahidine Lemaachi", "Chris Lettieri", "Charles Masson", "Hugo Miccinilli", "Elise Ramé", "Qiqi Ren", "Afshin Rostamizadeh", "Jean Ogier du Terrail", "Anna-Monica Toon", "Kan Wang", "Stephan Xie", "Zongzhe Xu", "Viktoriya Zhukova", "David Asker", "Ameet Talwalkar", "Othmane Abou-Amal"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-05-20", "url": "https://arxiv.org/abs/2505.14766", "pdf_url": "https://arxiv.org/pdf/2505.14766v2", "arxiv_id": "2505.14766", "doi": "10.48550/arXiv.2505.14766", "citation_count": 34, "influential_citation_count": 9, "has_code": true, "code_url": "https://github.com/DataDog/toto", "venue": "arXiv.org", "quality_score": 0.5}
{"id": "46d8a2c93450f8d7a0d481b3cb4471e0123d7e6bf45731d7faad7af1357640dc", "sources": ["arxiv", "semantic_scholar"], "title": "Multi-Order Wavelet Derivative Transform for Deep Time Series Forecasting", "abstract": "In deep time series forecasting, the Fourier Transform (FT) is extensively employed for frequency representation learning. However, it often struggles in capturing multi-scale, time-sensitive patterns. Although the Wavelet Transform (WT) can capture these patterns through frequency decomposition, its coefficients are insensitive to change points in time series, leading to suboptimal modeling. To mitigate these limitations, we introduce the multi-order Wavelet Derivative Transform (WDT) grounded in the WT, enabling the extraction of time-aware patterns spanning both the overall trend and subtle fluctuations. Compared with the standard FT and WT, which model the raw series, the WDT operates on the derivative of the series, selectively magnifying rate-of-change cues and exposing abrupt regime shifts that are particularly informative for time series modeling. Practically, we embed the WDT into a multi-branch framework named WaveTS, which decomposes the input series into multi-scale time-frequency coefficients, refines them via linear layers, and reconstructs them into the time domain via the inverse WDT. Extensive experiments on ten benchmark datasets demonstrate that WaveTS achieves state-of-the-art forecasting accuracy while retaining high computational efficiency.", "authors": ["Ziyu Zhou", "Jiaxi Hu", "Qingsong Wen", "James T. Kwok", "Yuxuan Liang"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-05-17", "url": "https://arxiv.org/abs/2505.11781", "pdf_url": "https://arxiv.org/pdf/2505.11781v2", "arxiv_id": "2505.11781", "doi": "10.48550/arXiv.2505.11781", "citation_count": 7, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.2258}
{"id": "acd92d851a4e234c11ef9f0be55650946392a2dbdebe7edf33d94d25a1f5a25a", "sources": ["arxiv", "semantic_scholar"], "title": "Foundation Time-Series AI Model for Realized Volatility Forecasting", "abstract": "Time series foundation models (FMs) have emerged as a popular paradigm for zero-shot multi-domain forecasting. These models are trained on numerous diverse datasets and claim to be effective forecasters across multiple different time series domains, including financial data. In this study, we evaluate the effectiveness of FMs, specifically the TimesFM model, for volatility forecasting, a core task in financial risk management. We first evaluate TimesFM in its pretrained (zero-shot) form, followed by our custom fine-tuning procedure based on incremental learning, and compare the resulting models against standard econometric benchmarks. While the pretrained model provides a reasonable baseline, our findings show that incremental fine-tuning, which allows the model to adapt to new financial return data over time, is essential for learning volatility patterns effectively. Fine-tuned variants not only improve forecast accuracy but also statistically outperform traditional models, as demonstrated through Diebold-Mariano and Giacomini-White tests. These results highlight the potential of foundation models as scalable and adaptive tools for financial forecasting-capable of delivering strong performance in dynamic market environments when paired with targeted fine-tuning strategies.", "authors": ["Anubha Goel", "Puneet Pasricha", "Martin Magris", "Juho Kanniainen"], "categories": ["q-fin.RM", "q-fin.ST"], "fields_of_study": ["Economics"], "published_date": "2025-05-16", "url": "https://arxiv.org/abs/2505.11163", "pdf_url": "https://arxiv.org/pdf/2505.11163v1", "arxiv_id": "2505.11163", "doi": null, "citation_count": 3, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.1505}
{"id": "a3868734aef90a1af68fd0505eb1e80f25a5b75b64f96ac2183f807d42f81e9c", "sources": ["arxiv", "semantic_scholar"], "title": "OLMA: One Loss for More Accurate Time Series Forecasting", "abstract": "Time series forecasting faces two important but often overlooked challenges. Firstly, the inherent random noise in the time series labels sets a theoretical lower bound for the forecasting error, which is positively correlated with the entropy of the labels. Secondly, neural networks exhibit a frequency bias when modeling the state-space of time series, that is, the model performs well in learning certain frequency bands but poorly in others, thus restricting the overall forecasting performance. To address the first challenge, we prove a theorem that there exists a unitary transformation that can reduce the marginal entropy of multiple correlated Gaussian processes, thereby providing guidance for reducing the lower bound of forecasting error. Furthermore, experiments confirm that Discrete Fourier Transform (DFT) can reduce the entropy in the majority of scenarios. Correspondingly, to alleviate the frequency bias, we jointly introduce supervision in the frequency domain along the temporal dimension through DFT and Discrete Wavelet Transform (DWT). This supervision-side strategy is highly general and can be seamlessly integrated into any supervised learning method. Moreover, we propose a novel loss function named OLMA, which utilizes the frequency domain transformation across both channel and temporal dimensions to enhance forecasting. Finally, the experimental results on multiple datasets demonstrate the effectiveness of OLMA in addressing the above two challenges and the resulting improvement in forecasting accuracy. The results also indicate that the perspectives of entropy and frequency bias provide a new and feasible research direction for time series forecasting. The code is available at: https://github.com/Yuyun1011/OLMA-One-Loss-for-More-Accurate-Time-Series-Forecasting.", "authors": ["Tianyi Shi", "Zhu Meng", "Yue Chen", "Siyang Zheng", "Fei Su", "Jin Huang", "Changrui Ren", "Zhicheng Zhao"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-05-16", "url": "https://arxiv.org/abs/2505.11567", "pdf_url": "https://arxiv.org/pdf/2505.11567v2", "arxiv_id": "2505.11567", "doi": null, "citation_count": 0, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/Yuyun1011/OLMA-One-Loss-for-More-Accurate-Time-Series-Forecasting", "venue": null, "quality_score": 0.1909}
{"id": "02821394854cb6239f7f97c51c044c6f307d09ac4dc914f1b311c81cbda63ffc", "sources": ["arxiv", "semantic_scholar"], "title": "Context-Aware Probabilistic Modeling with LLM for Multimodal Time Series Forecasting", "abstract": "Time series forecasting is important for applications spanning energy markets, climate analysis, and traffic management. However, existing methods struggle to effectively integrate exogenous texts and align them with the probabilistic nature of large language models (LLMs). Current approaches either employ shallow text-time series fusion via basic prompts or rely on deterministic numerical decoding that conflict with LLMs' token-generation paradigm, which limits contextual awareness and distribution modeling. To address these limitations, we propose CAPTime, a context-aware probabilistic multimodal time series forecasting method that leverages text-informed abstraction and autoregressive LLM decoding. Our method first encodes temporal patterns using a pretrained time series encoder, then aligns them with textual contexts via learnable interactions to produce joint multimodal representations. By combining a mixture of distribution experts with frozen LLMs, we enable context-aware probabilistic forecasting while preserving LLMs' inherent distribution modeling capabilities. Experiments on diverse time series forecasting tasks demonstrate the superior accuracy and generalization of CAPTime, particularly in multimodal scenarios. Additional analysis highlights its robustness in data-scarce scenarios through hybrid probabilistic decoding.", "authors": ["Yueyang Yao", "Jiajun Li", "Xingyuan Dai", "MengMeng Zhang", "Xiaoyan Gong", "Fei-Yue Wang", "Yisheng Lv"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-05-16", "url": "https://arxiv.org/abs/2505.10774", "pdf_url": "https://arxiv.org/pdf/2505.10774v2", "arxiv_id": "2505.10774", "doi": "10.48550/arXiv.2505.10774", "citation_count": 1, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1616}
{"id": "3fff9788de500ea471fb198c7a58b9b72deae2ba271122c8d0b9e03ae2dfa4fd", "sources": ["arxiv", "semantic_scholar"], "title": "OLinear: A Linear Model for Time Series Forecasting in Orthogonally Transformed Domain", "abstract": "This paper presents $\\mathbf{OLinear}$, a $\\mathbf{linear}$-based multivariate time series forecasting model that operates in an $\\mathbf{o}$rthogonally transformed domain. Recent forecasting models typically adopt the temporal forecast (TF) paradigm, which directly encode and decode time series in the time domain. However, the entangled step-wise dependencies in series data can hinder the performance of TF. To address this, some forecasters conduct encoding and decoding in the transformed domain using fixed, dataset-independent bases (e.g., sine and cosine signals in the Fourier transform). In contrast, we utilize $\\mathbf{OrthoTrans}$, a data-adaptive transformation based on an orthogonal matrix that diagonalizes the series' temporal Pearson correlation matrix. This approach enables more effective encoding and decoding in the decorrelated feature domain and can serve as a plug-in module to enhance existing forecasters. To enhance the representation learning for multivariate time series, we introduce a customized linear layer, $\\mathbf{NormLin}$, which employs a normalized weight matrix to capture multivariate dependencies. Empirically, the NormLin module shows a surprising performance advantage over multi-head self-attention, while requiring nearly half the FLOPs. Extensive experiments on 24 benchmarks and 140 forecasting tasks demonstrate that OLinear consistently achieves state-of-the-art performance with high efficiency. Notably, as a plug-in replacement for self-attention, the NormLin module consistently enhances Transformer-based forecasters. The code and datasets are available at https://anonymous.4open.science/r/OLinear", "authors": ["Wenzhen Yue", "Yong Liu", "Haoxuan Li", "Hao Wang", "Xianghua Ying", "Ruohao Guo", "Bowei Xing", "Ji Shi"], "categories": ["cs.LG", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2025-05-12", "url": "https://arxiv.org/abs/2505.08550", "pdf_url": "https://arxiv.org/pdf/2505.08550v2", "arxiv_id": "2505.08550", "doi": "10.48550/arXiv.2505.08550", "citation_count": 29, "influential_citation_count": 2, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.3693}
{"id": "a84921b4a388da965beb5795e1954fdacb5e54904c46f5758fc3a41ff9dcaccc", "sources": ["arxiv", "semantic_scholar"], "title": "Adaptive Bayesian Very Short-Term Wind Power Forecasting Based on the Generalised Logit Transformation", "abstract": "Wind power plays an increasingly significant role in achieving the 2050 Net Zero Strategy. Despite its rapid growth, its inherent variability presents challenges in forecasting. Accurately forecasting wind power generation is one key demand for the stable and controllable integration of renewable energy into existing grid operations. This paper proposes an adaptive method for very short-term forecasting that combines the generalised logit transformation with a Bayesian approach. The generalised logit transformation processes double-bounded wind power data to an unbounded domain, facilitating the application of Bayesian methods. A novel adaptive mechanism for updating the transformation shape parameter is introduced to leverage Bayesian updates by recovering a small sample of representative data. Four adaptive forecasting methods are investigated, evaluating their advantages and limitations through an extensive case study of over 100 wind farms ranging four years in the UK. The methods are evaluated using the Continuous Ranked Probability Score and we propose the use of functional reliability diagrams to assess calibration. Results indicate that the proposed Bayesian method with adaptive shape parameter updating outperforms benchmarks, yielding consistent improvements in CRPS and forecast reliability. The method effectively addresses uncertainty, ensuring robust and accurate probabilistic forecasting which is essential for grid integration and decision-making.", "authors": ["Tao Shen", "Jethro Browell", "Daniela Castro-Camilo"], "categories": ["stat.AP", "cs.LG"], "fields_of_study": ["Mathematics", "Computer Science"], "published_date": "2025-05-08", "url": "https://arxiv.org/abs/2505.06310", "pdf_url": "https://arxiv.org/pdf/2505.06310v1", "arxiv_id": "2505.06310", "doi": "10.48550/arXiv.2505.06310", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1524}
{"id": "3cd8629374025582c4080857855944ed752fe965ebfd8e8f9ff8af727bb1a0ac", "sources": ["arxiv", "semantic_scholar"], "title": "SCFormer: Structured Channel-wise Transformer with Cumulative Historical State for Multivariate Time Series Forecasting", "abstract": "The Transformer model has shown strong performance in multivariate time series forecasting by leveraging channel-wise self-attention. However, this approach lacks temporal constraints when computing temporal features and does not utilize cumulative historical series effectively.To address these limitations, we propose the Structured Channel-wise Transformer with Cumulative Historical state (SCFormer). SCFormer introduces temporal constraints to all linear transformations, including the query, key, and value matrices, as well as the fully connected layers within the Transformer. Additionally, SCFormer employs High-order Polynomial Projection Operators (HiPPO) to deal with cumulative historical time series, allowing the model to incorporate information beyond the look-back window during prediction. Extensive experiments on multiple real-world datasets demonstrate that SCFormer significantly outperforms mainstream baselines, highlighting its effectiveness in enhancing time series forecasting. The code is publicly available at https://github.com/ShiweiGuo1995/SCFormer", "authors": ["Shiwei Guo", "Ziang Chen", "Yupeng Ma", "Yunfei Han", "Yi Wang"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-05-05", "url": "https://arxiv.org/abs/2505.02655", "pdf_url": "https://arxiv.org/pdf/2505.02655v1", "arxiv_id": "2505.02655", "doi": "10.48550/arXiv.2505.02655", "citation_count": 1, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/ShiweiGuo1995/SCFormer", "venue": "International Conference on Database Systems for Advanced Applications", "quality_score": 0.2302}
{"id": "0cb7f15b17643387bbd437f04a19f17472ca56a773db51b36d339ef1dcc81a2e", "sources": ["arxiv", "semantic_scholar"], "title": "Empirical Comparison of Lightweight Forecasting Models for Seasonal and Non-Seasonal Time Series", "abstract": "Accurate time series forecasting is essential in many real-time applications that demand both high predictive accuracy and computational efficiency. This study provides an empirical comparison between a Polynomial Classifier and a Radial Basis Function Neural Network (RBFNN) across four real-world time series datasets (weather conditions, gold prices, crude oil prices, and beer production volumes) that cover both seasonal and nonseasonal patterns. Model performance is evaluated by forecasting accuracy (using Mean Absolute Error, Root Mean Squared Error, and Coefficient of Variation of Root Mean Squared Error) and computational time to assess each model's viability for real time forecasting. The results show that the PC yields more accurate and faster forecasts for non seasonal series, whereas the RBFNN performs better on series with pronounced seasonal patterns. From an interpretability standpoint, the polynomial model offers a simpler, more transparent structure (in contrast to the black box nature of neural network), which is advantageous for understanding and trust in real time decision making. The performance differences between PC and RBFNN are statistically significant, as confirmed by paired t tests and Wilcoxon signed rank tests. These findings provide practical guidance for model selection in time series forecasting, indicating that PC may be preferable for quick, interpretable forecasts in non-seasonal contexts, whereas RBFNN is superior for capturing complex seasonal behaviors", "authors": ["Thanh Son Nguyen", "Dang Minh Duc Nguyen", "Van Thanh Nguyen"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-05-02", "url": "https://arxiv.org/abs/2505.01163", "pdf_url": "https://arxiv.org/pdf/2505.01163v1", "arxiv_id": "2505.01163", "doi": "10.48550/arXiv.2505.01163", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1455}
{"id": "afb17a743a7f6e61df69536b22de7cdba795c8dc00a0ab656e6597ad0c404975", "sources": ["arxiv", "semantic_scholar"], "title": "Dual-Forecaster: A Multimodal Time Series Model Integrating Descriptive and Predictive Texts", "abstract": "Most existing single-modal time series models rely solely on numerical series, which suffer from the limitations imposed by insufficient information. Recent studies have revealed that multimodal models can address the core issue by integrating textual information. However, these models focus on either historical or future textual information, overlooking the unique contributions each plays in time series forecasting. Besides, these models fail to grasp the intricate relationships between textual and time series data, constrained by their moderate capacity for multimodal comprehension. To tackle these challenges, we propose Dual-Forecaster, a pioneering multimodal time series model that combines both descriptively historical textual information and predictive textual insights, leveraging advanced multimodal comprehension capability empowered by three well-designed cross-modality alignment techniques. Our comprehensive evaluations on fifteen multimodal time series datasets demonstrate that Dual-Forecaster is a distinctly effective multimodal time series model that outperforms or is comparable to other state-of-the-art models, highlighting the superiority of integrating textual information for time series forecasting. This work opens new avenues in the integration of textual information with numerical time series data for multimodal time series analysis.", "authors": ["Wenfa Wu", "Guanyu Zhang", "Zheng Tan", "Yi Wang", "Hongsheng Qi"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-05-02", "url": "https://arxiv.org/abs/2505.01135", "pdf_url": "https://arxiv.org/pdf/2505.01135v1", "arxiv_id": "2505.01135", "doi": "10.48550/arXiv.2505.01135", "citation_count": 5, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1945}
{"id": "19f8b0f64027c5d56e9b024859b14768723a28864e15bf12d199b461a87710e2", "sources": ["arxiv", "semantic_scholar"], "title": "UniBiomed: A Universal Foundation Model for Grounded Biomedical Image Interpretation", "abstract": "The integration of AI-assisted biomedical image analysis into clinical practice demands AI-generated findings that are not only accurate but also interpretable to clinicians. However, existing biomedical AI models generally lack the ability to simultaneously generate diagnostic findings and localize corresponding biomedical objects. This limitation makes it challenging for clinicians to correlate AI-generated findings with visual evidence (e.g., tiny lesions) in images and interpret the results of AI models. To address this challenge, we introduce UniBiomed, the first universal foundation model for grounded biomedical image interpretation, which is capable of generating accurate diagnostic findings and simultaneously segmenting the corresponding biomedical targets. UniBiomed is based on a novel integration of Multi-modal Large Language Model and Segment Anything Model, which can effectively unify diverse biomedical tasks in universal training for advancing grounded interpretation. To develop UniBiomed, we curate a large-scale dataset comprising over 27 million triplets of images, region annotations, and text descriptions across ten biomedical imaging modalities. Extensive validation on 70 internal and 14 external datasets demonstrated the state-of-the-art performance of UniBiomed in diverse biomedical tasks, including image segmentation, disease recognition, region-aware diagnosis, vision question answering, and report generation. In summary, UniBiomed is a powerful and versatile biomedical foundation model, unlocking the untapped grounded interpretation capability for optimizing AI-assisted biomedical image analysis.", "authors": ["Linshan Wu", "Yuxiang Nie", "Sunan He", "Jiaxin Zhuang", "Luyang Luo", "Tao Li", "Zhuoyao Xie", "Dexuan Chen", "Yinghua Zhao", "Neeraj Mahboobani", "Varut Vardhanabhuti", "Ronald Cheong Kin Chan", "Yifan Peng", "Pranav Rajpurkar", "Hao Chen"], "categories": ["cs.CV"], "fields_of_study": ["Computer Science"], "published_date": "2025-04-30", "url": "https://arxiv.org/abs/2504.21336", "pdf_url": "https://arxiv.org/pdf/2504.21336v3", "arxiv_id": "2504.21336", "doi": "10.48550/arXiv.2504.21336", "citation_count": 9, "influential_citation_count": 2, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.25}
{"id": "5b718b0683d57616592141319660452fbbab552c5ed7834e0b84c11750dcd305", "sources": ["arxiv", "semantic_scholar"], "title": "Evaluating Temporal Plasticity in Foundation Time Series Models for Incremental Fine-tuning", "abstract": "Time series foundation models excel at diverse time series forecasting tasks, but their capacity for continuous improvement through incremental learning remains unexplored. We present the first comprehensive study investigating these models' temporal plasticity - their ability to progressively enhance performance through continual learning while maintaining existing capabilities. Through experiments on real-world datasets exhibiting distribution shifts, we evaluate both conventional deep learning models and foundation models using a novel continual learning framework. Our findings reveal that while traditional models struggle with performance deterioration during incremental fine-tuning, foundation models like Time-MoE and Chronos demonstrate sustained improvement in predictive accuracy. This suggests that optimizing foundation model fine-tuning strategies may be more valuable than developing domain-specific small models. Our research introduces new evaluation methodologies and insights for developing foundation time series models with robust continuous learning capabilities.", "authors": ["Jia Liu", "Cheng Jinguo", "Xia Fang", "Zhenyuan Ma", "Yuankai Wu"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-04-20", "url": "https://arxiv.org/abs/2504.14677", "pdf_url": "https://arxiv.org/pdf/2504.14677v1", "arxiv_id": "2504.14677", "doi": "10.1109/IJCNN64981.2025.11229355", "citation_count": 2, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "IEEE International Joint Conference on Neural Network", "quality_score": 0.1318}
{"id": "991bc9a88af8c28c967fc5d24f35593e90c9becb7a7c8b80124abcda28a5d8ae", "sources": ["arxiv", "semantic_scholar"], "title": "Bridging Distribution Gaps in Time Series Foundation Model Pretraining with Prototype-Guided Normalization", "abstract": "Foundation models have achieved remarkable success across diverse machine-learning domains through large-scale pretraining on large, diverse datasets. However, pretraining on such datasets introduces significant challenges due to substantial mismatches in data distributions, a problem particularly pronounced with time series data. In this paper, we tackle this issue by proposing a domain-aware adaptive normalization strategy within the Transformer architecture. Specifically, we replace the traditional LayerNorm with a prototype-guided dynamic normalization mechanism (ProtoNorm), where learned prototypes encapsulate distinct data distributions, and sample-to-prototype affinity determines the appropriate normalization layer. This mechanism effectively captures the heterogeneity of time series characteristics, aligning pretrained representations with downstream tasks. Through comprehensive empirical evaluation, we demonstrate that our method significantly outperforms conventional pretraining techniques across both classification and forecasting tasks, while effectively mitigating the adverse effects of distribution shifts during pretraining. Incorporating ProtoNorm is as simple as replacing a single line of code. Extensive experiments on diverse real-world time series benchmarks validate the robustness and generalizability of our approach, advancing the development of more versatile time series foundation models.", "authors": ["Peiliang Gong", "Emadeldeen Eldele", "Min Wu", "Zhenghua Chen", "Xiaoli Li", "Daoqiang Zhang"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science", "Medicine"], "published_date": "2025-04-15", "url": "https://arxiv.org/abs/2504.10900", "pdf_url": "https://arxiv.org/pdf/2504.10900v1", "arxiv_id": "2504.10900", "doi": "10.48550/arXiv.2504.10900", "citation_count": 3, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "IEEE Transactions on Neural Networks and Learning Systems", "quality_score": 0.1505}
{"id": "478af868e49d7701b0058dbd400d50c2ff445dde5da86008fb9d7941bb950276", "sources": ["arxiv", "semantic_scholar"], "title": "Foundation Models for Time Series: A Survey", "abstract": "Transformer-based foundation models have emerged as a dominant paradigm in time series analysis, offering unprecedented capabilities in tasks such as forecasting, anomaly detection, classification, trend analysis and many more time series analytical tasks. This survey provides a comprehensive overview of the current state of the art pre-trained foundation models, introducing a novel taxonomy to categorize them across several dimensions. Specifically, we classify models by their architecture design, distinguishing between those leveraging patch-based representations and those operating directly on raw sequences. The taxonomy further includes whether the models provide probabilistic or deterministic predictions, and whether they are designed to work with univariate time series or can handle multivariate time series out of the box. Additionally, the taxonomy encompasses model scale and complexity, highlighting differences between lightweight architectures and large-scale foundation models. A unique aspect of this survey is its categorization by the type of objective function employed during training phase. By synthesizing these perspectives, this survey serves as a resource for researchers and practitioners, providing insights into current trends and identifying promising directions for future research in transformer-based time series modeling.", "authors": ["Siva Rama Krishna Kottapalli", "Karthik Hubli", "Sandeep Chandrashekhara", "Garima Jain", "Sunayana Hubli", "Gayathri Botla", "Ramesh Doddaiah"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-04-05", "url": "https://arxiv.org/abs/2504.04011", "pdf_url": "https://arxiv.org/pdf/2504.04011v1", "arxiv_id": "2504.04011", "doi": "10.48550/arXiv.2504.04011", "citation_count": 28, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.3656}
{"id": "f7a3b8c5b82fada11f5abfd3111926e43e36e0ed8b5da208290366dc38d334ef", "sources": ["arxiv", "semantic_scholar"], "title": "Integrating Quantum-Classical Attention in Patch Transformers for Enhanced Time Series Forecasting", "abstract": "QCAAPatchTF is a quantum attention network integrated with an advanced patch-based transformer, designed for multivariate time series forecasting, classification, and anomaly detection. Leveraging quantum superpositions, entanglement, and variational quantum eigensolver principles, the model introduces a quantum-classical hybrid self-attention mechanism to capture multivariate correlations across time points. For multivariate long-term time series, the quantum self-attention mechanism can reduce computational complexity while maintaining temporal relationships. It then applies the quantum-classical hybrid self-attention mechanism alongside a feed-forward network in the encoder stage of the advanced patch-based transformer. While the feed-forward network learns nonlinear representations for each variable frame, the quantum self-attention mechanism processes individual series to enhance multivariate relationships. The advanced patch-based transformer computes the optimized patch length by dividing the sequence length into a fixed number of patches instead of using an arbitrary set of values. The stride is then set to half of the patch length to ensure efficient overlapping representations while maintaining temporal continuity. QCAAPatchTF achieves state-of-the-art performance in both long-term and short-term forecasting, classification, and anomaly detection tasks, demonstrating state-of-the-art accuracy and efficiency on complex real-world datasets.", "authors": ["Sanjay Chakraborty", "Fredrik Heintz"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-03-31", "url": "https://arxiv.org/abs/2504.00068", "pdf_url": "https://arxiv.org/pdf/2504.00068v1", "arxiv_id": "2504.00068", "doi": "10.48550/arXiv.2504.00068", "citation_count": 1, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1505}
{"id": "989877ec7c238e496726b29964e2b0a812aa35d5deb22233dba56aa658a2973d", "sources": ["arxiv", "semantic_scholar"], "title": "Enhancing Time Series Forecasting with Fuzzy Attention-Integrated Transformers", "abstract": "This paper introduces FANTF (Fuzzy Attention Network-Based Transformers), a novel approach that integrates fuzzy logic with existing transformer architectures to advance time series forecasting, classification, and anomaly detection tasks. FANTF leverages a proposed fuzzy attention mechanism incorporating fuzzy membership functions to handle uncertainty and imprecision in noisy and ambiguous time series data. The FANTF approach enhances its ability to capture complex temporal dependencies and multivariate relationships by embedding fuzzy logic principles into the self-attention module of the existing transformer's architecture. The framework combines fuzzy-enhanced attention with a set of benchmark existing transformer-based architectures to provide efficient predictions, classification and anomaly detection. Specifically, FANTF generates learnable fuzziness attention scores that highlight the relative importance of temporal features and data points, offering insights into its decision-making process. Experimental evaluatios on some real-world datasets reveal that FANTF significantly enhances the performance of forecasting, classification, and anomaly detection tasks over traditional transformer-based models.", "authors": ["Sanjay Chakraborty", "Fredrik Heintz"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-03-31", "url": "https://arxiv.org/abs/2504.00070", "pdf_url": "https://arxiv.org/pdf/2504.00070v1", "arxiv_id": "2504.00070", "doi": "10.48550/arXiv.2504.00070", "citation_count": 1, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1505}
{"id": "fba3201f5ff5f0b246fb98ddfd91d8b320149632c3f7896c7e817f1a8b6e5946", "sources": ["arxiv", "semantic_scholar"], "title": "CITRAS: Covariate-Informed Transformer for Time Series Forecasting", "abstract": "In time series forecasting, covariates represent external factors that influence target variables. Some covariates are observable only in the past (observed covariates, such as recorded weather data), while others are known in advance (known covariates, such as calendar events or discount schedules). Although covariates have the potential to enhance forecasting performance, most deep learning-based forecasting models struggle to address the length discrepancy between variables caused by the future portion of known covariates and fail to leverage them flexibly. Moreover, capturing dependencies between target variables and covariates is non-trivial, as models must accurately reflect the local impact of covariates while simultaneously modeling global cross-variate dependencies. To address these challenges, we propose CITRAS, a decoder-only Transformer that flexibly integrates multiple target variables, observed covariates, and known covariates. While preserving strong autoregressive modeling capabilities, CITRAS introduces two novel mechanisms within patch-wise cross-variate attention: Key-Value (KV) Shift and Attention Score Smoothing. KV Shift seamlessly incorporates the future portion of known covariates into the forecasting process by aligning them with target variables based on their concurrent dependencies. Attention Score Smoothing refines locally accurate patch-wise cross-variate dependencies into global variate-level dependencies by smoothing the historical attention scores. Experimentally, CITRAS demonstrates strong performance across a wide range of real-world datasets in both covariate-informed and multivariate settings, showcasing its versatile ability to leverage cross-variate and cross-time dependencies for improved forecasting accuracy.", "authors": ["Yosuke Yamaguchi", "Issei Suemitsu", "Wenpeng Wei"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-03-31", "url": "https://arxiv.org/abs/2503.24007", "pdf_url": "https://arxiv.org/pdf/2503.24007v4", "arxiv_id": "2503.24007", "doi": "10.1109/ACCESS.2026.3695717", "citation_count": 5, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "IEEE Access", "quality_score": 0.1945}
{"id": "eaf5e55e1ffd924d1aeaf461171ee4b8babfdb951808c9d20535c11a836d791c", "sources": ["arxiv", "semantic_scholar"], "title": "Sentinel: Multi-Patch Transformer with Temporal and Channel Attention for Time Series Forecasting", "abstract": "Transformer-based time series forecasting has recently gained strong interest due to the ability of transformers to model sequential data. Most of the state-of-the-art architectures exploit either temporal or inter-channel dependencies, limiting their effectiveness in multivariate time-series forecasting where both types of dependencies are crucial. We propose Sentinel, a full transformer-based architecture composed of an encoder able to extract contextual information from the channel dimension, and a decoder designed to capture causal relations and dependencies across the temporal dimension. Additionally, we introduce a multi-patch attention mechanism, which leverages the patching process to structure the input sequence in a way that can be naturally integrated into the transformer architecture, replacing the multi-head splitting process. Extensive experiments on standard benchmarks demonstrate that Sentinel, because of its ability to \"monitor\" both the temporal and the inter-channel dimension, achieves better or comparable performance with respect to state-of-the-art approaches.", "authors": ["Davide Villaboni", "Alberto Castellini", "Ivan Luciano Danesi", "Alessandro Farinelli"], "categories": ["cs.LG", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2025-03-22", "url": "https://arxiv.org/abs/2503.17658", "pdf_url": "https://arxiv.org/pdf/2503.17658v1", "arxiv_id": "2503.17658", "doi": "10.48550/arXiv.2503.17658", "citation_count": 6, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.2113}
{"id": "33ea3438e93d397d29f3587b55ce3f36b3ffea5c5c755503ee7c344fff75ee27", "sources": ["arxiv", "semantic_scholar"], "title": "A Note on Local Linear Regression for Time Series in Banach Spaces", "abstract": "This work extends local linear regression to Banach space-valued time series for estimating smoothly varying means and their derivatives in non-stationary data. The asymptotic properties of both the standard and bias-reduced Jackknife estimators are analyzed under mild moment conditions, establishing their convergence rates. Simulation studies assess the finite sample performance of these estimators and compare them with the Nadaraya-Watson estimator. Additionally, the proposed methods are applied to smooth EEG recordings for reconstructing eye movements and to video analysis for detecting pedestrians and abandoned objects.", "authors": ["Florian Heinrichs"], "categories": ["math.ST", "stat.ME"], "fields_of_study": ["Mathematics"], "published_date": "2025-03-19", "url": "https://arxiv.org/abs/2503.15039", "pdf_url": "https://arxiv.org/pdf/2503.15039v1", "arxiv_id": "2503.15039", "doi": null, "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.0605}
{"id": "ac0e51a7be4733b3b450348b11d65f6326dd0c97821c6e9768f6360241903313", "sources": ["arxiv", "semantic_scholar"], "title": "Realized Volatility Forecasting for New Issues and Spin-Offs using Multi-Source Transfer Learning", "abstract": "Forecasting the volatility of financial assets is essential for various financial applications. This paper addresses the challenging task of forecasting the volatility of financial assets with limited historical data, such as new issues or spin-offs, by proposing a multi-source transfer learning approach. Specifically, we exploit complementary source data of assets with a substantial historical data record by selecting source time series instances that are most similar to the limited target data of the new issue/spin-off. Based on these instances and the target data, we estimate linear and non-linear realized volatility models and compare their forecasting performance to forecasts of models trained exclusively on the target data, and models trained on the entire source and target data. The results show that our transfer learning approach outperforms the alternative models and that the integration of complementary data is also beneficial immediately after the initial trading day of the new issue/spin-off.", "authors": ["Andreas Teller", "Uta Pigorsch", "Christian Pigorsch"], "categories": ["cs.LG", "q-fin.CP"], "fields_of_study": ["Computer Science", "Economics"], "published_date": "2025-03-16", "url": "https://arxiv.org/abs/2503.12648", "pdf_url": "https://arxiv.org/pdf/2503.12648v1", "arxiv_id": "2503.12648", "doi": "10.48550/arXiv.2503.12648", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.0917}
{"id": "32ef962053bd3c1180b31fe63f1272a82f9f89e3d72c08daa7d97f9e47079b3e", "sources": ["arxiv", "semantic_scholar"], "title": "Towards Efficient Large Scale Spatial-Temporal Time Series Forecasting via Improved Inverted Transformers", "abstract": "Time series forecasting at scale presents significant challenges for modern prediction systems, particularly when dealing with large sets of synchronized series, such as in a global payment network. In such systems, three key challenges must be overcome for accurate and scalable predictions: 1) emergence of new entities, 2) disappearance of existing entities, and 3) the large number of entities present in the data. The recently proposed Inverted Transformer (iTransformer) architecture has shown promising results by effectively handling variable entities. However, its practical application in large-scale settings is limited by quadratic time and space complexity ($O(N^2)$) with respect to the number of entities $N$. In this paper, we introduce EiFormer, an improved inverted transformer architecture that maintains the adaptive capabilities of iTransformer while reducing computational complexity to linear scale ($O(N)$). Our key innovation lies in restructuring the attention mechanism to eliminate redundant computations without sacrificing model expressiveness. Additionally, we incorporate a random projection mechanism that not only enhances efficiency but also improves prediction accuracy through better feature representation. Extensive experiments on the public LargeST benchmark dataset and a proprietary large-scale time series dataset demonstrate that EiFormer significantly outperforms existing methods in both computational efficiency and forecasting accuracy. Our approach enables practical deployment of transformer-based forecasting in industrial applications where handling time series at scale is essential.", "authors": ["Jiarui Sun", "Chin-Chia Michael Yeh", "Yujie Fan", "Xin Dai", "Xiran Fan", "Zhimeng Jiang", "Uday Singh Saini", "Vivian Lai", "Junpeng Wang", "Huiyuan Chen", "Zhongfang Zhuang", "Yan Zheng", "Girish Chowdhary"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-03-13", "url": "https://arxiv.org/abs/2503.10858", "pdf_url": "https://arxiv.org/pdf/2503.10858v1", "arxiv_id": "2503.10858", "doi": "10.48550/arXiv.2503.10858", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.0882}
{"id": "ffaf299ab8e53b403aed34092fda05b0a0538c79d842532ad0c2b1e3dbe22588", "sources": ["arxiv", "semantic_scholar"], "title": "Minimal Time Series Transformer", "abstract": "Transformer is the state-of-the-art model for many natural language processing, computer vision, and audio analysis problems. Transformer effectively combines information from the past input and output samples in auto-regressive manner so that each sample becomes aware of all inputs and outputs. In sequence-to-sequence (Seq2Seq) modeling, the transformer processed samples become effective in predicting the next output. Time series forecasting is a Seq2Seq problem. The original architecture is defined for discrete input and output sequence tokens, but to adopt it for time series, the model must be adapted for continuous data. This work introduces minimal adaptations to make the original transformer architecture suitable for continuous value time series data.", "authors": ["Joni-Kristian Kämäräinen"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-03-12", "url": "https://arxiv.org/abs/2503.09791", "pdf_url": "https://arxiv.org/pdf/2503.09791v1", "arxiv_id": "2503.09791", "doi": "10.48550/arXiv.2503.09791", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.0871}
{"id": "74dcb74242561e4ed6d767fbf32f573c35c5eca5eb3e50bf5db1cbdf5171dc86", "sources": ["arxiv", "semantic_scholar"], "title": "LangTime: A Language-Guided Unified Model for Time Series Forecasting with Proximal Policy Optimization", "abstract": "Recent research has shown an increasing interest in utilizing pre-trained large language models (LLMs) for a variety of time series applications. However, there are three main challenges when using LLMs as foundational models for time series forecasting: (1) Cross-domain generalization. (2) Cross-modality alignment. (3) Error accumulation in autoregressive frameworks. To address these challenges, we proposed LangTime, a language-guided unified model for time series forecasting that incorporates cross-domain pre-training with reinforcement learning-based fine-tuning. Specifically, LangTime constructs Temporal Comprehension Prompts (TCPs), which include dataset-wise and channel-wise instructions, to facilitate domain adaptation and condense time series into a single token, enabling LLMs to understand better and align temporal data. To improve autoregressive forecasting, we introduce TimePPO, a reinforcement learning-based fine-tuning algorithm. TimePPO mitigates error accumulation by leveraging a multidimensional rewards function tailored for time series and a repeat-based value estimation strategy. Extensive experiments demonstrate that LangTime achieves state-of-the-art cross-domain forecasting performance, while TimePPO fine-tuning effectively enhances the stability and accuracy of autoregressive forecasting.", "authors": ["Wenzhe Niu", "Zongxia Xie", "Yanru Sun", "Wei He", "Man Xu", "Chao Hao"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-03-11", "url": "https://arxiv.org/abs/2503.08271", "pdf_url": "https://arxiv.org/pdf/2503.08271v2", "arxiv_id": "2503.08271", "doi": "10.48550/arXiv.2503.08271", "citation_count": 28, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "International Conference on Machine Learning", "quality_score": 0.3656}
{"id": "3f0a757ef4bdea50e9841599bca677fb9e98fc6633409278987616d0f21822f9", "sources": ["arxiv", "semantic_scholar"], "title": "TimeFound: A Foundation Model for Time Series Forecasting", "abstract": "We present TimeFound, an encoder-decoder transformer-based time series foundation model for out-of-the-box zero-shot forecasting. To handle time series data from various domains, TimeFound employs a multi-resolution patching strategy to capture complex temporal patterns at multiple scales. We pre-train our model with two sizes (200M and 710M parameters) on a large time-series corpus comprising both real-world and synthetic datasets. Over a collection of unseen datasets across diverse domains and forecasting horizons, our empirical evaluations suggest that TimeFound can achieve superior or competitive zero-shot forecasting performance, compared to state-of-the-art time series foundation models.", "authors": ["Congxi Xiao", "Jingbo Zhou", "Yixiong Xiao", "Xinjiang Lu", "Le Zhang", "Hui Xiong"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-03-06", "url": "https://arxiv.org/abs/2503.04118", "pdf_url": "https://arxiv.org/pdf/2503.04118v1", "arxiv_id": "2503.04118", "doi": "10.48550/arXiv.2503.04118", "citation_count": 5, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1945}
{"id": "87ce96c8f4fd658fa28a6b087553ad9365882419f2301c87d46ee671e0dffdb0", "sources": ["arxiv", "semantic_scholar"], "title": "TS-RAG: Retrieval-Augmented Generation based Time Series Foundation Models are Stronger Zero-Shot Forecaster", "abstract": "Large Language Models (LLMs) and Foundation Models (FMs) have recently become prevalent for time series forecasting tasks. While fine-tuning LLMs enables domain adaptation, they often struggle to generalize across diverse and unseen datasets. Moreover, existing Time Series Foundation Models (TSFMs) still face challenges in handling non-stationary dynamics and distribution shifts, largely due to the lack of effective mechanisms for adaptation. To this end, we present TS-RAG, a retrieval-augmented generation framework for time series forecasting that enhances the generalization and interpretability of TSFMs. Specifically, TS-RAG leverages pre-trained time series encoders to retrieve semantically relevant segments from a dedicated knowledge base, enriching the contextual representation of the input query. Furthermore, we propose an Adaptive Retrieval Mixer (ARM) module that dynamically fuses the retrieved patterns with the TSFM's internal representation, improving forecasting accuracy without requiring task-specific fine-tuning. Thorough empirical studies on seven public benchmark datasets demonstrate that TS-RAG achieves state-of-the-art zero-shot forecasting performance, outperforming the existing TSFMs by up to 6.84% across diverse domains while also providing desirable interpretability. Our code and data are available at: https://github.com/UConn-DSIS/TS-RAG", "authors": ["Kanghui Ning", "Zijie Pan", "Yu Liu", "Yushan Jiang", "James Yiming Zhang", "Kashif Rasul", "Anderson Schneider", "Lintao Ma", "Yuriy Nevmyvaka", "Dongjin Song"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-03-06", "url": "https://arxiv.org/abs/2503.07649", "pdf_url": "https://arxiv.org/pdf/2503.07649v4", "arxiv_id": "2503.07649", "doi": "10.48550/arXiv.2503.07649", "citation_count": 35, "influential_citation_count": 3, "has_code": true, "code_url": "https://github.com/UConn-DSIS/TS-RAG", "venue": "arXiv.org", "quality_score": 0.3891}
{"id": "fbf3bca228a4ce7c992c637920301c1b224c8fbbef4c4395d2b5f50966258f98", "sources": ["arxiv", "semantic_scholar"], "title": "Unify and Anchor: A Context-Aware Transformer for Cross-Domain Time Series Forecasting", "abstract": "The rise of foundation models has revolutionized natural language processing and computer vision, yet their best practices to time series forecasting remains underexplored. Existing time series foundation models often adopt methodologies from these fields without addressing the unique characteristics of time series data. In this paper, we identify two key challenges in cross-domain time series forecasting: the complexity of temporal patterns and semantic misalignment. To tackle these issues, we propose the ``Unify and Anchor\" transfer paradigm, which disentangles frequency components for a unified perspective and incorporates external context as domain anchors for guided adaptation. Based on this framework, we introduce ContexTST, a Transformer-based model that employs a time series coordinator for structured representation and the Transformer blocks with a context-informed mixture-of-experts mechanism for effective cross-domain generalization. Extensive experiments demonstrate that ContexTST advances state-of-the-art forecasting performance while achieving strong zero-shot transferability across diverse domains.", "authors": ["Xiaobin Hong", "Jiawen Zhang", "Wenzhong Li", "Sanglu Lu", "Jia Li"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-03-03", "url": "https://arxiv.org/abs/2503.01157", "pdf_url": "https://arxiv.org/pdf/2503.01157v1", "arxiv_id": "2503.01157", "doi": "10.48550/arXiv.2503.01157", "citation_count": 4, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1747}
{"id": "6d216a5ea217ecb699a0b9b57d3dd0931cd65a3bd1dedc1d230c16f5f418d700", "sources": ["arxiv", "semantic_scholar"], "title": "Decision-Focused Fine-Tuning of Time Series Foundation Models for Dispatchable Feeder Optimization", "abstract": "Time series foundation models provide a universal solution for generating forecasts to support optimization problems in energy systems. Those foundation models are typically trained in a prediction-focused manner to maximize forecast quality. In contrast, decision-focused learning directly improves the resulting value of the forecast in downstream optimization rather than merely maximizing forecasting quality. The practical integration of forecast values into forecasting models is challenging, particularly when addressing complex applications with diverse instances, such as buildings. This becomes even more complicated when instances possess specific characteristics that require instance-specific, tailored predictions to increase the forecast value. To tackle this challenge, we use decision-focused fine-tuning within time series foundation models to offer a scalable and efficient solution for decision-focused learning applied to the dispatchable feeder optimization problem. To obtain more robust predictions for scarce building data, we use Moirai as a state-of-the-art foundation model, which offers robust and generalized results with few-shot parameter-efficient fine-tuning. Comparing the decision-focused fine-tuned Moirai with a state-of-the-art classical prediction-focused fine-tuning Morai, we observe an improvement of 9.45% in average total daily costs.", "authors": ["Maximilian Beichter", "Nils Friederich", "Janik Pinter", "Dorina Werling", "Kaleb Phipps", "Sebastian Beichter", "Oliver Neumann", "Ralf Mikut", "Veit Hagenmeyer", "Benedikt Heidrich"], "categories": ["cs.LG", "cs.AI", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2025-03-03", "url": "https://arxiv.org/abs/2503.01936", "pdf_url": "https://arxiv.org/pdf/2503.01936v1", "arxiv_id": "2503.01936", "doi": "10.1016/j.egyai.2025.100533", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "Energy and AI", "quality_score": 0.0768}
{"id": "38c8b89fffd6751d4e126c63f7ba168467bbf0585145090ff0dbdd419ee4ab0a", "sources": ["arxiv", "semantic_scholar"], "title": "Dynamic Gradient Sparsification Training for Few-Shot Fine-tuning of CT Lymph Node Segmentation Foundation Model", "abstract": "Accurate lymph node (LN) segmentation is critical in radiotherapy treatment and prognosis analysis, but is limited by the need for large annotated datasets. While deep learning-based segmentation foundation models show potential in developing high-performing models with fewer samples, their medical adaptation faces LN domain-specific prior deficiencies and inefficient few-shot fine-tuning for complex clinical practices, highlighting the necessity of an LN segmentation foundation model. In this work, we annotated 36,106 visible LNs from 3,346 publicly available head-and-neck CT scans to establish a robust LN segmentation model (nnUNetv2). Building on this, we propose Dynamic Gradient Sparsification Training (DGST), a few-shot fine-tuning approach that preserves foundational knowledge while dynamically updating the most critical parameters of the LN segmentation model with few annotations. We validate it on two publicly available LN segmentation datasets: SegRap2023 and LNQ2023. The results show that DGST outperforms existing few-shot fine-tuning methods, achieving satisfactory performance with limited labeled data. We release the dataset, models and all implementations to facilitate relevant research: https://github.com/Zihaoluoh/LN-Seg-FM.", "authors": ["Zihao Luo", "Zijun Gao", "Wenjun Liao", "Shichuan Zhang", "Guotai Wang", "Xiangde Luo"], "categories": ["cs.CV"], "fields_of_study": ["Computer Science"], "published_date": "2025-03-02", "url": "https://arxiv.org/abs/2503.00748", "pdf_url": "https://arxiv.org/pdf/2503.00748v1", "arxiv_id": "2503.00748", "doi": "10.48550/arXiv.2503.00748", "citation_count": 0, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/Zihaoluoh/LN-Seg-FM", "venue": "International Conference on Medical Image Computing and Computer-Assisted Intervention", "quality_score": 0.1169}
{"id": "b4c32fadb4bca800bf6b33152c04c604a7feffabda1bb9b26f3038f6869f0065", "sources": ["arxiv", "semantic_scholar"], "title": "TimesBERT: A BERT-Style Foundation Model for Time Series Understanding", "abstract": "Time series analysis is crucial in diverse scenarios. Beyond forecasting, considerable real-world tasks are categorized into classification, imputation, and anomaly detection, underscoring different capabilities termed time series understanding in this paper. While GPT-style models have been positioned as foundation models for time series forecasting, the BERT-style architecture, which has made significant advances in natural language understanding, has not been fully unlocked for time series understanding, possibly attributed to the undesirable dropout of essential elements of BERT. In this paper, inspired by the shared multi-granularity structure between multivariate time series and multisentence documents, we design TimesBERT to learn generic representations of time series including temporal patterns and variate-centric characteristics. In addition to a natural adaptation of masked modeling, we propose a parallel task of functional token prediction to embody vital multi-granularity structures. Our model is pre-trained on 260 billion time points across diverse domains. Leveraging multi-granularity representations, TimesBERT achieves state-of-the-art performance across four typical downstream understanding tasks, outperforming task-specific models and language pre-trained backbones, positioning it as a versatile foundation model for time series understanding.", "authors": ["Haoran Zhang", "Yong Liu", "Yunzhong Qiu", "Haixuan Liu", "Zhongyi Pei", "Jianmin Wang", "Mingsheng Long"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-02-28", "url": "https://arxiv.org/abs/2502.21245", "pdf_url": "https://arxiv.org/pdf/2502.21245v1", "arxiv_id": "2502.21245", "doi": "10.1145/3746027.3755238", "citation_count": 14, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "ACM Multimedia", "quality_score": 0.294}
{"id": "33a7345ac7541c0e9e379f1e44e2fc175039da253d8439ead503e176356e5754", "sources": ["arxiv", "semantic_scholar"], "title": "PFformer: A Position-Free Transformer Variant for Extreme-Adaptive Multivariate Time Series Forecasting", "abstract": "Multivariate time series (MTS) forecasting is vital in fields like weather, energy, and finance. However, despite deep learning advancements, traditional Transformer-based models often diminish the effect of crucial inter-variable relationships by singular token embedding and struggle to effectively capture complex dependencies among variables, especially in datasets with rare or extreme events. These events create significant imbalances and lead to high skewness, complicating accurate prediction efforts. This study introduces PFformer, a position-free Transformer-based model designed for single-target MTS forecasting, specifically for challenging datasets characterized by extreme variability. PFformer integrates two novel embedding strategies: Enhanced Feature-based Embedding (EFE) and Auto-Encoder-based Embedding (AEE). EFE effectively encodes inter-variable dependencies by mapping related sequence subsets to high-dimensional spaces without positional constraints, enhancing the encoder's functionality. PFformer shows superior forecasting accuracy without the traditional limitations of positional encoding in MTS modeling. We evaluated PFformer across four challenging datasets, focusing on two key forecasting scenarios: long sequence prediction for 3 days ahead and rolling predictions every four hours to reflect real-time decision-making processes in water management. PFformer demonstrated remarkable improvements, from 20% to 60%, compared with state-of-the-art models.", "authors": ["Yanhong Li", "David C. Anastasiu"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-02-27", "url": "https://arxiv.org/abs/2502.20571", "pdf_url": "https://arxiv.org/pdf/2502.20571v1", "arxiv_id": "2502.20571", "doi": "10.48550/arXiv.2502.20571", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "Pacific-Asia Conference on Knowledge Discovery and Data Mining", "quality_score": 0.0753}
{"id": "17fa64d602a9dbeeb0574d22a253fecb049fa55ed1a908c9b2a087cff13786f7", "sources": ["arxiv", "semantic_scholar"], "title": "Forecasting intermittent time series with Gaussian Processes and Tweedie likelihood", "abstract": "We adopt Gaussian Processes (GPs) as latent functions for probabilistic forecasting of intermittent time series. The model is trained in a Bayesian framework that accounts for the uncertainty about the latent function. We couple the latent GP variable with two types of forecast distributions: the negative binomial (NegBinGP) and the Tweedie distribution (TweedieGP). While the negative binomial has already been used in forecasting intermittent time series, this is the first time in which a fully parameterized Tweedie density is used for intermittent time series. We properly evaluate the Tweedie density, which has both a point mass at zero and heavy tails, avoiding simplifying assumptions made in existing models. We test our models on thousands of intermittent count time series. Results show that our models provide consistently better probabilistic forecasts than the competitors. In particular, TweedieGP obtains the best estimates of the highest quantiles, thus showing that it is more flexible than NegBinGP.", "authors": ["Stefano Damato", "Dario Azzimonti", "Giorgio Corani"], "categories": ["stat.ML", "cs.LG", "stat.AP"], "fields_of_study": ["Mathematics", "Computer Science"], "published_date": "2025-02-26", "url": "https://arxiv.org/abs/2502.19086", "pdf_url": "https://arxiv.org/pdf/2502.19086v5", "arxiv_id": "2502.19086", "doi": "10.1016/j.ijforecast.2025.10.001", "citation_count": 3, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "International Journal of Forecasting", "quality_score": 0.1505}
{"id": "fd40f0de00153a61000abc306d049fc978cd7efd9217ece8e38fe5785f8df1ef", "sources": ["arxiv", "semantic_scholar"], "title": "TSKANMixer: Kolmogorov-Arnold Networks with MLP-Mixer Model for Time Series Forecasting", "abstract": "Time series forecasting has long been a focus of research across diverse fields, including economics, energy, healthcare, and traffic management. Recent works have introduced innovative architectures for time series models, such as the Time-Series Mixer (TSMixer), which leverages multi-layer perceptrons (MLPs) to enhance prediction accuracy by effectively capturing both spatial and temporal dependencies within the data. In this paper, we investigate the capabilities of the Kolmogorov-Arnold Networks (KANs) for time-series forecasting by modifying TSMixer with a KAN layer (TSKANMixer). Experimental results demonstrate that TSKANMixer tends to improve prediction accuracy over the original TSMixer across multiple datasets, ranking among the top-performing models compared to other time series approaches. Our results show that the KANs are promising alternatives to improve the performance of time series forecasting by replacing or extending traditional MLPs.", "authors": ["Young-Chae Hong", "Bei Xiao", "Yangho Chen"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-02-25", "url": "https://arxiv.org/abs/2502.18410", "pdf_url": "https://arxiv.org/pdf/2502.18410v2", "arxiv_id": "2502.18410", "doi": "10.48550/arXiv.2502.18410", "citation_count": 3, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1505}
{"id": "7fa088fa9e2035ffbbabaef8639d417bef2b52a6250500d8b589419b8cc31d9d", "sources": ["arxiv", "semantic_scholar"], "title": "Mantis: Lightweight Calibrated Foundation Model for User-Friendly Time Series Classification", "abstract": "In recent years, there has been increasing interest in developing foundation models for time series data that can generalize across diverse downstream tasks. While numerous forecasting-oriented foundation models have been introduced, there is a notable scarcity of models tailored for time series classification. To address this gap, we present Mantis, a new open-source foundation model for time series classification based on the Vision Transformer (ViT) architecture that has been pre-trained using a contrastive learning approach. Our experimental results show that Mantis outperforms existing foundation models both when the backbone is frozen and when fine-tuned, while achieving the lowest calibration error. In addition, we propose several adapters to handle the multivariate setting, reducing memory requirements and modeling channel interdependence.", "authors": ["Vasilii Feofanov", "Songkang Wen", "Marius Alonso", "Romain Ilbert", "Hongbo Guo", "Malik Tiomoko", "Lujia Pan", "Jianfeng Zhang", "Ievgen Redko"], "categories": ["cs.LG", "cs.AI", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2025-02-21", "url": "https://arxiv.org/abs/2502.15637", "pdf_url": "https://arxiv.org/pdf/2502.15637v1", "arxiv_id": "2502.15637", "doi": "10.48550/arXiv.2502.15637", "citation_count": 27, "influential_citation_count": 4, "has_code": true, "code_url": null, "venue": "arXiv.org", "quality_score": 0.3618}
{"id": "2ad8b834ca27c2417a130898b64a96398defdf1af01c1a19721a522a1b1e3987", "sources": ["arxiv", "semantic_scholar"], "title": "Learning Novel Transformer Architecture for Time-series Forecasting", "abstract": "Despite the success of Transformer-based models in the time-series prediction (TSP) tasks, the existing Transformer architecture still face limitations and the literature lacks comprehensive explorations into alternative architectures. To address these challenges, we propose AutoFormer-TS, a novel framework that leverages a comprehensive search space for Transformer architectures tailored to TSP tasks. Our framework introduces a differentiable neural architecture search (DNAS) method, AB-DARTS, which improves upon existing DNAS approaches by enhancing the identification of optimal operations within the architecture. AutoFormer-TS systematically explores alternative attention mechanisms, activation functions, and encoding operations, moving beyond the traditional Transformer design. Extensive experiments demonstrate that AutoFormer-TS consistently outperforms state-of-the-art baselines across various TSP benchmarks, achieving superior forecasting accuracy while maintaining reasonable training efficiency.", "authors": ["Juyuan Zhang", "Wei Zhu", "Jiechao Gao"], "categories": ["cs.LG", "cs.CL"], "fields_of_study": ["Computer Science"], "published_date": "2025-02-19", "url": "https://arxiv.org/abs/2502.13721", "pdf_url": "https://arxiv.org/pdf/2502.13721v1", "arxiv_id": "2502.13721", "doi": "10.48550/arXiv.2502.13721", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.0753}
{"id": "7a3ace8bbb2e9694c631e933abd004c816906d94651c0d920e3feea93834f248", "sources": ["arxiv", "semantic_scholar"], "title": "Lightweight Online Adaption for Time Series Foundation Model Forecasts", "abstract": "Foundation models (FMs) have emerged as a promising approach for time series forecasting. While effective, FMs typically remain fixed during deployment due to the high computational costs of learning them online. Consequently, deployed FMs fail to adapt their forecasts to current data characteristics, despite the availability of online feedback from newly arriving data. This raises the question of whether FM performance can be enhanced by the efficient usage of this feedback. We propose ELF to answer this question. ELF is a lightweight mechanism for the online adaption of FM forecasts in response to online feedback. ELF consists of two parts: a) the ELF-Forecaster which is used to learn the current data distribution; and b) the ELF-Weighter which is used to combine the forecasts of the FM and the ELF-Forecaster. We evaluate the performance of ELF in conjunction with several recent FMs across a suite of standard time series datasets. In all of our experiments we find that using ELF improves performance. This work demonstrates how efficient usage of online feedback can be used to improve FM forecasts.", "authors": ["Thomas L. Lee", "William Toner", "Rajkarn Singh", "Artjom Joosen", "Martin Asenov"], "categories": ["cs.LG", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2025-02-18", "url": "https://arxiv.org/abs/2502.12920", "pdf_url": "https://arxiv.org/pdf/2502.12920v3", "arxiv_id": "2502.12920", "doi": "10.48550/arXiv.2502.12920", "citation_count": 10, "influential_citation_count": 3, "has_code": false, "code_url": null, "venue": "International Conference on Machine Learning", "quality_score": 0.301}
{"id": "98c201a68bc97b179498d98f9730b3eb3ac58ca19103549b107c7484b701483d", "sources": ["arxiv", "semantic_scholar"], "title": "S2TX: Cross-Attention Multi-Scale State-Space Transformer for Time Series Forecasting", "abstract": "Time series forecasting has recently achieved significant progress with multi-scale models to address the heterogeneity between long and short range patterns. Despite their state-of-the-art performance, we identify two potential areas for improvement. First, the variates of the multivariate time series are processed independently. Moreover, the multi-scale (long and short range) representations are learned separately by two independent models without communication. In light of these concerns, we propose State Space Transformer with cross-attention (S2TX). S2TX employs a cross-attention mechanism to integrate a Mamba model for extracting long-range cross-variate context and a Transformer model with local window attention to capture short-range representations. By cross-attending to the global context, the Transformer model further facilitates variate-level interactions as well as local/global communications. Comprehensive experiments on seven classic long-short range time-series forecasting benchmark datasets demonstrate that S2TX can achieve highly robust SOTA results while maintaining a low memory footprint.", "authors": ["Zihao Wu", "Juncheng Dong", "Haoming Yang", "Vahid Tarokh"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-02-17", "url": "https://arxiv.org/abs/2502.11340", "pdf_url": "https://arxiv.org/pdf/2502.11340v1", "arxiv_id": "2502.11340", "doi": "10.48550/arXiv.2502.11340", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.0753}
{"id": "981b28b451563972cad4e0d048cccfa267ef315807bbfd45b341d207c2698db1", "sources": ["arxiv", "semantic_scholar"], "title": "Vision-Enhanced Time Series Forecasting via Latent Diffusion Models", "abstract": "Diffusion models have recently emerged as powerful frameworks for generating high-quality images. While recent studies have explored their application to time series forecasting, these approaches face significant challenges in cross-modal modeling and transforming visual information effectively to capture temporal patterns. In this paper, we propose LDM4TS, a novel framework that leverages the powerful image reconstruction capabilities of latent diffusion models for vision-enhanced time series forecasting. Instead of introducing external visual data, we are the first to use complementary transformation techniques to convert time series into multi-view visual representations, allowing the model to exploit the rich feature extraction capabilities of the pre-trained vision encoder. Subsequently, these representations are reconstructed using a latent diffusion model with a cross-modal conditioning mechanism as well as a fusion module. Experimental results demonstrate that LDM4TS outperforms various specialized forecasting models for time series forecasting tasks.", "authors": ["Weilin Ruan", "Siru Zhong", "Haomin Wen", "Yuxuan Liang"], "categories": ["cs.CV", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-02-16", "url": "https://arxiv.org/abs/2502.14887", "pdf_url": "https://arxiv.org/pdf/2502.14887v1", "arxiv_id": "2502.14887", "doi": "10.48550/arXiv.2502.14887", "citation_count": 12, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.2785}
{"id": "06a8232edc8b85003bc0e9cd7a73f4f5f196b051bb27a20e521a614c432e8675", "sources": ["arxiv", "semantic_scholar"], "title": "AdaPTS: Adapting Univariate Foundation Models to Probabilistic Multivariate Time Series Forecasting", "abstract": "Pre-trained foundation models (FMs) have shown exceptional performance in univariate time series forecasting tasks. However, several practical challenges persist, including managing intricate dependencies among features and quantifying uncertainty in predictions. This study aims to tackle these critical limitations by introducing adapters; feature-space transformations that facilitate the effective use of pre-trained univariate time series FMs for multivariate tasks. Adapters operate by projecting multivariate inputs into a suitable latent space and applying the FM independently to each dimension. Inspired by the literature on representation learning and partially stochastic Bayesian neural networks, we present a range of adapters and optimization/inference strategies. Experiments conducted on both synthetic and real-world datasets confirm the efficacy of adapters, demonstrating substantial enhancements in forecasting accuracy and uncertainty quantification compared to baseline methods. Our framework, AdaPTS, positions adapters as a modular, scalable, and effective solution for leveraging time series FMs in multivariate contexts, thereby promoting their wider adoption in real-world applications. We release the code at https://github.com/abenechehab/AdaPTS.", "authors": ["Abdelhakim Benechehab", "Vasilii Feofanov", "Giuseppe Paolo", "Albert Thomas", "Maurizio Filippone", "Balázs Kégl"], "categories": ["stat.ML", "cs.LG"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2025-02-14", "url": "https://arxiv.org/abs/2502.10235", "pdf_url": "https://arxiv.org/pdf/2502.10235v1", "arxiv_id": "2502.10235", "doi": "10.48550/arXiv.2502.10235", "citation_count": 10, "influential_citation_count": 3, "has_code": true, "code_url": "https://github.com/abenechehab/AdaPTS", "venue": "International Conference on Machine Learning", "quality_score": 0.301}
{"id": "ee14006a920be3f17212eae93435789671b0cbf59aadd320bb91c32fd0986ad7", "sources": ["arxiv", "semantic_scholar"], "title": "Forecasting time series with constraints", "abstract": "Time series forecasting presents unique challenges that limit the effectiveness of traditional machine learning algorithms. To address these limitations, various approaches have incorporated linear constraints into learning algorithms, such as generalized additive models and hierarchical forecasting. In this paper, we propose a unified framework for integrating and combining linear constraints in time series forecasting. Within this framework, we show that the exact minimizer of the constrained empirical risk can be computed efficiently using linear algebra alone. This approach allows for highly scalable implementations optimized for GPUs. We validate the proposed methodology through extensive benchmarking on real-world tasks, including electricity demand forecasting and tourism forecasting, achieving state-of-the-art performance.", "authors": ["Nathan Doumèche", "Francis Bach", "Éloi Bedek", "Gérard Biau", "Claire Boyer", "Yannig Goude"], "categories": ["stat.ML", "cs.AI", "cs.LG", "math.ST", "stat.AP", "stat.ME"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2025-02-14", "url": "https://arxiv.org/abs/2502.10485", "pdf_url": "https://arxiv.org/pdf/2502.10485v1", "arxiv_id": "2502.10485", "doi": "10.48550/arXiv.2502.10485", "citation_count": 2, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1193}
{"id": "ee25370e778d2508da35092049c1a9384bb4bfa0130453e6968c067f62d5c0a1", "sources": ["arxiv", "semantic_scholar"], "title": "Fine-Tuning Foundation Models with Federated Learning for Privacy Preserving Medical Time Series Forecasting", "abstract": "Federated Learning (FL) provides a decentralized machine learning approach, where multiple devices or servers collaboratively train a model without sharing their raw data, thus enabling data privacy. This approach has gained significant interest in academia and industry due to its privacy-preserving properties, which are particularly valuable in the medical domain where data availability is often protected under strict regulations. A relatively unexplored area is the use of FL to fine-tune Foundation Models (FMs) for time series forecasting, potentially enhancing model efficacy by overcoming data limitation while maintaining privacy. In this paper, we fine-tuned time series FMs with Electrocardiogram (ECG) and Impedance Cardiography (ICG) data using different FL techniques. We then examined various scenarios and discussed the challenges FL faces under different data heterogeneity configurations. Our empirical results demonstrated that while FL can be effective for fine-tuning FMs on time series forecasting tasks, its benefits depend on the data distribution across clients. We highlighted the trade-offs in applying FL to FM fine-tuning.", "authors": ["Mahad Ali", "Curtis Lisle", "Patrick W. Moore", "Tammer Barkouki", "Brian J. Kirkwood", "Laura J. Brattain"], "categories": ["cs.LG", "cs.CR"], "fields_of_study": ["Computer Science", "Medicine"], "published_date": "2025-02-13", "url": "https://arxiv.org/abs/2502.09744", "pdf_url": "https://arxiv.org/pdf/2502.09744v1", "arxiv_id": "2502.09744", "doi": "10.1109/EMBC58623.2025.11254049", "citation_count": 2, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "Annual International Conference of the IEEE Engineering in Medicine and Biology Society", "quality_score": 0.1193}
{"id": "eaeefeeeb301aac9794a4ae01b559ab4c7b17acf026c188452a36f58be5b8a39", "sources": ["arxiv", "semantic_scholar"], "title": "HDT: Hierarchical Discrete Transformer for Multivariate Time Series Forecasting", "abstract": "Generative models have gained significant attention in multivariate time series forecasting (MTS), particularly due to their ability to generate high-fidelity samples. Forecasting the probability distribution of multivariate time series is a challenging yet practical task. Although some recent attempts have been made to handle this task, two major challenges persist: 1) some existing generative methods underperform in high-dimensional multivariate time series forecasting, which is hard to scale to higher dimensions; 2) the inherent high-dimensional multivariate attributes constrain the forecasting lengths of existing generative models. In this paper, we point out that discrete token representations can model high-dimensional MTS with faster inference time, and forecasting the target with long-term trends of itself can extend the forecasting length with high accuracy. Motivated by this, we propose a vector quantized framework called Hierarchical Discrete Transformer (HDT) that models time series into discrete token representations with l2 normalization enhanced vector quantized strategy, in which we transform the MTS forecasting into discrete tokens generation. To address the limitations of generative models in long-term forecasting, we propose a hierarchical discrete Transformer. This model captures the discrete long-term trend of the target at the low level and leverages this trend as a condition to generate the discrete representation of the target at the high level that introduces the features of the target itself to extend the forecasting length in high-dimensional MTS. Extensive experiments on five popular MTS datasets verify the effectiveness of our proposed method.", "authors": ["Shibo Feng", "Peilin Zhao", "Liu Liu", "Pengcheng Wu", "Zhiqi Shen"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-02-12", "url": "https://arxiv.org/abs/2502.08302", "pdf_url": "https://arxiv.org/pdf/2502.08302v1", "arxiv_id": "2502.08302", "doi": "10.48550/arXiv.2502.08302", "citation_count": 8, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "AAAI Conference on Artificial Intelligence", "quality_score": 0.2386}
{"id": "633040e4556306f2ced991a5796f481c23b016d21507de9a4fb439af33e31981", "sources": ["arxiv", "semantic_scholar"], "title": "Two-stage hybrid models for enhancing forecasting accuracy on heterogeneous time series", "abstract": "A time series forecasting model--which is typically built on a single time series--is known as a local time series model (tsLM). In contrast, a forecasting model trained on multiple time series is referred to as a global time series model (tsGM). tsGMs can enhance forecasting accuracy and improve generalisation by learning cross-series information. As such, developing tsGMs has become a prominent research focus within the time series forecasting community. However, the benefits of tsGMs may not always be realised if the given set of time series is heterogeneous. While increasing model complexity can help tsGMs adapt to such a set of data, it can also increase the risk of overfitting and forecasting error. Additionally, the definition of homogeneity remains ambiguous in the literature. To address these challenges, this paper explores how to define data heterogeneity and proposes a two-stage modelling framework: At stage one, a tsGM is learnt to identify homogeneous patterns; and at stage two, tsLMs (e.g., ARIMA) or sub-tsGMs tailored to different groups are learnt to capture the heterogeneity. Numerical experiments on four open datasets demonstrate that the proposed approach significantly outperforms six state-of-the-art models. These results highlight its effectiveness in unlocking the full potential of global forecasting models for heterogeneous datasets.", "authors": ["Junru Ren", "Shaomin Wu"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-02-12", "url": "https://arxiv.org/abs/2502.08600", "pdf_url": "https://arxiv.org/pdf/2502.08600v2", "arxiv_id": "2502.08600", "doi": "10.3233/FAIA251157", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "European Conference on Artificial Intelligence", "quality_score": 0.055}
{"id": "ddb11daae9e720a3b5d0c36ec828619c351b9e75468ba77e74c1665cbb687a84", "sources": ["arxiv", "semantic_scholar"], "title": "Transformers and Their Roles as Time Series Foundation Models", "abstract": "We give a comprehensive analysis of transformers as time series foundation models, focusing on their approximation and generalization capabilities. First, we demonstrate that there exist transformers that fit an autoregressive model on input univariate time series via gradient descent. We then analyze MOIRAI, a multivariate time series foundation model capable of handling an arbitrary number of covariates. We prove that it is capable of automatically fitting autoregressive models with an arbitrary number of covariates, offering insights into its design and empirical success. For generalization, we establish bounds for pretraining when the data satisfies Dobrushin's condition. Experiments support our theoretical findings, highlighting the efficacy of transformers as time series foundation models.", "authors": ["Dennis Wu", "Yihan He", "Yuan Cao", "Jianqing Fan", "Han Liu"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-02-05", "url": "https://arxiv.org/abs/2502.03383", "pdf_url": "https://arxiv.org/pdf/2502.03383v1", "arxiv_id": "2502.03383", "doi": "10.48550/arXiv.2502.03383", "citation_count": 4, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1747}
{"id": "68d35d229fa293ea3bbfd081c33cb03e07b34bf662539726ac3ff3df83df616f", "sources": ["arxiv", "semantic_scholar"], "title": "Benchmarking Time Series Forecasting Models: From Statistical Techniques to Foundation Models in Real-World Applications", "abstract": "Time series forecasting is essential for operational intelligence in the hospitality industry, and particularly challenging in large-scale, distributed systems. This study evaluates the performance of statistical, machine learning (ML), deep learning, and foundation models in forecasting hourly sales over a 14-day horizon using real-world data from a network of thousands of restaurants across Germany. The forecasting solution includes features such as weather conditions, calendar events, and time-of-day patterns. Results demonstrate the strong performance of ML-based meta-models and highlight the emerging potential of foundation models like Chronos and TimesFM, which deliver competitive performance with minimal feature engineering, leveraging only the pre-trained model (zero-shot inference). Additionally, a hybrid PySpark-Pandas approach proves to be a robust solution for achieving horizontal scalability in large-scale deployments.", "authors": ["Issar Arab", "Rodrigo Benitez"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-02-05", "url": "https://arxiv.org/abs/2502.03395", "pdf_url": "https://arxiv.org/pdf/2502.03395v1", "arxiv_id": "2502.03395", "doi": "10.48550/arXiv.2502.03395", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.0753}
{"id": "50ebf1b3f50467b5ac756d359b1ed31d5281383a4176a1578aec848fa3eefbe4", "sources": ["arxiv", "semantic_scholar"], "title": "Sundial: A Family of Highly Capable Time Series Foundation Models", "abstract": "We introduce Sundial, a family of native, flexible, and scalable time series foundation models. To predict the next-patch's distribution, we propose a TimeFlow Loss based on flow-matching, which facilitates native pre-training of Transformers on continuous-valued time series without discrete tokenization. Conditioned on arbitrary-length time series, our models are pre-trained without specifying any prior distribution and can generate multiple probable predictions, achieving more flexibility in representation learning than using parametric densities. Towards time series foundation models, we leverage minimal but crucial adaptations of Transformers and curate TimeBench with one trillion time points, comprising mostly real-world datasets and synthetic data. By mitigating mode collapse via TimeFlow Loss, we pre-train a family of Sundial models on TimeBench, which achieve unprecedented model capacity and generalization performance. In addition to excellent scalability, Sundial achieves state-of-the-art results on both point and probabilistic forecasting benchmarks with a just-in-time inference speed, i.e., making zero-shot predictions within a few milliseconds. We believe that Sundial's pioneering generative forecasting capability can improve model reliability in real-world decision-making. Code is available at: https://github.com/thuml/Sundial.", "authors": ["Yong Liu", "Guo Qin", "Zhiyuan Shi", "Zhi Chen", "Caiyin Yang", "Xiangdong Huang", "Jianmin Wang", "Mingsheng Long"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-02-02", "url": "https://arxiv.org/abs/2502.00816", "pdf_url": "https://arxiv.org/pdf/2502.00816v4", "arxiv_id": "2502.00816", "doi": "10.48550/arXiv.2502.00816", "citation_count": 140, "influential_citation_count": 21, "has_code": true, "code_url": "https://github.com/thuml/Sundial", "venue": "International Conference on Machine Learning", "quality_score": 0.6712}
{"id": "0f8f8b1dec0d593e60e8e9782f7dfbe7b6ef063ab286855cac518f25079f44ad", "sources": ["arxiv", "semantic_scholar"], "title": "Using Causality for Enhanced Prediction of Web Traffic Time Series", "abstract": "Predicting web service traffic has significant social value, as it can be applied to various practical scenarios, including but not limited to dynamic resource scaling, load balancing, system anomaly detection, service-level agreement compliance, and fraud detection. Web service traffic is characterized by frequent and drastic fluctuations over time and are influenced by heterogeneous web user behaviors, making accurate prediction a challenging task. Previous research has extensively explored statistical approaches, and neural networks to mine features from preceding service traffic time series for prediction. However, these methods have largely overlooked the causal relationships between services. Drawing inspiration from causality in ecological systems, we empirically recognize the causal relationships between web services. To leverage these relationships for improved web service traffic prediction, we propose an effective neural network module, CCMPlus, designed to extract causal relationship features across services. This module can be seamlessly integrated with existing time series models to consistently enhance the performance of web service traffic predictions. We theoretically justify that the causal correlation matrix generated by the CCMPlus module captures causal relationships among services. Empirical results on real-world datasets from Microsoft Azure, Alibaba Group, and Ant Group confirm that our method surpasses state-of-the-art approaches in Mean Squared Error (MSE) and Mean Absolute Error (MAE) for predicting service traffic time series. These findings highlight the efficacy of leveraging causal relationships for improved predictions.", "authors": ["Chang Tian", "Mingzhe Xing", "Zenglin Shi", "Matthew B. Blaschko", "Yinliang Yue", "Marie-Francine Moens"], "categories": ["cs.LG", "cs.NI"], "fields_of_study": ["Computer Science"], "published_date": "2025-02-02", "url": "https://arxiv.org/abs/2502.00612", "pdf_url": "https://arxiv.org/pdf/2502.00612v2", "arxiv_id": "2502.00612", "doi": "10.48550/arXiv.2502.00612", "citation_count": 4, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1747}
{"id": "b61064e5c6596e6bd7d11621ccd6ff37e003852b2c3e6b0b98bf904ad841ddef", "sources": ["arxiv", "semantic_scholar"], "title": "FreEformer: Frequency Enhanced Transformer for Multivariate Time Series Forecasting", "abstract": "This paper presents \\textbf{FreEformer}, a simple yet effective model that leverages a \\textbf{Fre}quency \\textbf{E}nhanced Trans\\textbf{former} for multivariate time series forecasting. Our work is based on the assumption that the frequency spectrum provides a global perspective on the composition of series across various frequencies and is highly suitable for robust representation learning. Specifically, we first convert time series into the complex frequency domain using the Discrete Fourier Transform (DFT). The Transformer architecture is then applied to the frequency spectra to capture cross-variate dependencies, with the real and imaginary parts processed independently. However, we observe that the vanilla attention matrix exhibits a low-rank characteristic, thus limiting representation diversity. This could be attributed to the inherent sparsity of the frequency domain and the strong-value-focused nature of Softmax in vanilla attention. To address this, we enhance the vanilla attention mechanism by introducing an additional learnable matrix to the original attention matrix, followed by row-wise L1 normalization. Theoretical analysis~demonstrates that this enhanced attention mechanism improves both feature diversity and gradient flow. Extensive experiments demonstrate that FreEformer consistently outperforms state-of-the-art models on eighteen real-world benchmarks covering electricity, traffic, weather, healthcare and finance. Notably, the enhanced attention mechanism also consistently improves the performance of state-of-the-art Transformer-based forecasters.", "authors": ["Wenzhen Yue", "Yong Liu", "Xianghua Ying", "Bowei Xing", "Ruohao Guo", "Ji Shi"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-01-23", "url": "https://arxiv.org/abs/2501.13989", "pdf_url": "https://arxiv.org/pdf/2501.13989v1", "arxiv_id": "2501.13989", "doi": "10.48550/arXiv.2501.13989", "citation_count": 33, "influential_citation_count": 2, "has_code": false, "code_url": null, "venue": "International Joint Conference on Artificial Intelligence", "quality_score": 0.3829}
{"id": "0c427fc4117e8172aa3395c999077a9e640bb8fb04e52c986aa605469f7e3e75", "sources": ["arxiv", "semantic_scholar"], "title": "Towards Lightweight Time Series Forecasting: a Patch-wise Transformer with Weak Data Enriching", "abstract": "Patch-wise Transformer based time series forecasting achieves superior accuracy. However, this superiority relies heavily on intricate model design with massive parameters, rendering both training and inference expensive, thus preventing their deployments on edge devices with limited resources and low latency requirements. In addition, existing methods often work in an autoregressive manner, which take into account only historical values, but ignore valuable, easy-to-obtain context information, such as weather forecasts, date and time of day. To contend with the two limitations, we propose LiPFormer, a novel Lightweight Patch-wise Transformer with weak data enriching. First, to simplify the Transformer backbone, LiPFormer employs a novel lightweight cross-patch attention and a linear transformation-based attention to eliminate Layer Normalization and Feed Forward Network, two heavy components in existing Transformers. Second, we propose a lightweight, weak data enriching module to provide additional, valuable weak supervision to the training. It enhances forecasting accuracy without significantly increasing model complexity as it does not involve expensive, human-labeling but using easily accessible context information. This facilitates the weak data enriching to plug-and-play on existing models. Extensive experiments on nine benchmark time series datasets demonstrate that LiPFormer outperforms state-of-the-art methods in accuracy, while significantly reducing parameter scale, training duration, and GPU memory usage. Deployment on an edge device reveals that LiPFormer takes only 1/3 inference time compared to classic Transformers. In addition, we demonstrate that the weak data enriching can integrate seamlessly into various Transformer based models to enhance their accuracy, suggesting its generality.", "authors": ["Meng Wang", "Jintao Yang", "Bin Yang", "Hui Li", "Tongxin Gong", "Bo Yang", "Jiangtao Cui"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-01-14", "url": "https://arxiv.org/abs/2501.10448", "pdf_url": "https://arxiv.org/pdf/2501.10448v1", "arxiv_id": "2501.10448", "doi": "10.1109/ICDE65448.2025.00100", "citation_count": 8, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "IEEE International Conference on Data Engineering", "quality_score": 0.2386}
{"id": "51997c7283356705c3e67c4d62c2f16af55da1b7638c908331524ffe1207342a", "sources": ["arxiv", "semantic_scholar"], "title": "Unveiling the Potential of Text in High-Dimensional Time Series Forecasting", "abstract": "Time series forecasting has traditionally focused on univariate and multivariate numerical data, often overlooking the benefits of incorporating multimodal information, particularly textual data. In this paper, we propose a novel framework that integrates time series models with Large Language Models to improve high-dimensional time series forecasting. Inspired by multimodal models, our method combines time series and textual data in the dual-tower structure. This fusion of information creates a comprehensive representation, which is then processed through a linear layer to generate the final forecast. Extensive experiments demonstrate that incorporating text enhances high-dimensional time series forecasting performance. This work paves the way for further research in multimodal time series forecasting.", "authors": ["Xin Zhou", "Weiqing Wang", "Shilin Qu", "Zhiqiang Zhang", "Christoph Bergmeir"], "categories": ["cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-01-13", "url": "https://arxiv.org/abs/2501.07048", "pdf_url": "https://arxiv.org/pdf/2501.07048v1", "arxiv_id": "2501.07048", "doi": "10.48550/arXiv.2501.07048", "citation_count": 4, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1747}
{"id": "891876b3d99592e13dfb017d1ee1e65ca2d6f9fae6d421c57a6cd20698face00", "sources": ["arxiv", "semantic_scholar"], "title": "From Tables to Time: Extending TabPFN-v2 to Time Series Forecasting", "abstract": "Recent progress in foundation models has enabled strong zero-shot performance for time series forecasting. In this work, we show that such capabilities can also emerge from tabular foundation models. We introduce TabPFN-TS, a simple method that treats forecasting as a tabular regression problem by combining lightweight temporal featurization with the pretrained TabPFN-v2. This formulation requires no time-series-specific pretraining and naturally supports both univariate and covariate-informed forecasting. Despite its compact size (11M parameters), TabPFN-TS achieves state-of-the-art performance on covariate-informed forecasting and competitive accuracy on univariate forecasting across the GIFT-Eval and fev-bench benchmarks. We further provide controlled analyses examining how the model interprets temporal structure, how featurization choices affect accuracy, and how forecasts change under alternative tabular backbones. Together, our results demonstrate that tabular foundation models--when paired with suitable temporal features--offer an efficient and versatile alternative for forecasting, bridging tabular and time-series learning within a unified framework. Code is available at https://github.com/PriorLabs/tabpfn-time-series.", "authors": ["Shi Bin Hoo", "Samuel Müller", "David Salinas", "Frank Hutter"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-01-06", "url": "https://arxiv.org/abs/2501.02945", "pdf_url": "https://arxiv.org/pdf/2501.02945v4", "arxiv_id": "2501.02945", "doi": null, "citation_count": 36, "influential_citation_count": 4, "has_code": true, "code_url": "https://github.com/PriorLabs/tabpfn-time-series", "venue": null, "quality_score": 0.3921}
{"id": "979847df6095c04a17f8890ef2f4de939da2c5c103aeaad4c50fe12a8b0e8b70", "sources": ["arxiv", "semantic_scholar"], "title": "Sequence Complementor: Complementing Transformers For Time Series Forecasting with Learnable Sequences", "abstract": "Since its introduction, the transformer has shifted the development trajectory away from traditional models (e.g., RNN, MLP) in time series forecasting, which is attributed to its ability to capture global dependencies within temporal tokens. Follow-up studies have largely involved altering the tokenization and self-attention modules to better adapt Transformers for addressing special challenges like non-stationarity, channel-wise dependency, and variable correlation in time series. However, we found that the expressive capability of sequence representation is a key factor influencing Transformer performance in time forecasting after investigating several representative methods, where there is an almost linear relationship between sequence representation entropy and mean square error, with more diverse representations performing better. In this paper, we propose a novel attention mechanism with Sequence Complementors and prove feasible from an information theory perspective, where these learnable sequences are able to provide complementary information beyond current input to feed attention. We further enhance the Sequence Complementors via a diversification loss that is theoretically covered. The empirical evaluation of both long-term and short-term forecasting has confirmed its superiority over the recent state-of-the-art methods.", "authors": ["Xiwen Chen", "Peijie Qiu", "Wenhui Zhu", "Huayu Li", "Hao Wang", "Aristeidis Sotiras", "Yalin Wang", "Abolfazl Razi"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-01-06", "url": "https://arxiv.org/abs/2501.02735", "pdf_url": "https://arxiv.org/pdf/2501.02735v1", "arxiv_id": "2501.02735", "doi": "10.48550/arXiv.2501.02735", "citation_count": 4, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "AAAI Conference on Artificial Intelligence", "quality_score": 0.1747}
{"id": "66e222d4e45bca7970959656872c6369238cfeef8859edeaf5c75e18d8820b12", "sources": ["arxiv", "semantic_scholar"], "title": "A Unified Hyperparameter Optimization Pipeline for Transformer-Based Time Series Forecasting Models", "abstract": "Transformer-based models for time series forecasting (TSF) have attracted significant attention in recent years due to their effectiveness and versatility. However, these models often require extensive hyperparameter optimization (HPO) to achieve the best possible performance, and a unified pipeline for HPO in transformer-based TSF remains lacking. In this paper, we present one such pipeline and conduct extensive experiments on several state-of-the-art (SOTA) transformer-based TSF models. These experiments are conducted on standard benchmark datasets to evaluate and compare the performance of different models, generating practical insights and examples. Our pipeline is generalizable beyond transformer-based architectures and can be applied to other SOTA models, such as Mamba and TimeMixer, as demonstrated in our experiments. The goal of this work is to provide valuable guidance to both industry practitioners and academic researchers in efficiently identifying optimal hyperparameters suited to their specific domain applications. The code and complete experimental results are available on GitHub.", "authors": ["Jingjing Xu", "Caesar Wu", "Yuan-Fang Li", "Grégoire Danoy", "Pascal Bouvry"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2025-01-02", "url": "https://arxiv.org/abs/2501.01394", "pdf_url": "https://arxiv.org/pdf/2501.01394v1", "arxiv_id": "2501.01394", "doi": "10.48550/arXiv.2501.01394", "citation_count": 2, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1193}
{"id": "ec363948984dedf028b42a73f07dbb0c918ca04a1e290893dd0777e6094bd7a9", "sources": ["arxiv", "semantic_scholar"], "title": "Evaluating Time Series Foundation Models on Noisy Periodic Time Series", "abstract": "While recent advancements in foundation models have significantly impacted machine learning, rigorous tests on the performance of time series foundation models (TSFMs) remain largely underexplored. This paper presents an empirical study evaluating the zero-shot, long-horizon forecasting abilities of several leading TSFMs over two synthetic datasets constituting noisy periodic time series. We assess model efficacy across different noise levels, underlying frequencies, and sampling rates. As benchmarks for comparison, we choose two statistical techniques: a Fourier transform (FFT)-based approach and a linear autoregressive (AR) model. Our findings demonstrate that while for time series with bounded periods and higher sampling rates, TSFMs can match or outperform the statistical approaches, their forecasting abilities deteriorate with longer periods, higher noise levels, lower sampling rates and more complex shapes of the time series.", "authors": ["Syamantak Datta Gupta"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2025-01-01", "url": "https://arxiv.org/abs/2501.00889", "pdf_url": "https://arxiv.org/pdf/2501.00889v2", "arxiv_id": "2501.00889", "doi": "10.48550/arXiv.2501.00889", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.0069}
{"id": "2c696af8126f07e04e8f5fec047385faa9fd4dea4731cc365782982dae72850a", "sources": ["arxiv", "semantic_scholar"], "title": "TimeRAF: Retrieval-Augmented Foundation model for Zero-shot Time Series Forecasting", "abstract": "Time series forecasting plays a crucial role in data mining, driving rapid advancements across numerous industries. With the emergence of large models, time series foundation models (TSFMs) have exhibited remarkable generalization capabilities, such as zero-shot learning, through large-scale pre-training. Meanwhile, Retrieval-Augmented Generation (RAG) methods have been widely employed to enhance the performance of foundation models on unseen data, allowing models to access to external knowledge. In this paper, we introduce TimeRAF, a Retrieval-Augmented Forecasting model that enhance zero-shot time series forecasting through retrieval-augmented techniques. We develop customized time series knowledge bases that are tailored to the specific forecasting tasks. TimeRAF employs an end-to-end learnable retriever to extract valuable information from the knowledge base. Additionally, we propose Channel Prompting for knowledge integration, which effectively extracts relevant information from the retrieved knowledge along the channel dimension. Extensive experiments demonstrate the effectiveness of our model, showing significant improvement across various domains and datasets.", "authors": ["Huanyu Zhang", "Chang Xu", "Yi-Fan Zhang", "Zhang Zhang", "Liang Wang", "Jiang Bian", "Tieniu Tan"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-12-30", "url": "https://arxiv.org/abs/2412.20810", "pdf_url": "https://arxiv.org/pdf/2412.20810v1", "arxiv_id": "2412.20810", "doi": "10.48550/arXiv.2412.20810", "citation_count": 18, "influential_citation_count": 2, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.3197}
{"id": "b74e41185b24e2f12bbc8216c84a18eaf9d81bfc775b19fee77b7a35d4e6c430", "sources": ["arxiv", "semantic_scholar"], "title": "Urban Water Consumption Forecasting Using Deep Learning and Correlated District Metered Areas", "abstract": "Accurate water consumption forecasting is a crucial tool for water utilities and policymakers, as it helps ensure a reliable supply, optimize operations, and support infrastructure planning. Urban Water Distribution Networks (WDNs) are divided into District Metered Areas (DMAs), where water flow is monitored to efficiently manage resources. This work focuses on short-term forecasting of DMA consumption using deep learning and aims to address two key challenging issues. First, forecasting based solely on a DMA's historical data may lack broader context and provide limited insights. Second, DMAs may experience sensor malfunctions providing incorrect data, or some DMAs may not be monitored at all due to computational costs, complicating accurate forecasting. We propose a novel method that first identifies DMAs with correlated consumption patterns and then uses these patterns, along with the DMA's local data, as input to a deep learning model for forecasting. In a real-world study with data from five DMAs, we show that: i) the deep learning model outperforms a classical statistical model; ii) accurate forecasting can be carried out using only correlated DMAs' consumption patterns; and iii) even when a DMA's local data is available, including correlated DMAs' data improves accuracy.", "authors": ["Kleanthis Malialis", "Nefeli Mavri", "Stelios G. Vrachimis", "Marios S. Kyriakou", "Demetrios G. Eliades", "Marios M. Polycarpou"], "categories": ["cs.LG", "cs.CY"], "fields_of_study": ["Computer Science"], "published_date": "2024-12-30", "url": "https://arxiv.org/abs/2501.00158", "pdf_url": "https://arxiv.org/pdf/2501.00158v1", "arxiv_id": "2501.00158", "doi": "10.1109/CIETESCompanion65203.2025.11003284", "citation_count": 3, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "2025 IEEE Symposia on Computational Intelligence for Energy, Transport and Environmental Sustainability", "quality_score": 0.1505}
{"id": "8c00115af663661ad3522c6f83144bf4c8e86796a4488ce8cc37e4a0c33da9dc", "sources": ["arxiv", "semantic_scholar"], "title": "Forecasting Malaria in Indian States: A Time Series Approach with R Shiny Integration", "abstract": "Malaria remains a significant public health challenge in many regions, necessitating robust predictive models to aid in its management and prevention. This study focuses on developing and evaluating time series models for forecasting malaria cases across eight Indian states: Jharkhand, Chhattisgarh, Maharashtra, Meghalaya, Mizoram, Odisha, Tripura, and Uttar Pradesh. We employed various modeling approaches, including polynomial regression with seasonal components, log-transformed polynomial regression, lagged difference models, and ARIMA models, to capture the temporal dynamics of malaria incidence. Comprehensive model fitting, residual analysis, and performance evaluation using metrics such as Root Mean Squared Error (RMSE) and Mean Absolute Percentage Error (MAPE) indicated that the log-transformed polynomial regression model consistently outperformed other models in terms of accuracy and robustness across all states. Rolling forecast validation further confirmed the superior predictive capability of the log-transformed model over time. Additionally, an interactive R Shiny tool was developed to facilitate the use of these predictive models by researchers and public health officials. This tool allows users to input data, select modeling approaches, and visualize predictions and performance metrics, providing a practical tool for real-time malaria forecasting and decision-making support. Our findings highlight the critical role of appropriate modeling techniques in malaria prediction and offer valuable resources for enhancing malaria surveillance and response efforts.", "authors": ["Sujit K. Ghosh", "Usha Ananthakumar", "Praveen D. Chougale", "Adithya B. Somaraj"], "categories": ["stat.AP"], "fields_of_study": ["Mathematics", "Medicine"], "published_date": "2024-12-28", "url": "https://arxiv.org/abs/2412.20121", "pdf_url": "https://arxiv.org/pdf/2412.20121v2", "arxiv_id": "2412.20121", "doi": "10.1186/s12936-025-05526-z", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "Malaria Journal", "quality_score": 0.0753}
{"id": "7f6f24317e9ffaa4b98576d0dc85a07041ba5fc022c4eacc6776daf168f12c6b", "sources": ["arxiv", "semantic_scholar"], "title": "Ister: Linear Transformer for Efficient Multivariate Time Series Forecasting", "abstract": "Transformer-based models have achieved remarkable success in multivariate time series forecasting (MTSF) by capturing long-range dependencies. However, their widespread adoption is hindered by the quadratic computational complexity of self-attention, which limits scalability on high-dimensional sequences. To address this challenge, we propose the Inverted Seasonal-Trend Decomposition Transformer (Ister), a novel architecture that enhances both predictive accuracy and computational efficiency. Central to Ister is Dot-attention, a linear-complexity attention mechanism that replaces conventional multi-head self-attention with element-wise dot-product operations to model inter-series dependencies. Furthermore, we introduce an inverted seasonal-trend decomposition strategy that isolates periodic components, enabling the model to focus learning on periodic patterns, thereby improving the performance of channel alignment. Extensive experiments across several real-world benchmarks demonstrate that Ister consistently achieves state-of-the-art performance. Code is available at https://github.com/macovaseas/Ister.", "authors": ["Fanpu Cao", "Shu Yang", "Zhengjian Chen", "Ye Liu", "Laizhong Cui"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-12-25", "url": "https://arxiv.org/abs/2412.18798", "pdf_url": "https://arxiv.org/pdf/2412.18798v3", "arxiv_id": "2412.18798", "doi": "10.1109/ICASSP55912.2026.11463971", "citation_count": 0, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/macovaseas/Ister", "venue": "IEEE International Conference on Acoustics, Speech, and Signal Processing", "quality_score": 0.0}
{"id": "91c16c5d7b0b6843d1da8a28921a74d2669ca2008b1f67b6a1a8c132ed17d439", "sources": ["arxiv", "semantic_scholar"], "title": "Enabling Time-series Foundation Model for Building Energy Forecasting via Contrastive Curriculum Learning", "abstract": "Advances in time-series forecasting are driving a shift from conventional machine learning models to foundation models (FMs) that are trained with generalized knowledge. However, existing FMs still perform poorly in the energy fields, such as building energy forecasting (BEF). This paper studies the adaptation of FM to BEF tasks. We demonstrate the shortcomings of fine-tuning FM straightforwardly from both the perspectives of FM and the data. To overcome these limitations, we propose a new \\textit{contrastive curriculum learning}-based training method. Our method optimizes the ordering of training data in the context of TSFM adaptation. Experiments show that our method can improve the zero/few-shot performance by 14.6\\% compared to the existing FMs. Our code and new TSFM will be available at <Anonymous Github Repo>.", "authors": ["Rui Liang", "Yang Deng", "Donghua Xie", "Fang He", "Dan Wang"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-12-23", "url": "https://arxiv.org/abs/2412.17285", "pdf_url": "https://arxiv.org/pdf/2412.17285v1", "arxiv_id": "2412.17285", "doi": "10.48550/arXiv.2412.17285", "citation_count": 2, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1505}
{"id": "06a0d5a7b580e08e2e175ed3a7a643176f15773eab9ac82a0cf45f4d9df6d18f", "sources": ["arxiv", "semantic_scholar"], "title": "EasyTime: Time Series Forecasting Made Easy", "abstract": "Time series forecasting has important applications across diverse domains. EasyTime, the system we demonstrate, facilitates easy use of time-series forecasting methods by researchers and practitioners alike. First, EasyTime enables one-click evaluation, enabling researchers to evaluate new forecasting methods using the suite of diverse time series datasets collected in the preexisting time series forecasting benchmark (TFB). This is achieved by leveraging TFB's flexible and consistent evaluation pipeline. Second, when practitioners must perform forecasting on a new dataset, a nontrivial first step is often to find an appropriate forecasting method. EasyTime provides an Automated Ensemble module that combines the promising forecasting methods to yield superior forecasting accuracy compared to individual methods. Third, EasyTime offers a natural language Q&A module leveraging large language models. Given a question like \"Which method is best for long term forecasting on time series with strong seasonality?\", EasyTime converts the question into SQL queries on the database of results obtained by TFB and then returns an answer in natural language and charts. By demonstrating EasyTime, we intend to show how it is possible to simplify the use of time series forecasting and to offer better support for the development of new generations of time series forecasting methods.", "authors": ["Xiangfei Qiu", "Xiuwen Li", "Ruiyang Pang", "Zhicheng Pan", "Xingjian Wu", "Liu Yang", "Jilin Hu", "Yang Shu", "Xuesong Lu", "Chengcheng Yang", "Chenjuan Guo", "Aoying Zhou", "Christian S. Jensen", "Bin Yang"], "categories": ["cs.LG", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2024-12-23", "url": "https://arxiv.org/abs/2412.17603", "pdf_url": "https://arxiv.org/pdf/2412.17603v1", "arxiv_id": "2412.17603", "doi": "10.1109/ICDE65448.2025.00353", "citation_count": 40, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "IEEE International Conference on Data Engineering", "quality_score": 0.4032}
{"id": "a460ad09a6a9484935b7869d08842e256c878ee1743de7d97ca6bb9267bb33c8", "sources": ["arxiv", "semantic_scholar"], "title": "WPMixer: Efficient Multi-Resolution Mixing for Long-Term Time Series Forecasting", "abstract": "Time series forecasting is crucial for various applications, such as weather forecasting, power load forecasting, and financial analysis. In recent studies, MLP-mixer models for time series forecasting have been shown as a promising alternative to transformer-based models. However, the performance of these models is still yet to reach its potential. In this paper, we propose Wavelet Patch Mixer (WPMixer), a novel MLP-based model, for long-term time series forecasting, which leverages the benefits of patching, multi-resolution wavelet decomposition, and mixing. Our model is based on three key components: (i) multi-resolution wavelet decomposition, (ii) patching and embedding, and (iii) MLP mixing. Multi-resolution wavelet decomposition efficiently extracts information in both the frequency and time domains. Patching allows the model to capture an extended history with a look-back window and enhances capturing local information while MLP mixing incorporates global information. Our model significantly outperforms state-of-the-art MLP-based and transformer-based models for long-term time series forecasting in a computationally efficient way, demonstrating its efficacy and potential for practical applications.", "authors": ["Md Mahmuddun Nabi Murad", "Mehmet Aktukmak", "Yasin Yilmaz"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-12-22", "url": "https://arxiv.org/abs/2412.17176", "pdf_url": "https://arxiv.org/pdf/2412.17176v1", "arxiv_id": "2412.17176", "doi": "10.48550/arXiv.2412.17176", "citation_count": 54, "influential_citation_count": 7, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.4515}
{"id": "e6e2cada74af575b3ce78db9b3d2ead0ba8af24b6b0b19456f3258127a45a682", "sources": ["arxiv", "semantic_scholar"], "title": "Zero Shot Time Series Forecasting Using Kolmogorov Arnold Networks", "abstract": "Accurate energy price forecasting is crucial for participants in day-ahead energy markets, as it significantly influences their decision-making processes. While machine learning-based approaches have shown promise in enhancing these forecasts, they often remain confined to the specific markets on which they are trained, thereby limiting their adaptability to new or unseen markets. In this paper, we introduce a cross-domain adaptation model designed to forecast energy prices by learning market-invariant representations across different markets during the training phase. We propose a doubly residual N-BEATS network with Kolmogorov Arnold networks at its core for time series forecasting. These networks, grounded in the Kolmogorov-Arnold representation theorem, offer a powerful way to approximate multivariate continuous functions. The cross domain adaptation model was generated with an adversarial framework. The model's effectiveness was tested in predicting day-ahead electricity prices in a zero shot fashion. In comparison with baseline models, our proposed framework shows promising results. By leveraging the Kolmogorov-Arnold networks, our model can potentially enhance its ability to capture complex patterns in energy price data, thus improving forecast accuracy across diverse market conditions. This addition not only enriches the model's representational capacity but also contributes to a more robust and flexible forecasting tool adaptable to various energy markets.", "authors": ["Abhiroop Bhattacharya", "Nandinee Haq"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-12-19", "url": "https://arxiv.org/abs/2412.17853", "pdf_url": "https://arxiv.org/pdf/2412.17853v2", "arxiv_id": "2412.17853", "doi": "10.48550/arXiv.2412.17853", "citation_count": 2, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1193}
{"id": "9da9c57952547c0daa0d1916f2747e013f6d2dc0b5932fafc8621f6bd46e53b2", "sources": ["arxiv", "semantic_scholar"], "title": "A Comparative Study of Pruning Methods in Transformer-based Time Series Forecasting", "abstract": "The current landscape in time-series forecasting is dominated by Transformer-based models. Their high parameter count and corresponding demand in computational resources pose a challenge to real-world deployment, especially for commercial and scientific applications with low-power embedded devices. Pruning is an established approach to reduce neural network parameter count and save compute. However, the implications and benefits of pruning Transformer-based models for time series forecasting are largely unknown. To close this gap, we provide a comparative benchmark study by evaluating unstructured and structured pruning on various state-of-the-art multivariate time series models. We study the effects of these pruning strategies on model predictive performance and computational aspects like model size, operations, and inference time. Our results show that certain models can be pruned even up to high sparsity levels, outperforming their dense counterpart. However, fine-tuning pruned models is necessary. Furthermore, we demonstrate that even with corresponding hardware and software support, structured pruning is unable to provide significant time savings.", "authors": ["Nicholas Kiefer", "Arvid Weyrauch", "Muhammed Öz", "Achim Streit", "Markus Götz", "Charlotte Debus"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-12-17", "url": "https://arxiv.org/abs/2412.12883", "pdf_url": "https://arxiv.org/pdf/2412.12883v1", "arxiv_id": "2412.12883", "doi": "10.1109/ICDMW69685.2025.00032", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.0753}
{"id": "67db86d3272f369e154c8d4024b330daa12cdee3c3c7da71e11bec1df3264bd3", "sources": ["arxiv", "semantic_scholar"], "title": "Federated Foundation Models on Heterogeneous Time Series", "abstract": "Training a general-purpose time series foundation models with robust generalization capabilities across diverse applications from scratch is still an open challenge. Efforts are primarily focused on fusing cross-domain time series datasets to extract shared subsequences as tokens for training models on Transformer architecture. However, due to significant statistical heterogeneity across domains, this cross-domain fusing approach doesn't work effectively as the same as fusing texts and images. To tackle this challenge, this paper proposes a novel federated learning approach to address the heterogeneity in time series foundation models training, namely FFTS. Specifically, each data-holding organization is treated as an independent client in a collaborative learning framework with federated settings, and then many client-specific local models will be trained to preserve the unique characteristics per dataset. Moreover, a new regularization mechanism will be applied to both client-side and server-side, thus to align the shared knowledge across heterogeneous datasets from different domains. Extensive experiments on benchmark datasets demonstrate the effectiveness of the proposed federated learning approach. The newly learned time series foundation models achieve superior generalization capabilities on cross-domain time series analysis tasks, including forecasting, imputation, and anomaly detection.", "authors": ["Shengchao Chen", "Guodong Long", "Jing Jiang", "Chengqi Zhang"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-12-12", "url": "https://arxiv.org/abs/2412.08906", "pdf_url": "https://arxiv.org/pdf/2412.08906v1", "arxiv_id": "2412.08906", "doi": "10.48550/arXiv.2412.08906", "citation_count": 23, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.3451}
{"id": "898ce88dde90bf25f1e7684125cf07bef164af1a54cd1c450596ac81a1c20143", "sources": ["arxiv", "semantic_scholar"], "title": "Auto-Regressive Moving Diffusion Models for Time Series Forecasting", "abstract": "Time series forecasting (TSF) is essential in various domains, and recent advancements in diffusion-based TSF models have shown considerable promise. However, these models typically adopt traditional diffusion patterns, treating TSF as a noise-based conditional generation task. This approach neglects the inherent continuous sequential nature of time series, leading to a fundamental misalignment between diffusion mechanisms and the TSF objective, thereby severely impairing performance. To bridge this misalignment, and inspired by the classic Auto-Regressive Moving Average (ARMA) theory, which views time series as continuous sequential progressions evolving from previous data points, we propose a novel Auto-Regressive Moving Diffusion (ARMD) model to first achieve the continuous sequential diffusion-based TSF. Unlike previous methods that start from white Gaussian noise, our model employs chain-based diffusion with priors, accurately modeling the evolution of time series and leveraging intermediate state information to improve forecasting accuracy and stability. Specifically, our approach reinterprets the diffusion process by considering future series as the initial state and historical series as the final state, with intermediate series generated using a sliding-based technique during the forward process. This design aligns the diffusion model's sampling procedure with the forecasting objective, resulting in an unconditional, continuous sequential diffusion TSF model. Extensive experiments conducted on seven widely used datasets demonstrate that our model achieves state-of-the-art performance, significantly outperforming existing diffusion-based TSF models. Our code is available on GitHub: https://github.com/daxin007/ARMD.", "authors": ["Jiaxin Gao", "Qinglong Cao", "Yuntian Chen"], "categories": ["cs.LG", "cs.AI", "stat.AP"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2024-12-12", "url": "https://arxiv.org/abs/2412.09328", "pdf_url": "https://arxiv.org/pdf/2412.09328v1", "arxiv_id": "2412.09328", "doi": "10.48550/arXiv.2412.09328", "citation_count": 16, "influential_citation_count": 1, "has_code": true, "code_url": "https://github.com/daxin007/ARMD", "venue": "arXiv.org", "quality_score": 0.3076}
{"id": "cb40b8ae8dccf52a690ff71d00e56cf2487b66743fb23e7f443a24377786208c", "sources": ["arxiv", "semantic_scholar"], "title": "Measuring Pre-training Data Quality without Labels for Time Series Foundation Models", "abstract": "Recently, there has been a growing interest in time series foundation models that generalize across different downstream tasks. A key to strong foundation models is a diverse pre-training dataset, which is particularly challenging to collect for time series classification. In this work, we explore the performance of a contrastive-learning-based foundation model as a function of the data used for pre-training. We introduce contrastive accuracy, a new measure to evaluate the quality of the representation space learned by the foundation model. Our experiments reveal the positive correlation between the proposed measure and the accuracy of the model on a collection of downstream tasks. This suggests that the contrastive accuracy can serve as a criterion to search for time series datasets that can enhance the pre-training and improve thereby the foundation model's generalization.", "authors": ["Songkang Wen", "Vasilii Feofanov", "Jianfeng Zhang"], "categories": ["cs.LG", "cs.AI", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2024-12-09", "url": "https://arxiv.org/abs/2412.06368", "pdf_url": "https://arxiv.org/pdf/2412.06368v1", "arxiv_id": "2412.06368", "doi": "10.48550/arXiv.2412.06368", "citation_count": 2, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1193}
{"id": "49b0ffbea9105815f25b80d208f46008a8fcc2d9df6425ed10c515aaf711a04f", "sources": ["arxiv", "semantic_scholar"], "title": "Enhancing Foundation Models for Time Series Forecasting via Wavelet-based Tokenization", "abstract": "How to best develop foundational models for time series forecasting remains an important open question. Tokenization is a crucial consideration in this effort: what is an effective discrete vocabulary for a real-valued sequential input? To address this question, we develop WaveToken, a wavelet-based tokenizer that allows models to learn complex representations directly in the space of time-localized frequencies. Our method first scales and decomposes the input time series, then thresholds and quantizes the wavelet coefficients, and finally pre-trains an autoregressive model to forecast coefficients for the forecast horizon. By decomposing coarse and fine structures in the inputs, wavelets provide an eloquent and compact language for time series forecasting that simplifies learning. Empirical results on a comprehensive benchmark, including 42 datasets for both in-domain and zero-shot settings, show that WaveToken: i) provides better accuracy than recently proposed foundation models for forecasting while using a much smaller vocabulary (1024 tokens), and performs on par or better than modern deep learning models trained specifically on each dataset; and ii) exhibits superior generalization capabilities, achieving the best average rank across all datasets for three complementary metrics. In addition, we show that our method can easily capture complex temporal patterns of practical relevance that are challenging for other recent pre-trained models, including trends, sparse spikes, and non-stationary time series with varying frequencies evolving over time.", "authors": ["Luca Masserano", "Abdul Fatir Ansari", "Boran Han", "Xiyuan Zhang", "Christos Faloutsos", "Michael W. Mahoney", "Andrew Gordon Wilson", "Youngsuk Park", "Syama Rangapuram", "Danielle C. Maddix", "Yuyang Wang"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-12-06", "url": "https://arxiv.org/abs/2412.05244", "pdf_url": "https://arxiv.org/pdf/2412.05244v1", "arxiv_id": "2412.05244", "doi": "10.48550/arXiv.2412.05244", "citation_count": 28, "influential_citation_count": 2, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.3656}
{"id": "2346e324a0b781612dfe2cf48e28f0aa4bb8e5f1620beb0d597c28ae13112d73", "sources": ["arxiv", "semantic_scholar"], "title": "A Foundation Model for Wearable Movement Data in Mental Health Research", "abstract": "Wearable movement data is collected by nearly all commercially available smartwatches and is a valuable resource for mental health research, reflecting fine-grained temporal behavioral trends. Despite its promise, the development of foundation models for health wearable modeling remains limited when compared to clinical image and text analysis. We designed transformers with patch embeddings and used self-supervised masked autoencoder pretraining on minute-level week-long actigraphy (physical activity intensity measurement) sequences to develop and evaluate the Pretrained Actigraphy Transformer (PAT). PAT is an open-source foundation model for wearable movement time series that combines week-long temporal modeling, psychiatric outcome evaluation, and reproducibility on public data. Pretrained on data from 21,538 U.S. participants in a nationally representative cohort from the National Health and Nutrition Examination Survey (NHANES), PAT consistently outperformed non-foundation-model baselines across mental health prediction tasks-including benzodiazepine and SSRI use, depression, and sleep abnormalities. During the benzodiazepine medication usage prediction task, PAT demonstrated the largest improvement over non-foundational deep learning models commonly used for time-series modeling (i.e., 55.6% improvement over the LSTM, 21.4% improvement over the 1-D CNN, 14.8% improvement over the ConvLSTM). Beyond predictive accuracy, PAT provides interpretable attention maps highlighting specific periods of daily activity most important for clinical predictions, offering model transparency and potential clinical insights. The results suggest that PAT offers an easy-to-deploy, adaptable and scalable solution to advance clinical insight from wearable sensor data for researchers and clinicians. GitHub: https://github.com/njacobsonlab/Pretrained-Actigraphy-Transformer/", "authors": ["Franklin Y. Ruan", "Aiwei Zhang", "Jenny Y. Oh", "SouYoung Jin", "Nicholas C. Jacobson"], "categories": ["cs.LG", "cs.AI", "cs.HC", "q-bio.QM"], "fields_of_study": ["Medicine", "Computer Science", "Biology"], "published_date": "2024-11-22", "url": "https://arxiv.org/abs/2411.15240", "pdf_url": "https://arxiv.org/pdf/2411.15240v5", "arxiv_id": "2411.15240", "doi": "10.1109/JBHI.2026.3694809", "citation_count": 2, "influential_citation_count": 1, "has_code": true, "code_url": "https://github.com/njacobsonlab/Pretrained-Actigraphy-Transformer/", "venue": "IEEE journal of biomedical and health informatics", "quality_score": 0.1505}
{"id": "de9da31d52f44b168eb9012965b0aa93f9c439d0f1f719cc0881127f528e8c54", "sources": ["arxiv", "semantic_scholar"], "title": "From RNNs to Foundation Models: An Empirical Study on Commercial Building Energy Consumption", "abstract": "Accurate short-term energy consumption forecasting for commercial buildings is crucial for smart grid operations. While smart meters and deep learning models enable forecasting using past data from multiple buildings, data heterogeneity from diverse buildings can reduce model performance. The impact of increasing dataset heterogeneity in time series forecasting, while keeping size and model constant, is understudied. We tackle this issue using the ComStock dataset, which provides synthetic energy consumption data for U.S. commercial buildings. Two curated subsets, identical in size and region but differing in building type diversity, are used to assess the performance of various time series forecasting models, including fine-tuned open-source foundation models (FMs). The results show that dataset heterogeneity and model architecture have a greater impact on post-training forecasting performance than the parameter count. Moreover, despite the higher computational cost, fine-tuned FMs demonstrate competitive performance compared to base models trained from scratch.", "authors": ["Shourya Bose", "Yijiang Li", "Amy Van Sant", "Yu Zhang", "Kibaek Kim"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-11-21", "url": "https://arxiv.org/abs/2411.14421", "pdf_url": "https://arxiv.org/pdf/2411.14421v2", "arxiv_id": "2411.14421", "doi": "10.48550/arXiv.2411.14421", "citation_count": 1, "influential_citation_count": 0, "has_code": true, "code_url": null, "venue": "arXiv.org", "quality_score": 0.0753}
{"id": "92f8634bd9930a7ec62542eadca5f64e0503fb489f38f1e809b2bcb23811380c", "sources": ["arxiv", "semantic_scholar"], "title": "Generalized Prompt Tuning: Adapting Frozen Univariate Time Series Foundation Models for Multivariate Healthcare Time Series", "abstract": "Time series foundation models are pre-trained on large datasets and are able to achieve state-of-the-art performance in diverse tasks. However, to date, there has been limited work demonstrating how well these models perform in medical applications, where labeled data can be scarce. Further, we observe that currently, the majority of time series foundation models either are univariate in nature, or assume channel independence, meaning that they handle multivariate time series but do not model how the different variables relate. In this paper, we propose a prompt-tuning-inspired fine-tuning technique, Generalized Prompt Tuning (Gen-P-Tuning), that enables us to adapt an existing univariate time series foundation model (treated as frozen) to handle multivariate time series prediction. Our approach provides a way to combine information across channels (variables) of multivariate time series. We demonstrate the effectiveness of our fine-tuning approach against various baselines on two MIMIC classification tasks, and on influenza-like illness forecasting.", "authors": ["Mingzhu Liu", "Angela H. Chen", "George H. Chen"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-11-19", "url": "https://arxiv.org/abs/2411.12824", "pdf_url": "https://arxiv.org/pdf/2411.12824v1", "arxiv_id": "2411.12824", "doi": "10.48550/arXiv.2411.12824", "citation_count": 4, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.1747}
{"id": "b7af301b90e75cf5c9abce6c2219c36c0d8f1b3cdcc914334a14266cc10823b2", "sources": ["arxiv", "semantic_scholar"], "title": "Approximate Probabilistic Inference for Time-Series Data A Robust Latent Gaussian Model With Temporal Awareness", "abstract": "The development of robust generative models for highly varied non-stationary time series data is a complex yet important problem. Traditional models for time series data prediction, such as Long Short-Term Memory (LSTM), are inefficient and generalize poorly as they cannot capture complex temporal relationships. In this paper, we present a probabilistic generative model that can be trained to capture temporal information, and that is robust to data errors. We call it Time Deep Latent Gaussian Model (tDLGM). Its novel architecture is inspired by Deep Latent Gaussian Model (DLGM). Our model is trained to minimize a loss function based on the negative log loss. One contributing factor to Time Deep Latent Gaussian Model (tDLGM) robustness is our regularizer, which accounts for data trends. Experiments conducted show that tDLGM is able to reconstruct and generate complex time series data, and that it is robust against to noise and faulty data.", "authors": ["Anton Johansson", "Arunselvan Ramaswamy"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-11-14", "url": "https://arxiv.org/abs/2411.09312", "pdf_url": "https://arxiv.org/pdf/2411.09312v2", "arxiv_id": "2411.09312", "doi": "10.48550/arXiv.2411.09312", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "International Conference on Agents and Artificial Intelligence", "quality_score": 0.0}
{"id": "ce47657d551040c6e7417713157c6eadc4e9895d6c1758adb3b3f92be2d0c379", "sources": ["arxiv", "semantic_scholar"], "title": "Retrieval Augmented Time Series Forecasting", "abstract": "Retrieval-augmented generation (RAG) is a central component of modern LLM systems, particularly in scenarios where up-to-date information is crucial for accurately responding to user queries or when queries exceed the scope of the training data. The advent of time-series foundation models (TSFM), such as Chronos, and the need for effective zero-shot forecasting performance across various time-series domains motivates the question: Do benefits of RAG similarly carry over to time series forecasting? In this paper, we advocate that the dynamic and event-driven nature of time-series data makes RAG a crucial component of TSFMs and introduce a principled RAG framework for time-series forecasting, called Retrieval Augmented Forecasting (RAF). Within RAF, we develop efficient strategies for retrieving related time-series examples and incorporating them into forecast. Through experiments and mechanistic studies, we demonstrate that RAF indeed improves the forecasting accuracy across diverse time series domains and the improvement is more significant for larger TSFM sizes.", "authors": ["Kutay Tire", "Ege Onur Taga", "Muhammed Emrullah Ildiz", "Samet Oymak"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-11-12", "url": "https://arxiv.org/abs/2411.08249", "pdf_url": "https://arxiv.org/pdf/2411.08249v2", "arxiv_id": "2411.08249", "doi": "10.48550/arXiv.2411.08249", "citation_count": 39, "influential_citation_count": 7, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.4515}
{"id": "4a3ec255a9bce206f13e97fe58871402bdb038a078869919256f4cf5cfc9b99d", "sources": ["arxiv", "semantic_scholar"], "title": "A Mamba Foundation Model for Time Series Forecasting", "abstract": "Time series foundation models have demonstrated strong performance in zero-shot learning, making them well-suited for predicting rapidly evolving patterns in real-world applications where relevant training data are scarce. However, most of these models rely on the Transformer architecture, which incurs quadratic complexity as input length increases. To address this, we introduce TSMamba, a linear-complexity foundation model for time series forecasting built on the Mamba architecture. The model captures temporal dependencies through both forward and backward Mamba encoders, achieving high prediction accuracy. To reduce reliance on large datasets and lower training costs, TSMamba employs a two-stage transfer learning process that leverages pretrained Mamba LLMs, allowing effective time series modeling with a moderate training set. In the first stage, the forward and backward backbones are optimized via patch-wise autoregressive prediction; in the second stage, the model trains a prediction head and refines other components for long-term forecasting. While the backbone assumes channel independence to manage varying channel numbers across datasets, a channel-wise compressed attention module is introduced to capture cross-channel dependencies during fine-tuning on specific multivariate datasets. Experiments show that TSMamba's zero-shot performance is comparable to state-of-the-art time series foundation models, despite using significantly less training data. It also achieves competitive or superior full-shot performance compared to task-specific prediction models. The code will be made publicly available.", "authors": ["Haoyu Ma", "Yushu Chen", "Wenlai Zhao", "Jinzhe Yang", "Yingsheng Ji", "Xinghua Xu", "Xiaozhu Liu", "Hao Jing", "Shengzhuo Liu", "Guangwen Yang"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-11-05", "url": "https://arxiv.org/abs/2411.02941", "pdf_url": "https://arxiv.org/pdf/2411.02941v1", "arxiv_id": "2411.02941", "doi": "10.48550/arXiv.2411.02941", "citation_count": 18, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.3197}
{"id": "4c8f7f0ec90b4ebfeb9be52f58dd9687df15411991ac25d0108b86ebdfa36eb0", "sources": ["arxiv", "semantic_scholar"], "title": "ElasTST: Towards Robust Varied-Horizon Forecasting with Elastic Time-Series Transformer", "abstract": "Numerous industrial sectors necessitate models capable of providing robust forecasts across various horizons. Despite the recent strides in crafting specific architectures for time-series forecasting and developing pre-trained universal models, a comprehensive examination of their capability in accommodating varied-horizon forecasting during inference is still lacking. This paper bridges this gap through the design and evaluation of the Elastic Time-Series Transformer (ElasTST). The ElasTST model incorporates a non-autoregressive design with placeholders and structured self-attention masks, warranting future outputs that are invariant to adjustments in inference horizons. A tunable version of rotary position embedding is also integrated into ElasTST to capture time-series-specific periods and enhance adaptability to different horizons. Additionally, ElasTST employs a multi-scale patch design, effectively integrating both fine-grained and coarse-grained information. During the training phase, ElasTST uses a horizon reweighting strategy that approximates the effect of random sampling across multiple horizons with a single fixed horizon setting. Through comprehensive experiments and comparisons with state-of-the-art time-series architectures and contemporary foundation models, we demonstrate the efficacy of ElasTST's unique design elements. Our findings position ElasTST as a robust solution for the practical necessity of varied-horizon forecasting.", "authors": ["Jiawen Zhang", "Shun Zheng", "Xumeng Wen", "Xiaofang Zhou", "Jiang Bian", "Jia Li"], "categories": ["cs.LG", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2024-11-04", "url": "https://arxiv.org/abs/2411.01842", "pdf_url": "https://arxiv.org/pdf/2411.01842v1", "arxiv_id": "2411.01842", "doi": "10.48550/arXiv.2411.01842", "citation_count": 13, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "Neural Information Processing Systems", "quality_score": 0.2865}
{"id": "039268936bcda4144bad2287ec0aa1fc738a1973af1343df34f76a2b4aa6cf92", "sources": ["arxiv", "semantic_scholar"], "title": "PSformer: Parameter-efficient Transformer with Segment Attention for Time Series Forecasting", "abstract": "Time series forecasting remains a critical challenge across various domains, often complicated by high-dimensional data and long-term dependencies. This paper presents a novel transformer architecture for time series forecasting, incorporating two key innovations: parameter sharing (PS) and Spatial-Temporal Segment Attention (SegAtt). We also define the time series segment as the concatenation of sequence patches from the same positions across different variables. The proposed model, PSformer, reduces the number of training parameters through the parameter sharing mechanism, thereby improving model efficiency and scalability. The introduction of SegAtt could enhance the capability of capturing local spatio-temporal dependencies by computing attention over the segments, and improve global representation by integrating information across segments. The combination of parameter sharing and SegAtt significantly improves the forecasting performance. Extensive experiments on benchmark datasets demonstrate that PSformer outperforms popular baselines and other transformer-based approaches in terms of accuracy and scalability, establishing itself as an accurate and scalable tool for time series forecasting.", "authors": ["Yanlong Wang", "Jian Xu", "Fei Ma", "Shao-Lun Huang", "Danny Dongning Sun", "Xiao-Ping Zhang"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-11-03", "url": "https://arxiv.org/abs/2411.01419", "pdf_url": "https://arxiv.org/pdf/2411.01419v2", "arxiv_id": "2411.01419", "doi": "10.48550/arXiv.2411.01419", "citation_count": 4, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1747}
{"id": "b9840168ac232243daae6db8ddc3f868cd07d04571f6fcd99c9a78fdd030dac5", "sources": ["arxiv", "semantic_scholar"], "title": "LATST: Are Transformers Necessarily Complex for Time-Series Forecasting", "abstract": "Transformer-based architectures have achieved remarkable success in natural language processing and computer vision. However, their performance in multivariate long-term forecasting often falls short compared to simpler linear baselines. Previous research has identified the traditional attention mechanism as a key factor limiting their effectiveness in this domain. To bridge this gap, we introduce LATST, a novel approach designed to mitigate entropy collapse and training instability common challenges in Transformer-based time series forecasting. We rigorously evaluate LATST across multiple real-world multivariate time series datasets, demonstrating its ability to outperform existing state-of-the-art Transformer models. Notably, LATST manages to achieve competitive performance with fewer parameters than some linear models on certain datasets, highlighting its efficiency and effectiveness.", "authors": ["Dizhen Liang"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-10-31", "url": "https://arxiv.org/abs/2410.23749", "pdf_url": "https://arxiv.org/pdf/2410.23749v9", "arxiv_id": "2410.23749", "doi": null, "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.0}
{"id": "505280469815296fe1b0d261de39af379fec95718c3660147743bd3a76a4c7f4", "sources": ["arxiv", "semantic_scholar"], "title": "In-Context Fine-Tuning for Time-Series Foundation Models", "abstract": "Motivated by the recent success of time-series foundation models for zero-shot forecasting, we present a methodology for $\\textit{in-context fine-tuning}$ of a time-series foundation model. In particular, we design a pretrained foundation model that can be prompted (at inference time) with multiple time-series examples, in order to forecast a target time-series into the future. Our foundation model is specifically trained to utilize examples from multiple related time-series in its context window (in addition to the history of the target time-series) to help it adapt to the specific distribution of the target domain at inference time. We show that such a foundation model that uses in-context examples at inference time can obtain much better performance on popular forecasting benchmarks compared to supervised deep learning methods, statistical models, as well as other time-series foundation models. Interestingly, our in-context fine-tuning approach even rivals the performance of a foundation model that is explicitly fine-tuned on the target domain.", "authors": ["Abhimanyu Das", "Matthew Faw", "Rajat Sen", "Yichen Zhou"], "categories": ["cs.LG", "cs.AI", "cs.CL"], "fields_of_study": ["Computer Science"], "published_date": "2024-10-31", "url": "https://arxiv.org/abs/2410.24087", "pdf_url": "https://arxiv.org/pdf/2410.24087v1", "arxiv_id": "2410.24087", "doi": "10.48550/arXiv.2410.24087", "citation_count": 27, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "International Conference on Machine Learning", "quality_score": 0.3618}
{"id": "9b81b06463cf4f94c1ac4e41f2acb60cb0e714935092bef60b82a82523b639b8", "sources": ["arxiv", "semantic_scholar"], "title": "Ada-MSHyper: Adaptive Multi-Scale Hypergraph Transformer for Time Series Forecasting", "abstract": "Although transformer-based methods have achieved great success in multi-scale temporal pattern interaction modeling, two key challenges limit their further development: (1) Individual time points contain less semantic information, and leveraging attention to model pair-wise interactions may cause the information utilization bottleneck. (2) Multiple inherent temporal variations (e.g., rising, falling, and fluctuating) entangled in temporal patterns. To this end, we propose Adaptive Multi-Scale Hypergraph Transformer (Ada-MSHyper) for time series forecasting. Specifically, an adaptive hypergraph learning module is designed to provide foundations for modeling group-wise interactions, then a multi-scale interaction module is introduced to promote more comprehensive pattern interactions at different scales. In addition, a node and hyperedge constraint mechanism is introduced to cluster nodes with similar semantic information and differentiate the temporal variations within each scales. Extensive experiments on 11 real-world datasets demonstrate that Ada-MSHyper achieves state-of-the-art performance, reducing prediction errors by an average of 4.56%, 10.38%, and 4.97% in MSE for long-range, short-range, and ultra-long-range time series forecasting, respectively. Code is available at https://github.com/shangzongjiang/Ada-MSHyper.", "authors": ["Zongjiang Shang", "Ling Chen", "Binqing wu", "Dongliang Cui"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-10-31", "url": "https://arxiv.org/abs/2410.23992", "pdf_url": "https://arxiv.org/pdf/2410.23992v1", "arxiv_id": "2410.23992", "doi": "10.48550/arXiv.2410.23992", "citation_count": 32, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/shangzongjiang/Ada-MSHyper", "venue": "Neural Information Processing Systems", "quality_score": 0.3796}
{"id": "fca274ebcba001581dd8f5bef1b09cfd8d629e3cd9d60c4dfd119e86d9711e9c", "sources": ["arxiv", "semantic_scholar"], "title": "Sequential Order-Robust Mamba for Time Series Forecasting", "abstract": "Mamba has recently emerged as a promising alternative to Transformers, offering near-linear complexity in processing sequential data. However, while channels in time series (TS) data have no specific order in general, recent studies have adopted Mamba to capture channel dependencies (CD) in TS, introducing a sequential order bias. To address this issue, we propose SOR-Mamba, a TS forecasting method that 1) incorporates a regularization strategy to minimize the discrepancy between two embedding vectors generated from data with reversed channel orders, thereby enhancing robustness to channel order, and 2) eliminates the 1D-convolution originally designed to capture local information in sequential data. Furthermore, we introduce channel correlation modeling (CCM), a pretraining task aimed at preserving correlations between channels from the data space to the latent space in order to enhance the ability to capture CD. Extensive experiments demonstrate the efficacy of the proposed method across standard and transfer learning scenarios. Code is available at https://github.com/seunghan96/SOR-Mamba.", "authors": ["Seunghan Lee", "Juri Hong", "Kibok Lee", "Taeyoung Park"], "categories": ["cs.LG", "cs.AI", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2024-10-30", "url": "https://arxiv.org/abs/2410.23356", "pdf_url": "https://arxiv.org/pdf/2410.23356v1", "arxiv_id": "2410.23356", "doi": "10.48550/arXiv.2410.23356", "citation_count": 3, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/seunghan96/SOR-Mamba", "venue": "arXiv.org", "quality_score": 0.1505}
{"id": "fab8ac914c06385645c55cf310c122de7dcc3871ede6cdb40a09d432911eb51f", "sources": ["arxiv", "semantic_scholar"], "title": "Dataset-Driven Channel Masks in Transformers for Multivariate Time Series", "abstract": "Recent advancements in foundation models have been successfully extended to the time series (TS) domain, facilitated by the emergence of large-scale TS datasets. However, previous efforts have primarily Capturing channel dependency (CD) is essential for modeling multivariate time series (TS), and attention-based methods have been widely employed for this purpose. Nonetheless, these methods primarily focus on modifying the architecture, often neglecting the importance of dataset-specific characteristics. In this work, we introduce the concept of partial channel dependence (PCD) to enhance CD modeling in Transformer-based models by leveraging dataset-specific information to refine the CD captured by the model. To achieve PCD, we propose channel masks (CMs), which are integrated into the attention matrices of Transformers via element-wise multiplication. CMs consist of two components: 1) a similarity matrix that captures relationships between the channels, and 2) dataset-specific and learnable domain parameters that refine the similarity matrix. We validate the effectiveness of PCD across diverse tasks and datasets with various backbones. Code is available at this repository: https://github.com/YonseiML/pcd.", "authors": ["Seunghan Lee", "Taeyoung Park", "Kibok Lee"], "categories": ["cs.LG", "cs.AI", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2024-10-30", "url": "https://arxiv.org/abs/2410.23222", "pdf_url": "https://arxiv.org/pdf/2410.23222v4", "arxiv_id": "2410.23222", "doi": "10.1109/icassp55912.2026.11464024", "citation_count": 2, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/YonseiML/pcd", "venue": "IEEE International Conference on Acoustics, Speech, and Signal Processing", "quality_score": 0.1193}
{"id": "2fd8d89de45f1af26175631ffe6e36b2a548e60275e81ae1b6eeb6af6eb1859e", "sources": ["arxiv", "semantic_scholar"], "title": "FlexTSF: A Flexible Forecasting Model for Time Series with Variable Regularities", "abstract": "Forecasting time series with irregular temporal structures remains challenging for universal pre-trained models. Existing approaches often assume regular sampling or depend heavily on imputation, limiting their applicability in real-world scenarios where irregularities are prevalent due to diverse sensing devices and recording practices. We introduce FlexTSF, a flexible forecasting model specifically designed for time series data with variable temporal regularities. At its foundation lies the IVP Patcher, a continuous-time patching module leveraging Initial Value Problems (IVPs) to inherently support uneven time intervals, variable sequence lengths, and missing values. FlexTSF employs a decoder-only architecture that integrates normalized timestamp inputs and domain-specific statistics through a specialized causal self-attention mechanism, enabling adaptability across domains. Extensive experiments on 16 datasets demonstrate FlexTSF's effectiveness, significantly outperforming existing models in classic forecasting scenarios, zero-shot generalization, and low-resource fine-tuning conditions. Ablation studies confirm the contributions of each design component and the advantage of not relying on predefined fixed patch lengths.", "authors": ["Jingge Xiao", "Yile Chen", "Gao Cong", "Wolfgang Nejdl", "Simon Gottschalk"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-10-30", "url": "https://arxiv.org/abs/2410.23160", "pdf_url": "https://arxiv.org/pdf/2410.23160v2", "arxiv_id": "2410.23160", "doi": null, "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.0}
{"id": "c1a13e3a75e4a5a1d7f961c9306f6c89ced29994e2618bce55837442f6a2cc16", "sources": ["arxiv", "semantic_scholar"], "title": "WaveRoRA: Wavelet Rotary Route Attention for Multivariate Time Series Forecasting", "abstract": "In recent years, Transformer-based models (Transformers) have achieved significant success in multivariate time series forecasting (MTSF). However, previous works focus on extracting features either from the time domain or the frequency domain, which inadequately captures the trends and periodic characteristics. To address this issue, we propose a wavelet learning framework to model complex temporal dependencies of the time series data. The wavelet domain integrates both time and frequency information, allowing for the analysis of local characteristics of signals at different scales. Additionally, the Softmax self-attention mechanism used by Transformers has quadratic complexity, which leads to excessive computational costs when capturing long-term dependencies. Therefore, we propose a novel attention mechanism: Rotary Route Attention (RoRA). Unlike Softmax attention, RoRA utilizes rotary position embeddings to inject relative positional information to sequence tokens and introduces a small number of routing tokens $r$ to aggregate information from the $KV$ matrices and redistribute it to the $Q$ matrix, offering linear complexity. We further propose WaveRoRA, which leverages RoRA to capture inter-series dependencies in the wavelet domain. We conduct extensive experiments on eight real-world datasets. The results indicate that WaveRoRA outperforms existing state-of-the-art models while maintaining lower computational costs. Our code is available at https://github.com/Leopold2333/WaveRoRA.", "authors": ["Aobo Liang", "Yan Sun", "Nadra Guizani"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-10-30", "url": "https://arxiv.org/abs/2410.22649", "pdf_url": "https://arxiv.org/pdf/2410.22649v2", "arxiv_id": "2410.22649", "doi": "10.1109/TMC.2025.3599406", "citation_count": 7, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/Leopold2333/WaveRoRA", "venue": "IEEE Transactions on Mobile Computing", "quality_score": 0.2258}
{"id": "7214c8d844d30253d818eded4bfb19f306057c5e508df1244eb7825f0574ed13", "sources": ["arxiv", "semantic_scholar"], "title": "TEAFormers: TEnsor-Augmented Transformers for Multi-Dimensional Time Series Forecasting", "abstract": "Multi-dimensional time series data, such as matrix and tensor-variate time series, are increasingly prevalent in fields such as economics, finance, and climate science. Traditional Transformer models, though adept with sequential data, do not effectively preserve these multi-dimensional structures, as their internal operations in effect flatten multi-dimensional observations into vectors, thereby losing critical multi-dimensional relationships and patterns. To address this, we introduce the Tensor-Augmented Transformer (TEAFormer), a novel method that incorporates tensor expansion and compression within the Transformer framework to maintain and leverage the inherent multi-dimensional structures, thus reducing computational costs and improving prediction accuracy. The core feature of the TEAFormer, the Tensor-Augmentation (TEA) module, utilizes tensor expansion to enhance multi-view feature learning and tensor compression for efficient information aggregation and reduced computational load. The TEA module is not just a specific model architecture but a versatile component that is highly compatible with the attention mechanism and the encoder-decoder structure of Transformers, making it adaptable to existing Transformer architectures. Our comprehensive experiments, which integrate the TEA module into three popular time series Transformer models across three real-world benchmarks, show significant performance enhancements, highlighting the potential of TEAFormers for cutting-edge time series forecasting.", "authors": ["Linghang Kong", "Elynn Chen", "Yuzhou Chen", "Yuefeng Han"], "categories": ["cs.LG", "cs.AI", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2024-10-27", "url": "https://arxiv.org/abs/2410.20439", "pdf_url": "https://arxiv.org/pdf/2410.20439v1", "arxiv_id": "2410.20439", "doi": "10.48550/arXiv.2410.20439", "citation_count": 3, "influential_citation_count": 2, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.2386}
{"id": "93692a31fc2ee9324b7e8034f857264657034331c31c1e47ca9ebf23f985f71e", "sources": ["arxiv", "semantic_scholar"], "title": "Large Language Models for Financial Aid in Financial Time-series Forecasting", "abstract": "Considering the difficulty of financial time series forecasting in financial aid, much of the current research focuses on leveraging big data analytics in financial services. One modern approach is to utilize \"predictive analysis\", analogous to forecasting financial trends. However, many of these time series data in Financial Aid (FA) pose unique challenges due to limited historical datasets and high dimensional financial information, which hinder the development of effective predictive models that balance accuracy with efficient runtime and memory usage. Pre-trained foundation models are employed to address these challenging tasks. We use state-of-the-art time series models including pre-trained LLMs (GPT-2 as the backbone), transformers, and linear models to demonstrate their ability to outperform traditional approaches, even with minimal (\"few-shot\") or no fine-tuning (\"zero-shot\"). Our benchmark study, which includes financial aid with seven other time series tasks, shows the potential of using LLMs for scarce financial datasets.", "authors": ["Md Khairul Islam", "Ayush Karmacharya", "Timothy Sue", "Judy Fox"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-10-24", "url": "https://arxiv.org/abs/2410.19025", "pdf_url": "https://arxiv.org/pdf/2410.19025v1", "arxiv_id": "2410.19025", "doi": "10.1109/BigData62323.2024.10824953", "citation_count": 9, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/UVA-MLSys/Financial-Time-Series", "venue": "BigData Congress [Services Society]", "quality_score": 0.25}
{"id": "bacfa4c098db8fc41479fee239a90462eff47a71467f4193f858939ad641d4ab", "sources": ["arxiv", "semantic_scholar"], "title": "A Comprehensive Survey of Deep Learning for Time Series Forecasting: Architectural Diversity and Open Challenges", "abstract": "Time series forecasting is a critical task that provides key information for decision-making. After traditional statistical and machine learning approaches, various fundamental deep learning architectures such as MLPs, CNNs, RNNs, and GNNs have been developed. However, the structural limitations caused by the inductive biases of each deep learning architecture constrained their performance. Transformer models, which excel at handling long-term dependencies, have become significant architectural components for time series forecasting. However, recent research has shown that alternatives such as simple linear layers can outperform Transformers. These findings have opened up new possibilities for using diverse architectures, ranging from fundamental deep learning models to emerging architectures and hybrid approaches. In this context, architectural modeling of time series forecasting has now entered a renaissance. This survey not only provides a historical context for time series forecasting but also offers comprehensive and timely analysis of the movement toward architectural diversification. By comparing and re-examining deep learning models, we uncover new perspectives and present recent trends, including hybrid, diffusion, Mamba, and foundation models. By focusing on the inherent characteristics of time series data, we also address open challenges that have gained attention in time series forecasting, such as channel dependency, distribution shift, causality, and feature extraction. These contributions help lower entry barriers for newcomers by providing a systematic understanding of the diverse research areas in time series forecasting (TSF), while offering seasoned researchers broader perspectives and new opportunities through in-depth exploration of TSF challenges. (Shortened due to arXiv's 1,920-character limit. Full version in the paper.)", "authors": ["Jongseon Kim", "Hyungjoon Kim", "HyunGi Kim", "Dongjun Lee", "Sungroh Yoon"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-10-24", "url": "https://arxiv.org/abs/2411.05793", "pdf_url": "https://arxiv.org/pdf/2411.05793v3", "arxiv_id": "2411.05793", "doi": null, "citation_count": 89, "influential_citation_count": 3, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.4886}
{"id": "cf059be380871ee33f8c53a3d481b0bac5efbc99927e8488be69756edac2db4a", "sources": ["arxiv", "semantic_scholar"], "title": "On the Regularization of Learnable Embeddings for Time Series Forecasting", "abstract": "In forecasting multiple time series, accounting for the individual features of each sequence can be challenging. To address this, modern deep learning methods for time series analysis combine a shared (global) model with local layers, specific to each time series, often implemented as learnable embeddings. Ideally, these local embeddings should encode meaningful representations of the unique dynamics of each sequence. However, when these are learned end-to-end as parameters of a forecasting model, they may end up acting as mere sequence identifiers. Shared processing blocks may then become reliant on such identifiers, limiting their transferability to new contexts. In this paper, we address this issue by investigating methods to regularize the learning of local learnable embeddings for time series processing. Specifically, we perform the first extensive empirical study on the subject and show how such regularizations consistently improve performance in widely adopted architectures. Furthermore, we show that methods attempting to prevent the co-adaptation of local and global parameters by means of embeddings perturbation are particularly effective in this context. In this regard, we include in the comparison several perturbation-based regularization methods, going as far as periodically resetting the embeddings during training. The obtained results provide an important contribution to understanding the interplay between learnable local parameters and shared processing layers: a key challenge in modern time series processing models and a step toward developing effective foundation models for time series.", "authors": ["Luca Butera", "Giovanni De Felice", "Andrea Cini", "Cesare Alippi"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-10-18", "url": "https://arxiv.org/abs/2410.14630", "pdf_url": "https://arxiv.org/pdf/2410.14630v2", "arxiv_id": "2410.14630", "doi": null, "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "L. Butera, G. D. Felice, A. Cini, and C. Alippi. On the regularization of learnable embeddings for time series forecasting. Transactions on Machine Learning Research, 2025. ISSN 2835-8856. URL https://openreview.net/forum?id=F5ALCh3GWG", "quality_score": 0.0753}
{"id": "26e6857929ea77e9b29e9bf4e9bd0362fbeba0b82ba7873afe52c6065b411b4f", "sources": ["arxiv", "semantic_scholar"], "title": "TimeSeriesExam: A time series understanding exam", "abstract": "Large Language Models (LLMs) have recently demonstrated a remarkable ability to model time series data. These capabilities can be partly explained if LLMs understand basic time series concepts. However, our knowledge of what these models understand about time series data remains relatively limited. To address this gap, we introduce TimeSeriesExam, a configurable and scalable multiple-choice question exam designed to assess LLMs across five core time series understanding categories: pattern recognition, noise understanding, similarity analysis, anomaly detection, and causality analysis. TimeSeriesExam comprises of over 700 questions, procedurally generated using 104 carefully curated templates and iteratively refined to balance difficulty and their ability to discriminate good from bad models. We test 7 state-of-the-art LLMs on the TimeSeriesExam and provide the first comprehensive evaluation of their time series understanding abilities. Our results suggest that closed-source models such as GPT-4 and Gemini understand simple time series concepts significantly better than their open-source counterparts, while all models struggle with complex concepts such as causality analysis. We believe that the ability to programatically generate questions is fundamental to assessing and improving LLM's ability to understand and reason about time series data.", "authors": ["Yifu Cai", "Arjun Choudhry", "Mononito Goswami", "Artur Dubrawski"], "categories": ["cs.AI", "cs.CL"], "fields_of_study": ["Computer Science"], "published_date": "2024-10-18", "url": "https://arxiv.org/abs/2410.14752", "pdf_url": "https://arxiv.org/pdf/2410.14752v1", "arxiv_id": "2410.14752", "doi": "10.48550/arXiv.2410.14752", "citation_count": 41, "influential_citation_count": 9, "has_code": true, "code_url": null, "venue": "arXiv.org", "quality_score": 0.5}
{"id": "1d4e3c740fd49bfa66828a4a8d22a033da723b2890975d67ce5e406785f239a6", "sources": ["arxiv", "semantic_scholar"], "title": "Analyzing Deep Transformer Models for Time Series Forecasting via Manifold Learning", "abstract": "Transformer models have consistently achieved remarkable results in various domains such as natural language processing and computer vision. However, despite ongoing research efforts to better understand these models, the field still lacks a comprehensive understanding. This is particularly true for deep time series forecasting methods, where analysis and understanding work is relatively limited. Time series data, unlike image and text information, can be more challenging to interpret and analyze. To address this, we approach the problem from a manifold learning perspective, assuming that the latent representations of time series forecasting models lie next to a low-dimensional manifold. In our study, we focus on analyzing the geometric features of these latent data manifolds, including intrinsic dimension and principal curvatures. Our findings reveal that deep transformer models exhibit similar geometric behavior across layers, and these geometric features are correlated with model performance. Additionally, we observe that untrained models initially have different structures, but they rapidly converge during training. By leveraging our geometric analysis and differentiable tools, we can potentially design new and improved deep forecasting neural networks. This approach complements existing analysis studies and contributes to a better understanding of transformer models in the context of time series forecasting. Code is released at https://github.com/azencot-group/GATLM.", "authors": ["Ilya Kaufman", "Omri Azencot"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-10-17", "url": "https://arxiv.org/abs/2410.13792", "pdf_url": "https://arxiv.org/pdf/2410.13792v1", "arxiv_id": "2410.13792", "doi": "10.48550/arXiv.2410.13792", "citation_count": 5, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/azencot-group/GATLM", "venue": null, "quality_score": 0.1945}
{"id": "8cf37a940ba54db63c55099b129fa3313d48eae2cc2a57c49047099a80c6bdb7", "sources": ["arxiv", "semantic_scholar"], "title": "LLM-Mixer: Multiscale Mixing in LLMs for Time Series Forecasting", "abstract": "Time series forecasting remains a challenging task, particularly in the context of complex multiscale temporal patterns. This study presents LLM-Mixer, a framework that improves forecasting accuracy through the combination of multiscale time-series decomposition with pre-trained LLMs (Large Language Models). LLM-Mixer captures both short-term fluctuations and long-term trends by decomposing the data into multiple temporal resolutions and processing them with a frozen LLM, guided by a textual prompt specifically designed for time-series data. Extensive experiments conducted on multivariate and univariate datasets demonstrate that LLM-Mixer achieves competitive performance, outperforming recent state-of-the-art models across various forecasting horizons. This work highlights the potential of combining multiscale analysis and LLMs for effective and scalable time-series forecasting.", "authors": ["Md Kowsher", "Md. Shohanur Islam Sobuj", "Nusrat Jahan Prottasha", "E. Alejandro Alanis", "Ozlem Ozmen Garibay", "Niloofar Yousefi"], "categories": ["cs.LG", "cs.CL"], "fields_of_study": ["Computer Science"], "published_date": "2024-10-15", "url": "https://arxiv.org/abs/2410.11674", "pdf_url": "https://arxiv.org/pdf/2410.11674v2", "arxiv_id": "2410.11674", "doi": "10.48550/arXiv.2410.11674", "citation_count": 6, "influential_citation_count": 2, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.2386}
{"id": "894debd07b5b00e37b5f5fdff8681571e57f5e8f7ad27a965b0868d93c291bf8", "sources": ["arxiv", "semantic_scholar"], "title": "TSFM-Bench: A Comprehensive and Unified Benchmark of Foundation Models for Time Series Forecasting", "abstract": "Time Series Forecasting (TSF) is key functionality in numerous fields, such as financial investment, weather services, and energy management. Although increasingly capable TSF methods occur, many of them require domain-specific data collection and model training and do not generalize well when applied in other domains. Time Series Foundation Models (TSFMs) that are pre-trained on massive heterogeneous time series data aim to overcome these limitations. The prospects for generalizability have spurred the development of a new generation of TSFMs. This study proposes a benchmark, TSFM-Bench, to facilitate comprehensive and unified evaluation of TSFMs. TSFM-Bench covers a wide range of TSFMs, including those based on large language models and those pre-trained on time series data. TSFM-Bench supports multiple forecasting scenarios, including zero-shot, few-shot, and full-shot, enabling assessment across the full range of adaptation strategies. TSFM-Bench also provides a standardized experimental protocols for critical evaluation processes such as dataset splitting, loading, normalization, and few-shot sampling, facilitating consistency and fairness. We report on an extensive evaluation of TSFMs across a diverse range of datasets spanning multiple domains and exhibiting varied statistical characteristics. Specifically, we identify pros and cons and inherent limitations of existing TSFMs, and we propose potential directions for new model designs.", "authors": ["Zhe Li", "Xiangfei Qiu", "Peng Chen", "Yihang Wang", "Hanyin Cheng", "Yang Shu", "Jilin Hu", "Chenjuan Guo", "Aoying Zhou", "Christian S. Jensen", "Bin Yang"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-10-15", "url": "https://arxiv.org/abs/2410.11802", "pdf_url": "https://arxiv.org/pdf/2410.11802v6", "arxiv_id": "2410.11802", "doi": "10.1145/3711896.3737442", "citation_count": 48, "influential_citation_count": 5, "has_code": false, "code_url": null, "venue": "Knowledge Discovery and Data Mining", "quality_score": 0.4225}
{"id": "2adbf3b728948e73222e53aa34751cca3a31ae341bb3e40c1ac356fc202dfb31", "sources": ["arxiv", "semantic_scholar"], "title": "Time-Series Foundation AI Model for Value-at-Risk Forecasting", "abstract": "This study is the first to analyze the performance of a time-series foundation AI model for Value-at-Risk (VaR), which essentially forecasts the left-tail quantiles of returns. Foundation models, pre-trained on diverse datasets, can be applied in a zero-shot setting with minimal data or further improved through finetuning. We compare Google's TimesFM model to conventional parametric and non-parametric models, including GARCH and Generalized Autoregressive Score (GAS), using 19 years of daily returns from the SP 100 index and its constituents. Backtesting with over 8.5 years of out-of-sample data shows that the fine-tuned foundation model consistently outperforms traditional methods in actual-over-expected ratios. For the quantile score loss function, it performs comparably to the best econometric model, GAS. Overall, the foundation model ranks as the best or among the top performers across the 0.01, 0.025, 0.05, and 0.1 quantile forecasting. Fine-tuning significantly improves accuracy, showing that zero-shot use is not optimal for VaR.", "authors": ["Anubha Goel", "Puneet Pasricha", "Juho Kanniainen"], "categories": ["q-fin.RM", "cs.AI"], "fields_of_study": ["Economics", "Computer Science"], "published_date": "2024-10-15", "url": "https://arxiv.org/abs/2410.11773", "pdf_url": "https://arxiv.org/pdf/2410.11773v7", "arxiv_id": "2410.11773", "doi": null, "citation_count": 5, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.1945}
{"id": "ab4861bed91762987eef717bf5b908cb5579178555e2aa84d4072e9c6982c728", "sources": ["arxiv", "semantic_scholar"], "title": "Transfer Learning with Foundational Models for Time Series Forecasting using Low-Rank Adaptations", "abstract": "Foundational Models are an emerging widely used technique of GenAI. These models are distinguished by their scalability and the ease with which they can be adapted through the exploitation of Transfer Learning. The availability of high computational power and large datasets have supported their development, achieving a high generalization capacity due to the enormous and heterogeneous amounts of data used in their initial training. These characteristics contribute to a solid base that can be adapted or adjusted to a wide range of tasks, increasing their applicability. This study proposes the methodology LLIAM, a straightforward adaptation of a kind of FM, Large Language Models, for the Time Series Forecasting task. An adequate time-series prompting schema and Low-Rank Adaptations are used to enhance the knowledge of the model with diverse time series datasets, known as the fine-tuning phase. A study divided in two stages has been performed for evaluating the effectiveness of the proposed methodology. Initially, a comparison was made between the performance of LLIAM and different state-of-the-art DL algorithms, including Recurrent Neural Networks and Temporal Convolutional Networks, as well as a LLM-based method, TimeLLM. Following this, a zero-shot study is presented in order to evaluate the generalization capacity of the proposed methodology with time series datasets from unknown domains not considered in the model training. The outcomes of this investigation demonstrate the efficacy of LLIAM, highlighting that this straightforward and general approach can attain competent results without the necessity for applying complex modifications. This work also encourages the use of available resources (such as these pre-trained models) and efficient fine-tuning techniques to avoid unnecessary and costly training, narrowing the gap between the goals of traditional AI and Green AI.", "authors": ["M. Germán-Morales", "A. J. Rivera-Rivas", "M. J. del Jesus Díaz", "C. J. Carmona"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-10-15", "url": "https://arxiv.org/abs/2410.11539", "pdf_url": "https://arxiv.org/pdf/2410.11539v3", "arxiv_id": "2410.11539", "doi": "10.1016/j.inffus.2025.103247", "citation_count": 14, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "Information Fusion", "quality_score": 0.294}
{"id": "17ed053d4556033bac4b4e8723515e0c1db51b05702be6ade8c9b4ab741fc59b", "sources": ["arxiv", "semantic_scholar"], "title": "GIFT-Eval: A Benchmark For General Time Series Forecasting Model Evaluation", "abstract": "Time series foundation models excel in zero-shot forecasting, handling diverse tasks without explicit training. However, the advancement of these models has been hindered by the lack of comprehensive benchmarks. To address this gap, we introduce the General Time Series Forecasting Model Evaluation, GIFT-Eval, a pioneering benchmark aimed at promoting evaluation across diverse datasets. GIFT-Eval encompasses 23 datasets over 144,000 time series and 177 million data points, spanning seven domains, 10 frequencies, multivariate inputs, and prediction lengths ranging from short to long-term forecasts. To facilitate the effective pretraining and evaluation of foundation models, we also provide a non-leaking pretraining dataset containing approximately 230 billion data points. Additionally, we provide a comprehensive analysis of 17 baselines, which includes statistical models, deep learning models, and foundation models. We discuss each model in the context of various benchmark characteristics and offer a qualitative analysis that spans both deep learning and foundation models. We believe the insights from this analysis, along with access to this new standard zero-shot time series forecasting benchmark, will guide future developments in time series foundation models. Code, data, and the leaderboard can be found at https://github.com/SalesforceAIResearch/gift-eval .", "authors": ["Taha Aksu", "Gerald Woo", "Juncheng Liu", "Xu Liu", "Chenghao Liu", "Silvio Savarese", "Caiming Xiong", "Doyen Sahoo"], "categories": ["cs.LG", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2024-10-14", "url": "https://arxiv.org/abs/2410.10393", "pdf_url": "https://arxiv.org/pdf/2410.10393v2", "arxiv_id": "2410.10393", "doi": "10.48550/arXiv.2410.10393", "citation_count": 129, "influential_citation_count": 55, "has_code": true, "code_url": "https://github.com/SalesforceAIResearch/gift-eval", "venue": "arXiv.org", "quality_score": 0.8741}
{"id": "3bdc1e286a3aebf5e8a7717a35362233fdc698c853f916be062afdf475541d98", "sources": ["arxiv", "semantic_scholar"], "title": "Moirai-MoE: Empowering Time Series Foundation Models with Sparse Mixture of Experts", "abstract": "Time series foundation models have demonstrated impressive performance as zero-shot forecasters. However, achieving effectively unified training on time series remains an open challenge. Existing approaches introduce some level of model specialization to account for the highly heterogeneous nature of time series data. For instance, Moirai pursues unified training by employing multiple input/output projection layers, each tailored to handle time series at a specific frequency. Similarly, TimesFM maintains a frequency embedding dictionary for this purpose. We identify two major drawbacks to this human-imposed frequency-level model specialization: (1) Frequency is not a reliable indicator of the underlying patterns in time series. For example, time series with different frequencies can display similar patterns, while those with the same frequency may exhibit varied patterns. (2) Non-stationarity is an inherent property of real-world time series, leading to varied distributions even within a short context window of a single time series. Frequency-level specialization is too coarse-grained to capture this level of diversity. To address these limitations, this paper introduces Moirai-MoE, using a single input/output projection layer while delegating the modeling of diverse time series patterns to the sparse mixture of experts (MoE) within Transformers. With these designs, Moirai-MoE reduces reliance on human-defined heuristics and enables automatic token-level specialization. Extensive experiments on 39 datasets demonstrate the superiority of Moirai-MoE over existing foundation models in both in-distribution and zero-shot scenarios. Furthermore, this study conducts comprehensive model analyses to explore the inner workings of time series MoE foundation models and provides valuable insights for future research.", "authors": ["Xu Liu", "Juncheng Liu", "Gerald Woo", "Taha Aksu", "Yuxuan Liang", "Roger Zimmermann", "Chenghao Liu", "Silvio Savarese", "Caiming Xiong", "Doyen Sahoo"], "categories": ["cs.LG", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2024-10-14", "url": "https://arxiv.org/abs/2410.10469", "pdf_url": "https://arxiv.org/pdf/2410.10469v1", "arxiv_id": "2410.10469", "doi": "10.48550/arXiv.2410.10469", "citation_count": 96, "influential_citation_count": 10, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.5207}
{"id": "6ace902a51023c98520e291388bc2701491ab2e7d832dc6a1b6f0f6c4abc162b", "sources": ["arxiv", "semantic_scholar"], "title": "Benchmarking Time Series Foundation Models for Short-Term Household Electricity Load Forecasting", "abstract": "Accurate household electricity short-term load forecasting (STLF) is key to future and sustainable energy systems. While various studies have analyzed statistical, machine learning, or deep learning approaches for household electricity STLF, recently proposed time series foundation models such as Chronos, TimesFM or Time-MoE promise a new approach for household electricity STLF. These models are trained on a vast amount of time series data and are able to forecast time series without explicit task-specific training (zero-shot learning). In this study, we benchmark the forecasting capabilities of time series foundation models compared to Trained-from-Scratch (TFS) Transformer-based approaches. Our results suggest that foundation models perform comparably to TFS Transformer models, while certain time series foundation models outperform all TFS models when the input size increases. At the same time, they require less effort, as they need no domain-specific training and only limited contextual data for inference.", "authors": ["Marcel Meyer", "David Zapata", "Sascha Kaltenpoth", "Oliver Müller"], "categories": ["cs.CE"], "fields_of_study": ["Computer Science"], "published_date": "2024-10-12", "url": "https://arxiv.org/abs/2410.09487", "pdf_url": "https://arxiv.org/pdf/2410.09487v3", "arxiv_id": "2410.09487", "doi": "10.1109/ACCESS.2025.3648056", "citation_count": 16, "influential_citation_count": 2, "has_code": false, "code_url": null, "venue": "IEEE Access", "quality_score": 0.3076}
{"id": "9b726525fc87f90c59071aaa1a3153fb880011f1799d822a96b34262071bdfdf", "sources": ["arxiv", "semantic_scholar"], "title": "Mamba4Cast: Efficient Zero-Shot Time Series Forecasting with State Space Models", "abstract": "This paper introduces Mamba4Cast, a zero-shot foundation model for time series forecasting. Based on the Mamba architecture and inspired by Prior-data Fitted Networks (PFNs), Mamba4Cast generalizes robustly across diverse time series tasks without the need for dataset specific fine-tuning. Mamba4Cast's key innovation lies in its ability to achieve strong zero-shot performance on real-world datasets while having much lower inference times than time series foundation models based on the transformer architecture. Trained solely on synthetic data, the model generates forecasts for entire horizons in a single pass, outpacing traditional auto-regressive approaches. Our experiments show that Mamba4Cast performs competitively against other state-of-the-art foundation models in various data sets while scaling significantly better with the prediction length. The source code can be accessed at https://github.com/automl/Mamba4Cast.", "authors": ["Sathya Kamesh Bhethanabhotla", "Omar Swelam", "Julien Siems", "David Salinas", "Frank Hutter"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-10-12", "url": "https://arxiv.org/abs/2410.09385", "pdf_url": "https://arxiv.org/pdf/2410.09385v1", "arxiv_id": "2410.09385", "doi": "10.48550/arXiv.2410.09385", "citation_count": 22, "influential_citation_count": 1, "has_code": true, "code_url": "https://github.com/automl/Mamba4Cast", "venue": "arXiv.org", "quality_score": 0.3404}
{"id": "67c950a3fbe14600ac1ce632cab34b268c12596eb836820cabf83e0b8a50c84c", "sources": ["arxiv", "semantic_scholar"], "title": "Timer-XL: Long-Context Transformers for Unified Time Series Forecasting", "abstract": "We present Timer-XL, a causal Transformer for unified time series forecasting. To uniformly predict multidimensional time series, we generalize next token prediction, predominantly adopted for 1D token sequences, to multivariate next token prediction. The paradigm formulates various forecasting tasks as a long-context prediction problem. We opt for decoder-only Transformers that capture causal dependencies from varying-length contexts for unified forecasting, making predictions on non-stationary univariate time series, multivariate series with complicated dynamics and correlations, as well as covariate-informed contexts that include exogenous variables. Technically, we propose a universal TimeAttention to capture fine-grained intra- and inter-series dependencies of flattened time series tokens (patches), which is further enhanced by deft position embedding for temporal causality and variable equivalence. Timer-XL achieves state-of-the-art performance across task-specific forecasting benchmarks through a unified approach. Based on large-scale pre-training, Timer-XL achieves state-of-the-art zero-shot performance, making it a promising architecture for pre-trained time series models. Code is available at this repository: https://github.com/thuml/Timer-XL.", "authors": ["Yong Liu", "Guo Qin", "Xiangdong Huang", "Jianmin Wang", "Mingsheng Long"], "categories": ["cs.LG", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2024-10-07", "url": "https://arxiv.org/abs/2410.04803", "pdf_url": "https://arxiv.org/pdf/2410.04803v4", "arxiv_id": "2410.04803", "doi": "10.48550/arXiv.2410.04803", "citation_count": 85, "influential_citation_count": 16, "has_code": true, "code_url": "https://github.com/thuml/Timer-XL", "venue": "International Conference on Learning Representations", "quality_score": 0.6152}
{"id": "5a005622350858bf47c076275efdf16d829e7c5d7ffa5842c7463163f24da54f", "sources": ["arxiv", "semantic_scholar"], "title": "Continuous Ensemble Weather Forecasting with Diffusion models", "abstract": "Weather forecasting has seen a shift in methods from numerical simulations to data-driven systems. While initial research in the area focused on deterministic forecasting, recent works have used diffusion models to produce skillful ensemble forecasts. These models are trained on a single forecasting step and rolled out autoregressively. However, they are computationally expensive and accumulate errors for high temporal resolution due to the many rollout steps. We address these limitations with Continuous Ensemble Forecasting, a novel and flexible method for sampling ensemble forecasts in diffusion models. The method can generate temporally consistent ensemble trajectories completely in parallel, with no autoregressive steps. Continuous Ensemble Forecasting can also be combined with autoregressive rollouts to yield forecasts at an arbitrary fine temporal resolution without sacrificing accuracy. We demonstrate that the method achieves competitive results for global weather forecasting with good probabilistic properties.", "authors": ["Martin Andrae", "Tomas Landelius", "Joel Oskarsson", "Fredrik Lindsten"], "categories": ["cs.LG", "physics.ao-ph"], "fields_of_study": ["Computer Science", "Physics"], "published_date": "2024-10-07", "url": "https://arxiv.org/abs/2410.05431", "pdf_url": "https://arxiv.org/pdf/2410.05431v2", "arxiv_id": "2410.05431", "doi": "10.48550/arXiv.2410.05431", "citation_count": 28, "influential_citation_count": 1, "has_code": true, "code_url": "https://github.com/martinandrae/Continuous-Ensemble-Forecasting", "venue": "International Conference on Learning Representations", "quality_score": 0.3656}
{"id": "d7bcaad301bf08c74624087e5e1d7c60005dc17385e02ae2e7c8c202e955e90c", "sources": ["arxiv", "semantic_scholar"], "title": "Metadata Matters for Time Series: Informative Forecasting with Transformers", "abstract": "Time series forecasting is prevalent in extensive real-world applications, such as financial analysis and energy planning. Previous studies primarily focus on time series modality, endeavoring to capture the intricate variations and dependencies inherent in time series. Beyond numerical time series data, we notice that metadata (e.g.~dataset and variate descriptions) also carries valuable information essential for forecasting, which can be used to identify the application scenario and provide more interpretable knowledge than digit sequences. Inspired by this observation, we propose a Metadata-informed Time Series Transformer (MetaTST), which incorporates multiple levels of context-specific metadata into Transformer forecasting models to enable informative time series forecasting. To tackle the unstructured nature of metadata, MetaTST formalizes them into natural languages by pre-designed templates and leverages large language models (LLMs) to encode these texts into metadata tokens as a supplement to classic series tokens, resulting in an informative embedding. Further, a Transformer encoder is employed to communicate series and metadata tokens, which can extend series representations by metadata information for more accurate forecasting. This design also allows the model to adaptively learn context-specific patterns across various scenarios, which is particularly effective in handling large-scale, diverse-scenario forecasting tasks. Experimentally, MetaTST achieves state-of-the-art compared to advanced time series models and LLM-based methods on widely acknowledged short- and long-term forecasting benchmarks, covering both single-dataset individual and multi-dataset joint training settings.", "authors": ["Jiaxiang Dong", "Haixu Wu", "Yuxuan Wang", "Li Zhang", "Jianmin Wang", "Mingsheng Long"], "categories": ["cs.LG", "cs.CL"], "fields_of_study": ["Computer Science"], "published_date": "2024-10-04", "url": "https://arxiv.org/abs/2410.03806", "pdf_url": "https://arxiv.org/pdf/2410.03806v1", "arxiv_id": "2410.03806", "doi": "10.48550/arXiv.2410.03806", "citation_count": 7, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.2258}
{"id": "8c4d9b69a90ccb0dd70f8ef888ba770343cfbd6480683abb8a105494566be4cb", "sources": ["arxiv", "semantic_scholar"], "title": "Local Attention Mechanism: Boosting the Transformer Architecture for Long-Sequence Time Series Forecasting", "abstract": "Transformers have become the leading choice in natural language processing over other deep learning architectures. This trend has also permeated the field of time series analysis, especially for long-horizon forecasting, showcasing promising results both in performance and running time. In this paper, we introduce Local Attention Mechanism (LAM), an efficient attention mechanism tailored for time series analysis. This mechanism exploits the continuity properties of time series to reduce the number of attention scores computed. We present an algorithm for implementing LAM in tensor algebra that runs in time and memory O(nlogn), significantly improving upon the O(n^2) time and memory complexity of traditional attention mechanisms. We also note the lack of proper datasets to evaluate long-horizon forecast models. Thus, we propose a novel set of datasets to improve the evaluation of models addressing long-horizon forecasting challenges. Our experimental analysis demonstrates that the vanilla transformer architecture magnified with LAM surpasses state-of-the-art models, including the vanilla attention mechanism. These results confirm the effectiveness of our approach and highlight a range of future challenges in long-sequence time series forecasting.", "authors": ["Ignacio Aguilera-Martos", "Andrés Herrera-Poyatos", "Julián Luengo", "Francisco Herrera"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-10-04", "url": "https://arxiv.org/abs/2410.03805", "pdf_url": "https://arxiv.org/pdf/2410.03805v3", "arxiv_id": "2410.03805", "doi": "10.48550/arXiv.2410.03805", "citation_count": 7, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.2258}
{"id": "fc45dbecfa5e443f3aba39373b07527ede3ba6665b6aeb582430fe4b21133cfb", "sources": ["arxiv", "semantic_scholar"], "title": "Continuous-Time Linear Positional Embedding for Irregular Time Series Forecasting", "abstract": "Irregularly sampled time series forecasting, characterized by non-uniform intervals, is prevalent in practical applications. However, previous research have been focused on regular time series forecasting, typically relying on transformer architectures. To extend transformers to handle irregular time series, we tackle the positional embedding which represents the temporal information of the data. We propose CTLPE, a method learning a continuous linear function for encoding temporal information. The two challenges of irregular time series, inconsistent observation patterns and irregular time gaps, are solved by learning a continuous-time function and concise representation of position. Additionally, the linear continuous function is empirically shown superior to other continuous functions by learning a neural controlled differential equation-based positional embedding, and theoretically supported with properties of ideal positional embedding. CTLPE outperforms existing techniques across various irregularly-sampled time series datasets, showcasing its enhanced efficacy.", "authors": ["Byunghyun Kim", "Jae-Gil Lee"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-09-30", "url": "https://arxiv.org/abs/2409.20092", "pdf_url": "https://arxiv.org/pdf/2409.20092v1", "arxiv_id": "2409.20092", "doi": "10.48550/arXiv.2409.20092", "citation_count": 2, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1193}
{"id": "99c5ac93285588e4d5e0f628a261b37d877c4004901724d2f58cffef1daad641", "sources": ["arxiv", "semantic_scholar"], "title": "CycleNet: Enhancing Time Series Forecasting through Modeling Periodic Patterns", "abstract": "The stable periodic patterns present in time series data serve as the foundation for conducting long-horizon forecasts. In this paper, we pioneer the exploration of explicitly modeling this periodicity to enhance the performance of models in long-term time series forecasting (LTSF) tasks. Specifically, we introduce the Residual Cycle Forecasting (RCF) technique, which utilizes learnable recurrent cycles to model the inherent periodic patterns within sequences, and then performs predictions on the residual components of the modeled cycles. Combining RCF with a Linear layer or a shallow MLP forms the simple yet powerful method proposed in this paper, called CycleNet. CycleNet achieves state-of-the-art prediction accuracy in multiple domains including electricity, weather, and energy, while offering significant efficiency advantages by reducing over 90% of the required parameter quantity. Furthermore, as a novel plug-and-play technique, the RCF can also significantly improve the prediction accuracy of existing models, including PatchTST and iTransformer. The source code is available at: https://github.com/ACAT-SCUT/CycleNet.", "authors": ["Shengsheng Lin", "Weiwei Lin", "Xinyi Hu", "Wentai Wu", "Ruichao Mo", "Haocheng Zhong"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-09-27", "url": "https://arxiv.org/abs/2409.18479", "pdf_url": "https://arxiv.org/pdf/2409.18479v2", "arxiv_id": "2409.18479", "doi": "10.48550/arXiv.2409.18479", "citation_count": 165, "influential_citation_count": 20, "has_code": true, "code_url": "https://github.com/ACAT-SCUT/CycleNet", "venue": "Neural Information Processing Systems", "quality_score": 0.6611}
{"id": "e14e7ff954f458f59bfbf14909b40d2c7573bf4e2a12a1cadd9598ad40c2b9cf", "sources": ["arxiv", "semantic_scholar"], "title": "Optimal starting point for time series forecasting", "abstract": "Recent advances on time series forecasting mainly focus on improving the forecasting models themselves. However, when the time series data suffer from potential structural breaks or concept drifts, the forecasting performance might be significantly reduced. In this paper, we introduce a novel approach called Optimal Starting Point Time Series Forecast (OSP-TSP) for optimal forecasting, which can be combined with existing time series forecasting models. By adjusting the sequence length via leveraging the XGBoost and LightGBM models, the proposed approach can determine the optimal starting point (OSP) of the time series and then enhance the prediction performances of the base forecasting models. To illustrate the effectiveness of the proposed approach, comprehensive empirical analysis have been conducted on the M4 dataset and other real world datasets. Empirical results indicate that predictions based on the OSP-TSP approach consistently outperform those using the complete time series dataset. Moreover, comparison results reveals that combining our approach with existing forecasting models can achieve better prediction accuracy, which also reflect the advantages of the proposed approach.", "authors": ["Yiming Zhong", "Yinuo Ren", "Guangyao Cao", "Feng Li", "Haobo Qi"], "categories": ["stat.AP", "cs.LG"], "fields_of_study": ["Mathematics", "Computer Science"], "published_date": "2024-09-25", "url": "https://arxiv.org/abs/2409.16843", "pdf_url": "https://arxiv.org/pdf/2409.16843v2", "arxiv_id": "2409.16843", "doi": "10.48550/arXiv.2409.16843", "citation_count": 4, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "Expert systems with applications", "quality_score": 0.1747}
{"id": "1e17f74f2d3c7c612c6b4160ff8efc985d1b4d7a8ad8e9420e8fde1dc4333648", "sources": ["arxiv", "semantic_scholar"], "title": "Time-MoE: Billion-Scale Time Series Foundation Models with Mixture of Experts", "abstract": "Deep learning for time series forecasting has seen significant advancements over the past decades. However, despite the success of large-scale pre-training in language and vision domains, pre-trained time series models remain limited in scale and operate at a high cost, hindering the development of larger capable forecasting models in real-world applications. In response, we introduce Time-MoE, a scalable and unified architecture designed to pre-train larger, more capable forecasting foundation models while reducing inference costs. By leveraging a sparse mixture-of-experts (MoE) design, Time-MoE enhances computational efficiency by activating only a subset of networks for each prediction, reducing computational load while maintaining high model capacity. This allows Time-MoE to scale effectively without a corresponding increase in inference costs. Time-MoE comprises a family of decoder-only transformer models that operate in an auto-regressive manner and support flexible forecasting horizons with varying input context lengths. We pre-trained these models on our newly introduced large-scale data Time-300B, which spans over 9 domains and encompassing over 300 billion time points. For the first time, we scaled a time series foundation model up to 2.4 billion parameters, achieving significantly improved forecasting precision. Our results validate the applicability of scaling laws for training tokens and model size in the context of time series forecasting. Compared to dense models with the same number of activated parameters or equivalent computation budgets, our models consistently outperform them by large margin. These advancements position Time-MoE as a state-of-the-art solution for tackling real-world time series forecasting challenges with superior capability, efficiency, and flexibility.", "authors": ["Xiaoming Shi", "Shiyu Wang", "Yuqi Nie", "Dianqi Li", "Zhou Ye", "Qingsong Wen", "Ming Jin"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-09-24", "url": "https://arxiv.org/abs/2409.16040", "pdf_url": "https://arxiv.org/pdf/2409.16040v4", "arxiv_id": "2409.16040", "doi": "10.48550/arXiv.2409.16040", "citation_count": 285, "influential_citation_count": 59, "has_code": false, "code_url": null, "venue": "International Conference on Learning Representations", "quality_score": 0.8891}
{"id": "ced96b5c8633c331d91c0226ee2df3e757d9888d066c7c1012118d707ff5286d", "sources": ["arxiv", "semantic_scholar"], "title": "Towards Long-Context Time Series Foundation Models", "abstract": "Time series foundation models have shown impressive performance on a variety of tasks, across a wide range of domains, even in zero-shot settings. However, most of these models are designed to handle short univariate time series as an input. This limits their practical use, especially in domains such as healthcare with copious amounts of long and multivariate data with strong temporal and intra-variate dependencies. Our study bridges this gap by cataloging and systematically comparing various context expansion techniques from both language and time series domains, and introducing a novel compressive memory mechanism to allow encoder-only TSFMs to effectively model intra-variate dependencies. We demonstrate the benefits of our approach by imbuing MOMENT, a recent family of multi-task time series foundation models, with the multivariate context.", "authors": ["Nina Żukowska", "Mononito Goswami", "Michał Wiliński", "Willa Potosnak", "Artur Dubrawski"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-09-20", "url": "https://arxiv.org/abs/2409.13530", "pdf_url": "https://arxiv.org/pdf/2409.13530v1", "arxiv_id": "2409.13530", "doi": "10.48550/arXiv.2409.13530", "citation_count": 6, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.2113}
{"id": "24d181e156a97613fd220ad5acffd8cfa8b6e5db7f4362284114fbaf91513fa3", "sources": ["arxiv", "semantic_scholar"], "title": "Fine-Tuning a Time Series Foundation Model with Wasserstein Loss", "abstract": "Inspired by recent advancements in large language models (LLMs) for Natural Language Processing (NLP), there has been a surge in research focused on developing foundational models for time series forecasting. One approach involves training LLM architectures on tokenized time series data using cross-entropy loss. Although this method has demonstrated promising results, cross-entropy loss is primarily designed for classification tasks and does not account for the distance between classes. To address this limitation, we propose using the Wasserstein loss for such architectures. To validate our approach, we fine-tuned a foundational time series model on $22$ zero-shot datasets, comparing the performance of cross-entropy loss with that of Wasserstein loss. Our results demonstrate that replacing cross-entropy loss with Wasserstein loss significantly improves point estimation.", "authors": ["Andrei Chernov"], "categories": ["cs.LG", "cs.AI", "cs.CL"], "fields_of_study": ["Computer Science"], "published_date": "2024-09-18", "url": "https://arxiv.org/abs/2409.15367", "pdf_url": "https://arxiv.org/pdf/2409.15367v2", "arxiv_id": "2409.15367", "doi": "10.48550/arXiv.2409.15367", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.0753}
{"id": "10c04e8b385d2b09777436234e33717c4bbb532db209758968cfdf45a0d3b82c", "sources": ["arxiv", "semantic_scholar"], "title": "Towards Time Series Reasoning with LLMs", "abstract": "Multi-modal large language models (MLLMs) have enabled numerous advances in understanding and reasoning in domains like vision, but we have not yet seen this broad success for time-series. Although prior works on time-series MLLMs have shown promising performance in time-series forecasting, very few works show how an LLM could be used for time-series reasoning in natural language. We propose a novel multi-modal time-series LLM approach that learns generalizable information across various domains with powerful zero-shot performance. First, we train a lightweight time-series encoder on top of an LLM to directly extract time-series information. Then, we fine-tune our model with chain-of-thought augmented time-series tasks to encourage the model to generate reasoning paths. We show that our model learns a latent representation that reflects specific time-series features (e.g. slope, frequency), as well as outperforming GPT-4o on a set of zero-shot reasoning tasks on a variety of domains.", "authors": ["Winnie Chow", "Lauren Gardiner", "Haraldur T. Hallgrímsson", "Maxwell A. Xu", "Shirley You Ren"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-09-17", "url": "https://arxiv.org/abs/2409.11376", "pdf_url": "https://arxiv.org/pdf/2409.11376v2", "arxiv_id": "2409.11376", "doi": "10.48550/arXiv.2409.11376", "citation_count": 43, "influential_citation_count": 5, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.4109}
{"id": "acc224100f2434678ccf14eb45281317abd9a51a2d39a3e15e3c3225be7b4ecb", "sources": ["arxiv", "semantic_scholar"], "title": "Implicit Reasoning in Deep Time Series Forecasting", "abstract": "Recently, time series foundation models have shown promising zero-shot forecasting performance on time series from a wide range of domains. However, it remains unclear whether their success stems from a true understanding of temporal dynamics or simply from memorizing the training data. While implicit reasoning in language models has been studied, similar evaluations for time series models have been largely unexplored. This work takes an initial step toward assessing the reasoning abilities of deep time series forecasting models. We find that certain linear, MLP-based, and patch-based Transformer models generalize effectively in systematically orchestrated out-of-distribution scenarios, suggesting underexplored reasoning capabilities beyond simple pattern memorization.", "authors": ["Willa Potosnak", "Cristian Challu", "Mononito Goswami", "Michał Wiliński", "Nina Żukowska", "Artur Dubrawski"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-09-17", "url": "https://arxiv.org/abs/2409.10840", "pdf_url": "https://arxiv.org/pdf/2409.10840v4", "arxiv_id": "2409.10840", "doi": "10.48550/arXiv.2409.10840", "citation_count": 7, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.2258}
{"id": "169c3efa120b94fa5a96c572b70e36e8ab8ff752204377b0a8ffd6a9c8840f48", "sources": ["arxiv", "semantic_scholar"], "title": "Integration of Mamba and Transformer -- MAT for Long-Short Range Time Series Forecasting with Application to Weather Dynamics", "abstract": "Long-short range time series forecasting is essential for predicting future trends and patterns over extended periods. While deep learning models such as Transformers have made significant strides in advancing time series forecasting, they often encounter difficulties in capturing long-term dependencies and effectively managing sparse semantic features. The state-space model, Mamba, addresses these issues through its adept handling of selective input and parallel computing, striking a balance between computational efficiency and prediction accuracy. This article examines the advantages and disadvantages of both Mamba and Transformer models, and introduces a combined approach, MAT, which leverages the strengths of each model to capture unique long-short range dependencies and inherent evolutionary patterns in multivariate time series. Specifically, MAT harnesses the long-range dependency capabilities of Mamba and the short-range characteristics of Transformers. Experimental results on benchmark weather datasets demonstrate that MAT outperforms existing comparable methods in terms of prediction accuracy, scalability, and memory efficiency.", "authors": ["Wenqing Zhang", "Junming Huang", "Ruotong Wang", "Changsong Wei", "Wenqian Huang", "Yuxin Qiao"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-09-13", "url": "https://arxiv.org/abs/2409.08530", "pdf_url": "https://arxiv.org/pdf/2409.08530v1", "arxiv_id": "2409.08530", "doi": "10.1109/ICECCE63537.2024.10823516", "citation_count": 25, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.3537}
{"id": "2711c8e69fad0d7e8b7c96857a1866deabf4d2d779e865f730413558f6331d41", "sources": ["arxiv", "semantic_scholar"], "title": "ReAugment: Model Zoo-Guided RL for Few-Shot Time Series Augmentation and Forecasting", "abstract": "Time series forecasting, particularly in few-shot learning scenarios, is challenging due to the limited availability of high-quality training data. To address this, we present a pilot study on using reinforcement learning (RL) for time series data augmentation. Our method, ReAugment, tackles three critical questions: which parts of the training set should be augmented, how the augmentation should be performed, and what advantages RL brings to the process. Specifically, our approach maintains a forecasting model zoo, and by measuring prediction diversity across the models, we identify samples with higher probabilities for overfitting and use them as the anchor points for augmentation. Leveraging RL, our method adaptively transforms the overfit-prone samples into new data that not only enhances training set diversity but also directs the augmented data to target regions where the forecasting models are prone to overfitting. We validate the effectiveness of ReAugment across a wide range of base models, showing its advantages in both standard time series forecasting and few-shot learning tasks.", "authors": ["Haochen Yuan", "Yutong Wang", "Yihong Chen", "Yunbo Wang", "Xiaokang Yang"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-09-10", "url": "https://arxiv.org/abs/2409.06282", "pdf_url": "https://arxiv.org/pdf/2409.06282v4", "arxiv_id": "2409.06282", "doi": null, "citation_count": 3, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.1505}
{"id": "126b335f0d5ce3da0f83647d965158dc151e0c0219f2a501044171f68fdc0ab3", "sources": ["arxiv", "semantic_scholar"], "title": "TimeDiT: General-purpose Diffusion Transformers for Time Series Foundation Model", "abstract": "Foundation models, particularly Large Language Models (LLMs), have revolutionized text and video processing, yet time series data presents distinct challenges for such approaches due to domain-specific features such as missing values, multi-resolution characteristics, etc. Furthermore, the de-facto autoregressive transformers tend to learn deterministic temporal dependencies within pre-trained data while overlooking inherent uncertainties and lacking integration of physical constraints. In this paper, we introduce TimeDiT, a diffusion transformer model that synergistically combines transformer-based temporal dependency learning with diffusion-based probabilistic sampling. TimeDiT employs a unified masking mechanism to harmonize the training and inference process across diverse tasks while introducing a theoretically grounded, finetuning-free model editing strategy that enables flexible integration of external knowledge during sampling. Acknowledging the challenges of unifying multiple downstream tasks under a single model, our systematic evaluation demonstrates TimeDiT's effectiveness both in fundamental tasks, i.e., forecasting and imputation, through zero-shot/fine-tuning; and in domain tasks, i.e., multi-resolution forecasting, anomaly detection, and data generation, establishing it as a \\textit{proto-foundation model} that bridges the gap between general-purpose and domain-specific models.", "authors": ["Defu Cao", "Wen Ye", "Yizhou Zhang", "Yan Liu"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-09-03", "url": "https://arxiv.org/abs/2409.02322", "pdf_url": "https://arxiv.org/pdf/2409.02322v2", "arxiv_id": "2409.02322", "doi": "10.48550/arXiv.2409.02322", "citation_count": 41, "influential_citation_count": 7, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.4515}
{"id": "7806763c9ca1a13e2dca3358b0fac966b7d75d94e255e01b0fcf836fc27ef1a0", "sources": ["arxiv", "semantic_scholar"], "title": "VisionTS: Visual Masked Autoencoders Are Free-Lunch Zero-Shot Time Series Forecasters", "abstract": "Foundation models have emerged as a promising approach in time series forecasting (TSF). Existing approaches either repurpose large language models (LLMs) or build large-scale time series datasets to develop TSF foundation models for universal forecasting. However, these methods face challenges due to the severe cross-domain gap or in-domain heterogeneity. This paper explores a new road to building a TSF foundation model from rich, high-quality natural images. Our key insight is that a visual masked autoencoder, pre-trained on the ImageNet dataset, can naturally be a numeric series forecaster. By reformulating TSF as an image reconstruction task, we bridge the gap between image pre-training and TSF downstream tasks. Surprisingly, without further adaptation in the time series domain, the proposed VisionTS could achieve better zero-shot forecast performance than existing TSF foundation models. With fine-tuning for one epoch, VisionTS could further improve the forecasting and achieve state-of-the-art performance in most cases. Extensive experiments reveal intrinsic similarities between images and real-world time series, suggesting that visual models may offer a \"free lunch\" for TSF and highlight the potential for future cross-modality research. Our code is publicly available at https://github.com/Keytoyze/VisionTS.", "authors": ["Mouxiang Chen", "Lefei Shen", "Zhuo Li", "Xiaoyun Joy Wang", "Jianling Sun", "Chenghao Liu"], "categories": ["cs.CV", "cs.AI", "cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-08-30", "url": "https://arxiv.org/abs/2408.17253", "pdf_url": "https://arxiv.org/pdf/2408.17253v4", "arxiv_id": "2408.17253", "doi": "10.48550/arXiv.2408.17253", "citation_count": 83, "influential_citation_count": 19, "has_code": true, "code_url": "https://github.com/Keytoyze/VisionTS", "venue": "International Conference on Machine Learning", "quality_score": 0.6505}
{"id": "27e5db83ab5b94ef4fece28aa64da8c465c79768b0580bcc8efd53459c266f68", "sources": ["arxiv", "semantic_scholar"], "title": "Distributed Lag Transformer based on Time-Variable-Aware Learning for Explainable Multivariate Time Series Forecasting", "abstract": "Time series data is a key element of big data analytics, commonly found in domains such as finance, healthcare, climate forecasting, and transportation. In large scale real world settings, such data is often high dimensional and multivariate, requiring advanced forecasting methods that are both accurate and interpretable. Although Transformer based models perform well in multivariate time series forecasting (MTSF), their lack of explainability limits their use in critical applications. To overcome this, we propose Distributed Lag Transformer (DLFormer), a novel Transformer architecture for explainable and scalable MTSF. DLFormer integrates a distributed lag embedding and a time variable aware learning (TVAL) mechanism to structurally model both local and global temporal dependencies and explicitly capture the influence of past variables on future outcomes. Experiments on ten benchmark and real world datasets show that DLFormer achieves state of the art predictive accuracy while offering robust, interpretable insights into variable wise and temporal dynamics. These results highlight ability of DLFormer to bridge the gap between performance and explainability, making it highly suitable for practical big data forecasting tasks.", "authors": ["Younghwi Kim", "Dohee Kim", "Joongrock Kim", "Sunghyun Sim"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-08-29", "url": "https://arxiv.org/abs/2408.16896", "pdf_url": "https://arxiv.org/pdf/2408.16896v2", "arxiv_id": "2408.16896", "doi": null, "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.0}
{"id": "31ecf871871b035ceb62296316f0978d26b3f8ce3fae628bfc9fa6921db9821d", "sources": ["arxiv", "semantic_scholar"], "title": "Mamba or Transformer for Time Series Forecasting? Mixture of Universals (MoU) Is All You Need", "abstract": "Time series forecasting requires balancing short-term and long-term dependencies for accurate predictions. Existing methods mainly focus on long-term dependency modeling, neglecting the complexities of short-term dynamics, which may hinder performance. Transformers are superior in modeling long-term dependencies but are criticized for their quadratic computational cost. Mamba provides a near-linear alternative but is reported less effective in time series longterm forecasting due to potential information loss. Current architectures fall short in offering both high efficiency and strong performance for long-term dependency modeling. To address these challenges, we introduce Mixture of Universals (MoU), a versatile model to capture both short-term and long-term dependencies for enhancing performance in time series forecasting. MoU is composed of two novel designs: Mixture of Feature Extractors (MoF), an adaptive method designed to improve time series patch representations for short-term dependency, and Mixture of Architectures (MoA), which hierarchically integrates Mamba, FeedForward, Convolution, and Self-Attention architectures in a specialized order to model long-term dependency from a hybrid perspective. The proposed approach achieves state-of-the-art performance while maintaining relatively low computational costs. Extensive experiments on seven real-world datasets demonstrate the superiority of MoU. Code is available at https://github.com/lunaaa95/mou/.", "authors": ["Sijia Peng", "Yun Xiong", "Yangyong Zhu", "Zhiqiang Shen"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-08-28", "url": "https://arxiv.org/abs/2408.15997", "pdf_url": "https://arxiv.org/pdf/2408.15997v1", "arxiv_id": "2408.15997", "doi": "10.48550/arXiv.2408.15997", "citation_count": 2, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/lunaaa95/mou/", "venue": "arXiv.org", "quality_score": 0.1193}
{"id": "48428d2fa7c25dd4aa526e85904be7b02cec10df95b1e6d3ff9dabc0b65e84a9", "sources": ["arxiv", "semantic_scholar"], "title": "Time Series Analysis for Education: Methods, Applications, and Future Directions", "abstract": "Recent advancements in the collection and analysis of sequential educational data have brought time series analysis to a pivotal position in educational research, highlighting its essential role in facilitating data-driven decision-making. However, there is a lack of comprehensive summaries that consolidate these advancements. To the best of our knowledge, this paper is the first to provide a comprehensive review of time series analysis techniques specifically within the educational context. We begin by exploring the landscape of educational data analytics, categorizing various data sources and types relevant to education. We then review four prominent time series methods-forecasting, classification, clustering, and anomaly detection-illustrating their specific application points in educational settings. Subsequently, we present a range of educational scenarios and applications, focusing on how these methods are employed to address diverse educational tasks, which highlights the practical integration of multiple time series methods to solve complex educational problems. Finally, we conclude with a discussion on future directions, including personalized learning analytics, multimodal data fusion, and the role of large language models (LLMs) in educational time series. The contributions of this paper include a detailed taxonomy of educational data, a synthesis of time series techniques with specific educational applications, and a forward-looking perspective on emerging trends and future research opportunities in educational analysis. The related papers and resources are available and regularly updated at the project page.", "authors": ["Shengzhong Mao", "Chaoli Zhang", "Yichi Song", "Jindong Wang", "Xiao-Jun Zeng", "Zenglin Xu", "Qingsong Wen"], "categories": ["cs.LG", "cs.AI", "cs.CY"], "fields_of_study": ["Computer Science"], "published_date": "2024-08-25", "url": "https://arxiv.org/abs/2408.13960", "pdf_url": "https://arxiv.org/pdf/2408.13960v2", "arxiv_id": "2408.13960", "doi": "10.48550/arXiv.2408.13960", "citation_count": 16, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/ai-for-edu/time-series-analysis-for-education", "venue": "arXiv.org", "quality_score": 0.3076}
{"id": "36f4e3056f79d2df10f955ed0b9f90acd91e64c9bf66a17ea1bb58fcade7de36", "sources": ["arxiv", "semantic_scholar"], "title": "Image Segmentation in Foundation Model Era: A Survey", "abstract": "Image segmentation is a long-standing challenge in computer vision, studied continuously over several decades, as evidenced by seminal algorithms such as N-Cut, FCN, and MaskFormer. With the advent of foundation models (FMs), contemporary segmentation methodologies have embarked on a new epoch by either adapting FMs (e.g., CLIP, Stable Diffusion, DINO) for image segmentation or developing dedicated segmentation foundation models (e.g., SAM). These approaches not only deliver superior segmentation performance, but also herald newfound segmentation capabilities previously unseen in deep learning context. However, current research in image segmentation lacks a detailed analysis of distinct characteristics, challenges, and solutions associated with these advancements. This survey seeks to fill this gap by providing a thorough review of cutting-edge research centered around FM-driven image segmentation. We investigate two basic lines of research -- generic image segmentation (i.e., semantic segmentation, instance segmentation, panoptic segmentation), and promptable image segmentation (i.e., interactive segmentation, referring segmentation, few-shot segmentation) -- by delineating their respective task settings, background concepts, and key challenges. Furthermore, we provide insights into the emergence of segmentation knowledge from FMs like CLIP, Stable Diffusion, and DINO. An exhaustive overview of over 300 segmentation approaches is provided to encapsulate the breadth of current research efforts. Subsequently, we engage in a discussion of open issues and potential avenues for future research. We envisage that this fresh, comprehensive, and systematic survey catalyzes the evolution of advanced image segmentation systems. A public website is created to continuously track developments in this fast advancing field: \\url{https://github.com/stanley-313/ImageSegFM-Survey}.", "authors": ["Tianfei Zhou", "Wang Xia", "Fei Zhang", "Boyu Chang", "Wenguan Wang", "Ye Yuan", "Ender Konukoglu", "Daniel Cremers"], "categories": ["cs.CV"], "fields_of_study": ["Computer Science"], "published_date": "2024-08-23", "url": "https://arxiv.org/abs/2408.12957", "pdf_url": "https://arxiv.org/pdf/2408.12957v3", "arxiv_id": "2408.12957", "doi": "10.48550/arXiv.2408.12957", "citation_count": 49, "influential_citation_count": 1, "has_code": true, "code_url": "https://github.com/stanley-313/ImageSegFM-Survey}", "venue": "arXiv.org", "quality_score": 0.4247}
{"id": "d687048f36441b8523419cfa2a9e305c62cfb2849471c0d7626f4fee3a5565a2", "sources": ["arxiv", "semantic_scholar"], "title": "SDE: A Simplified and Disentangled Dependency Encoding Framework for State Space Models in Time Series Forecasting", "abstract": "In recent years, advancements in deep learning have spurred the development of numerous models for Long-term Time Series Forecasting (LTSF). However, most existing approaches struggle to fully capture the complex and structured dependencies inherent in time series data. In this work, we identify and formally define three critical dependencies that are fundamental to forecasting accuracy: order dependency and semantic dependency along the temporal dimension, as well as cross-variate dependency across the feature dimension. These dependencies are often treated in isolation, and improper handling can introduce noise and degrade forecasting performance. To bridge this gap, we investigate the potential of State Space Models (SSMs) for LTSF and emphasize their inherent advantages in capturing these essential dependencies. Additionally, we empirically observe that excessive nonlinearity in conventional SSMs introduce redundancy when applied to semantically sparse time series data. Motivated by this insight, we propose SDE (Simplified and Disentangled Dependency Encoding), a novel framework designed to enhance the capability of SSMs for LTSF. Specifically, we first eliminate unnecessary nonlinearities in vanilla SSMs, thereby improving the suitability for time series forecasting. Building on this foundation, we introduce a disentangled encoding strategy, which empowers SSMs to efficiently model cross-variate dependencies while mitigating interference between the temporal and feature dimensions. Furthermore, we provide rigorous theoretical justifications to substantiate our design choices. Extensive experiments on nine real-world benchmark datasets demonstrate that SDE-enhanced SSMs consistently outperform state-of-the-art time series forecasting models.Our code is available at https://github.com/YukinoAsuna/SAMBA.", "authors": ["Zixuan Weng", "Jindong Han", "Wenzhao Jiang", "Hao Liu"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-08-22", "url": "https://arxiv.org/abs/2408.12068", "pdf_url": "https://arxiv.org/pdf/2408.12068v3", "arxiv_id": "2408.12068", "doi": "10.1145/3711896.3737119", "citation_count": 4, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/YukinoAsuna/SAMBA", "venue": "Knowledge Discovery and Data Mining", "quality_score": 0.1747}
{"id": "1c8f5c6b4425ca0f4049a73e39520090fd080d411df248fb84240b1290e7246c", "sources": ["arxiv", "semantic_scholar"], "title": "Scalable Transformer for High Dimensional Multivariate Time Series Forecasting", "abstract": "Deep models for Multivariate Time Series (MTS) forecasting have recently demonstrated significant success. Channel-dependent models capture complex dependencies that channel-independent models cannot capture. However, the number of channels in real-world applications outpaces the capabilities of existing channel-dependent models, and contrary to common expectations, some models underperform the channel-independent models in handling high-dimensional data, which raises questions about the performance of channel-dependent models. To address this, our study first investigates the reasons behind the suboptimal performance of these channel-dependent models on high-dimensional MTS data. Our analysis reveals that two primary issues lie in the introduced noise from unrelated series that increases the difficulty of capturing the crucial inter-channel dependencies, and challenges in training strategies due to high-dimensional data. To address these issues, we propose STHD, the Scalable Transformer for High-Dimensional Multivariate Time Series Forecasting. STHD has three components: a) Relation Matrix Sparsity that limits the noise introduced and alleviates the memory issue; b) ReIndex applied as a training strategy to enable a more flexible batch size setting and increase the diversity of training data; and c) Transformer that handles 2-D inputs and captures channel dependencies. These components jointly enable STHD to manage the high-dimensional MTS while maintaining computational feasibility. Furthermore, experimental results show STHD's considerable improvement on three high-dimensional datasets: Crime-Chicago, Wiki-People, and Traffic. The source code and dataset are publicly available https://github.com/xinzzzhou/ScalableTransformer4HighDimensionMTSF.git.", "authors": ["Xin Zhou", "Weiqing Wang", "Wray Buntine", "Shilin Qu", "Abishek Sriramulu", "Weicong Tan", "Christoph Bergmeir"], "categories": ["cs.LG", "cs.AI", "cs.IR"], "fields_of_study": ["Computer Science"], "published_date": "2024-08-08", "url": "https://arxiv.org/abs/2408.04245", "pdf_url": "https://arxiv.org/pdf/2408.04245v1", "arxiv_id": "2408.04245", "doi": "10.1145/3627673.3679757", "citation_count": 19, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/xinzzzhou/ScalableTransformer4HighDimensionMTSF.git", "venue": "International Conference on Information and Knowledge Management", "quality_score": 0.3253}
{"id": "cd76185cbfea047416e8c835c68a28190e97bd7d6249d700bf7c7a13439b3bc6", "sources": ["arxiv", "semantic_scholar"], "title": "Inter-Series Transformer: Attending to Products in Time Series Forecasting", "abstract": "Time series forecasting is an important task in many fields ranging from supply chain management to weather forecasting. Recently, Transformer neural network architectures have shown promising results in forecasting on common time series benchmark datasets. However, application to supply chain demand forecasting, which can have challenging characteristics such as sparsity and cross-series effects, has been limited. In this work, we explore the application of Transformer-based models to supply chain demand forecasting. In particular, we develop a new Transformer-based forecasting approach using a shared, multi-task per-time series network with an initial component applying attention across time series, to capture interactions and help address sparsity. We provide a case study applying our approach to successfully improve demand prediction for a medical device manufacturing company. To further validate our approach, we also apply it to public demand forecasting datasets as well and demonstrate competitive to superior performance compared to a variety of baseline and state-of-the-art forecast methods across the private and public datasets.", "authors": ["Rares Cristian", "Pavithra Harsha", "Clemente Ocejo", "Georgia Perakis", "Brian Quanz", "Ioannis Spantidakis", "Hamza Zerhouni"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-08-07", "url": "https://arxiv.org/abs/2408.03872", "pdf_url": "https://arxiv.org/pdf/2408.03872v1", "arxiv_id": "2408.03872", "doi": "10.48550/arXiv.2408.03872", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.0753}
{"id": "4b86f366ce2e4a57aca19b896baba53a7ec0666b49687758339ed3f6b20a9fc4", "sources": ["arxiv", "semantic_scholar"], "title": "Time series forecasting with high stakes: A field study of the air cargo industry", "abstract": "Time series forecasting in the air cargo industry presents unique challenges due to volatile market dynamics and the significant impact of accurate forecasts on generated revenue. This paper explores a comprehensive approach to demand forecasting at the origin-destination (O\\&D) level, focusing on the development and implementation of machine learning models in decision-making for the air cargo industry. We leverage a mixture of experts framework, combining statistical and advanced deep learning models to provide reliable forecasts for cargo demand over a six-month horizon. The results demonstrate that our approach outperforms industry benchmarks, offering actionable insights for cargo capacity allocation and strategic decision-making in the air cargo industry. While this work is applied in the airline industry, the methodology is broadly applicable to any field where forecast-based decision-making in a volatile environment is crucial.", "authors": ["Abhinav Garg", "Naman Shukla", "Maarten Wormer"], "categories": ["cs.LG", "eess.SY"], "fields_of_study": ["Computer Science", "Engineering"], "published_date": "2024-07-29", "url": "https://arxiv.org/abs/2407.20192", "pdf_url": "https://arxiv.org/pdf/2407.20192v2", "arxiv_id": "2407.20192", "doi": "10.48550/arXiv.2407.20192", "citation_count": 2, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1193}
{"id": "d8349102f7de581c5b6105a4de5de486b148836a0771feab8e3be7096de143fc", "sources": ["arxiv", "semantic_scholar"], "title": "Survey and Taxonomy: The Role of Data-Centric AI in Transformer-Based Time Series Forecasting", "abstract": "Alongside the continuous process of improving AI performance through the development of more sophisticated models, researchers have also focused their attention to the emerging concept of data-centric AI, which emphasizes the important role of data in a systematic machine learning training process. Nonetheless, the development of models has also continued apace. One result of this progress is the development of the Transformer Architecture, which possesses a high level of capability in multiple domains such as Natural Language Processing (NLP), Computer Vision (CV) and Time Series Forecasting (TSF). Its performance is, however, heavily dependent on input data preprocessing and output data evaluation, justifying a data-centric approach to future research. We argue that data-centric AI is essential for training AI models, particularly for transformer-based TSF models efficiently. However, there is a gap regarding the integration of transformer-based TSF and data-centric AI. This survey aims to pin down this gap via the extensive literature review based on the proposed taxonomy. We review the previous research works from a data-centric AI perspective and we intend to lay the foundation work for the future development of transformer-based architecture and data-centric AI.", "authors": ["Jingjing Xu", "Caesar Wu", "Yuan-Fang Li", "Gregoire Danoy", "Pascal Bouvry"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-07-29", "url": "https://arxiv.org/abs/2407.19784", "pdf_url": "https://arxiv.org/pdf/2407.19784v1", "arxiv_id": "2407.19784", "doi": "10.48550/arXiv.2407.19784", "citation_count": 3, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1505}
{"id": "36111d5041d86335ba15ecd09012ee75605c8b7c0bf8d25c00e9c6e10c301e73", "sources": ["arxiv", "semantic_scholar"], "title": "DAM: Towards A Foundation Model for Time Series Forecasting", "abstract": "It is challenging to scale time series forecasting models such that they forecast accurately for multiple distinct domains and datasets, all with potentially different underlying collection procedures (e.g., sample resolution), patterns (e.g., periodicity), and prediction requirements (e.g., reconstruction vs. forecasting). We call this general task universal forecasting. Existing methods usually assume that input data is regularly sampled, and they forecast to pre-determined horizons, resulting in failure to generalise outside of the scope of their training. We propose the DAM - a neural model that takes randomly sampled histories and outputs an adjustable basis composition as a continuous function of time for forecasting to non-fixed horizons. It involves three key components: (1) a flexible approach for using randomly sampled histories from a long-tail distribution, that enables an efficient global perspective of the underlying temporal dynamics while retaining focus on the recent history; (2) a transformer backbone that is trained on these actively sampled histories to produce, as representational output, (3) the basis coefficients of a continuous function of time. We show that a single univariate DAM, trained on 25 time series datasets, either outperformed or closely matched existing SoTA models at multivariate long-term forecasting across 18 datasets, including 8 held-out for zero-shot transfer, even though these models were trained to specialise for each dataset-horizon combination. This single DAM excels at zero-shot transfer and very-long-term forecasting, performs well at imputation, is interpretable via basis function composition and attention, can be tuned for different inference-cost requirements, is robust to missing and irregularly sampled data {by design}.", "authors": ["Luke Darlow", "Qiwen Deng", "Ahmed Hassan", "Martin Asenov", "Rajkarn Singh", "Artjom Joosen", "Adam Barker", "Amos Storkey"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-07-25", "url": "https://arxiv.org/abs/2407.17880", "pdf_url": "https://arxiv.org/pdf/2407.17880v1", "arxiv_id": "2407.17880", "doi": "10.48550/arXiv.2407.17880", "citation_count": 10, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.2603}
{"id": "b585403e914789b8c5ca24aa6291cf9a07d8ce104f63ef48e92681545d2414cf", "sources": ["arxiv", "semantic_scholar"], "title": "On the optimal prediction of extreme events in heavy-tailed time series with applications to solar flare forecasting", "abstract": "The prediction of extreme events in time series is a fundamental problem arising in many financial, scientific, engineering, and other applications. We begin by establishing a general Neyman-Pearson-type characterization of optimal extreme event predictors in terms of density ratios. This yields new insights and several closed-form optimal extreme event predictors for additive models. These results naturally extend to time series, where we study optimal extreme event prediction for both light- and heavy-tailed autoregressive and moving average models. Using a uniform law of large numbers for ergodic time series, we establish the asymptotic optimality of an empirical version of the optimal predictor for autoregressive models. Using multivariate regular variation, we obtain an expression for the optimal extremal precision in heavy-tailed infinite moving averages, which provides theoretical bounds on the ability to predict extremes in this general class of models. We address the important problem of predicting solar flares by applying our theory and methodology to a state-of-the-art time series consisting of solar soft X-ray flux measurements. Our results demonstrate the success and limitations in solar flare forecasting of long-memory autoregressive models and long-range-dependent, heavy-tailed FARIMA models.", "authors": ["Victor Verma", "Stilian Stoev", "Yang Chen"], "categories": ["math.ST", "stat.AP", "stat.ME"], "fields_of_study": ["Mathematics"], "published_date": "2024-07-16", "url": "https://arxiv.org/abs/2407.11887", "pdf_url": "https://arxiv.org/pdf/2407.11887v2", "arxiv_id": "2407.11887", "doi": "10.1111/jtsa.12820", "citation_count": 7, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "Journal of Time Series Analysis", "quality_score": 0.2258}
{"id": "4552fb7dbe3ac802fefcace08122e9dc16e702fc56c282601c2160fdd97cafbb", "sources": ["arxiv", "semantic_scholar"], "title": "ViTime: Foundation Model for Time Series Forecasting Powered by Vision Intelligence", "abstract": "Time series forecasting (TSF) possesses great practical values in various fields, including power and energy, transportation, etc. TSF methods have been studied based on knowledge from classical statistics to modern deep learning. Yet, all of them were developed based on one fundamental concept, the numerical data fitting. Thus, the models developed have long been known to be problem-specific and lacking application generalizability. Practitioners expect a TSF foundation model that serves TSF tasks in different applications. The central question is then how to develop such a TSF foundation model. This paper offers one pioneering study in the TSF foundation model development method and proposes a vision intelligence-powered framework, ViTime, for the first time. ViTime fundamentally shifts TSF from numerical fitting to operations based on a binary image-based time series metric space and naturally supports both point and probabilistic forecasting. We also provide rigorous theoretical analyses of ViTime, including quantization-induced system error bounds and principled strategies for optimal parameter selection. Furthermore, we propose RealTS, an innovative synthesis algorithm generating diverse and realistic training samples, effectively enriching the training data and significantly enhancing model generalizability. Extensive experiments demonstrate ViTime's state-of-the-art performance. In zero-shot scenarios, ViTime outperforms TimesFM by 9-15\\%. With just 10\\% fine-tuning data, ViTime surpasses both leading foundation models and fully-supervised benchmarks, a gap that widens with 100\\% fine-tuning. ViTime also exhibits exceptional robustness, effectively handling missing data and outperforming TimesFM by 20-30\\% under various data perturbations, validating the power of its visual space data operation paradigm.", "authors": ["Luoxiao Yang", "Yun Wang", "Xinqi Fan", "Israel Cohen", "Jingdong Chen", "Zijun Zhang"], "categories": ["cs.LG", "cs.AI", "cs.CV"], "fields_of_study": ["Computer Science"], "published_date": "2024-07-10", "url": "https://arxiv.org/abs/2407.07311", "pdf_url": "https://arxiv.org/pdf/2407.07311v4", "arxiv_id": "2407.07311", "doi": null, "citation_count": 10, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "Transactions on Machine Learning Research,2025", "quality_score": 0.2603}
{"id": "513b101d148e83d4cf959531e94862650a8dd955fd24e5ca80490b2a7c53d08d", "sources": ["arxiv", "semantic_scholar"], "title": "Toto: Time Series Optimized Transformer for Observability", "abstract": "This technical report describes the Time Series Optimized Transformer for Observability (Toto), a new state of the art foundation model for time series forecasting developed by Datadog. In addition to advancing the state of the art on generalized time series benchmarks in domains such as electricity and weather, this model is the first general-purpose time series forecasting foundation model to be specifically tuned for observability metrics. Toto was trained on a dataset of one trillion time series data points, the largest among all currently published time series foundation models. Alongside publicly available time series datasets, 75% of the data used to train Toto consists of fully anonymous numerical metric data points from the Datadog platform. In our experiments, Toto outperforms existing time series foundation models on observability data. It does this while also excelling at general-purpose forecasting tasks, achieving state-of-the-art zero-shot performance on multiple open benchmark datasets.", "authors": ["Ben Cohen", "Emaad Khwaja", "Kan Wang", "Charles Masson", "Elise Ramé", "Youssef Doubli", "Othmane Abou-Amal"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-07-10", "url": "https://arxiv.org/abs/2407.07874", "pdf_url": "https://arxiv.org/pdf/2407.07874v2", "arxiv_id": "2407.07874", "doi": "10.48550/arXiv.2407.07874", "citation_count": 31, "influential_citation_count": 4, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.3763}
{"id": "8c1371a6dbefd0d1c7a12dc8c671130ac9df1f8c5f7f9d7e462377188885e693", "sources": ["arxiv", "semantic_scholar"], "title": "JANET: Joint Adaptive predictioN-region Estimation for Time-series", "abstract": "Conformal prediction provides machine learning models with prediction sets that offer theoretical guarantees, but the underlying assumption of exchangeability limits its applicability to time series data. Furthermore, existing approaches struggle to handle multi-step ahead prediction tasks, where uncertainty estimates across multiple future time points are crucial. We propose JANET (Joint Adaptive predictioN-region Estimation for Time-series), a novel framework for constructing conformal prediction regions that are valid for both univariate and multivariate time series. JANET generalises the inductive conformal framework and efficiently produces joint prediction regions with controlled K-familywise error rates, enabling flexible adaptation to specific application needs. Our empirical evaluation demonstrates JANET's superior performance in multi-step prediction tasks across diverse time series datasets, highlighting its potential for reliable and interpretable uncertainty quantification in sequential data.", "authors": ["Eshant English", "Eliot Wong-Toi", "Matteo Fontana", "Stephan Mandt", "Padhraic Smyth", "Christoph Lippert"], "categories": ["stat.ML", "cs.LG", "stat.ME"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2024-07-08", "url": "https://arxiv.org/abs/2407.06390", "pdf_url": "https://arxiv.org/pdf/2407.06390v2", "arxiv_id": "2407.06390", "doi": "10.1007/s10994-025-06812-2", "citation_count": 3, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "Machine-mediated learning", "quality_score": 0.1505}
{"id": "c1375debacf9afcf81ab2ee8519f1ab8a9d5d53b02ff4a30c31164c4c9f5b7e6", "sources": ["arxiv", "semantic_scholar"], "title": "On the Workflows and Smells of Leaderboard Operations (LBOps): An Exploratory Study of Foundation Model Leaderboards", "abstract": "Foundation models (FM), such as large language models (LLMs), which are large-scale machine learning (ML) models, have demonstrated remarkable adaptability in various downstream software engineering (SE) tasks, such as code completion, code understanding, and software development. As a result, FM leaderboards have become essential tools for SE teams to compare and select the best third-party FMs for their specific products and purposes. However, the lack of standardized guidelines for FM evaluation and comparison threatens the transparency of FM leaderboards and limits stakeholders' ability to perform effective FM selection. As a first step towards addressing this challenge, our research focuses on understanding how these FM leaderboards operate in real-world scenarios (\"leaderboard operations\") and identifying potential pitfalls and areas for improvement (\"leaderboard smells\"). In this regard, we collect up to 1,045 FM leaderboards from five different sources: GitHub, Hugging Face Spaces, Papers With Code, spreadsheet and independent platform, to examine their documentation and engage in direct communication with leaderboard operators to understand their workflows. Through card sorting and negotiated agreement, we identify five distinct workflow patterns and develop a domain model that captures the key components and their interactions within these workflows. We then identify eight unique types of leaderboard smells in LBOps. By mitigating these smells, SE teams can improve transparency, accountability, and collaboration in current LBOps practices, fostering a more robust and responsible ecosystem for FM comparison and selection.", "authors": ["Zhimin Zhao", "Abdul Ali Bangash", "Filipe Roseiro Côgo", "Bram Adams", "Ahmed E. Hassan"], "categories": ["cs.SE", "cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-07-04", "url": "https://arxiv.org/abs/2407.04065", "pdf_url": "https://arxiv.org/pdf/2407.04065v4", "arxiv_id": "2407.04065", "doi": "10.1109/TSE.2025.3533972", "citation_count": 5, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/SAILResearch/awesome-foundation-model-leaderboards;", "venue": "IEEE Transactions on Software Engineering", "quality_score": 0.1945}
{"id": "c40c79dae9bb78d7b2e9b7efbb2d98b33f0216a8ef5b666b0f1651bd88f1e5d6", "sources": ["arxiv", "semantic_scholar"], "title": "Are Language Models Actually Useful for Time Series Forecasting?", "abstract": "Large language models (LLMs) are being applied to time series forecasting. But are language models actually useful for time series? In a series of ablation studies on three recent and popular LLM-based time series forecasting methods, we find that removing the LLM component or replacing it with a basic attention layer does not degrade forecasting performance -- in most cases, the results even improve! We also find that despite their significant computational cost, pretrained LLMs do no better than models trained from scratch, do not represent the sequential dependencies in time series, and do not assist in few-shot settings. Additionally, we explore time series encoders and find that patching and attention structures perform similarly to LLM-based forecasters.", "authors": ["Mingtian Tan", "Mike A. Merrill", "Vinayak Gupta", "Tim Althoff", "Thomas Hartvigsen"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-06-22", "url": "https://arxiv.org/abs/2406.16964", "pdf_url": "https://arxiv.org/pdf/2406.16964v2", "arxiv_id": "2406.16964", "doi": "10.48550/arXiv.2406.16964", "citation_count": 250, "influential_citation_count": 12, "has_code": false, "code_url": null, "venue": "Neural Information Processing Systems", "quality_score": 0.5999}
{"id": "6fb3a1fcadadb19168d33bebac2ba04bf4dd2a32cc82e55f290d1df35df24dde", "sources": ["arxiv", "semantic_scholar"], "title": "WindowMixer: Intra-Window and Inter-Window Modeling for Time Series Forecasting", "abstract": "Time series forecasting (TSF) is crucial in fields like economic forecasting, weather prediction, traffic flow analysis, and public health surveillance. Real-world time series data often include noise, outliers, and missing values, making accurate forecasting challenging. Traditional methods model point-to-point relationships, which limits their ability to capture complex temporal patterns and increases their susceptibility to noise.To address these issues, we introduce the WindowMixer model, built on an all-MLP framework. WindowMixer leverages the continuous nature of time series by examining temporal variations from a window-based perspective. It decomposes time series into trend and seasonal components, handling them individually. For trends, a fully connected (FC) layer makes predictions. For seasonal components, time windows are projected to produce window tokens, processed by Intra-Window-Mixer and Inter-Window-Mixer modules. The Intra-Window-Mixer models relationships within each window, while the Inter-Window-Mixer models relationships between windows. This approach captures intricate patterns and long-range dependencies in the data.Experiments show WindowMixer consistently outperforms existing methods in both long-term and short-term forecasting tasks.", "authors": ["Quangao Liu", "Ruiqi Li", "Maowei Jiang", "Wei Yang", "Chen Liang", "LongLong Pang", "Zhuozhang Zou"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-06-14", "url": "https://arxiv.org/abs/2406.12921", "pdf_url": "https://arxiv.org/pdf/2406.12921v2", "arxiv_id": "2406.12921", "doi": "10.48550/arXiv.2406.12921", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.0}
{"id": "dc98cf124a8405452153b4c2ef619daa8ad50cd2bd174f171590a8d848974508", "sources": ["arxiv", "semantic_scholar"], "title": "Fredformer: Frequency Debiased Transformer for Time Series Forecasting", "abstract": "The Transformer model has shown leading performance in time series forecasting. Nevertheless, in some complex scenarios, it tends to learn low-frequency features in the data and overlook high-frequency features, showing a frequency bias. This bias prevents the model from accurately capturing important high-frequency data features. In this paper, we undertook empirical analyses to understand this bias and discovered that frequency bias results from the model disproportionately focusing on frequency features with higher energy. Based on our analysis, we formulate this bias and propose Fredformer, a Transformer-based framework designed to mitigate frequency bias by learning features equally across different frequency bands. This approach prevents the model from overlooking lower amplitude features important for accurate forecasting. Extensive experiments show the effectiveness of our proposed approach, which can outperform other baselines in different real-world time-series datasets. Furthermore, we introduce a lightweight variant of the Fredformer with an attention matrix approximation, which achieves comparable performance but with much fewer parameters and lower computation costs. The code is available at: https://github.com/chenzRG/Fredformer", "authors": ["Xihao Piao", "Zheng Chen", "Taichi Murayama", "Yasuko Matsubara", "Yasushi Sakurai"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-06-13", "url": "https://arxiv.org/abs/2406.09009", "pdf_url": "https://arxiv.org/pdf/2406.09009v4", "arxiv_id": "2406.09009", "doi": "10.1145/3637528.3671928", "citation_count": 116, "influential_citation_count": 7, "has_code": true, "code_url": "https://github.com/chenzRG/Fredformer", "venue": "Knowledge Discovery and Data Mining", "quality_score": 0.517}
{"id": "aeb369519b09159d136b1d1887aeb64623a7dc832e4264d2de8490e6d2bc767a", "sources": ["arxiv", "semantic_scholar"], "title": "Time Series Analysis: yesterday, today, tomorrow", "abstract": "Forecasts of various processes have always been a sophisticated problem for statistics and data science. Over the past decades the solution procedures were updated by deep learning and kernel methods. According to many specialists, these approaches are much more precise, stable, and suitable compared to the classical statistical linear time series methods. Here we investigate how true this point of view is.", "authors": ["Igor Mackarov"], "categories": ["cs.CY"], "fields_of_study": ["Computer Science"], "published_date": "2024-06-10", "url": "https://arxiv.org/abs/2406.06453", "pdf_url": "https://arxiv.org/pdf/2406.06453v1", "arxiv_id": "2406.06453", "doi": "10.48550/arXiv.2406.06453", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.0}
{"id": "bd332d5fc079cf19a67882441808b50088d4b583aa35cecb19cd3aca98b124b5", "sources": ["arxiv", "semantic_scholar"], "title": "Data Augmentation for Multivariate Time Series Classification: An Experimental Study", "abstract": "Our study investigates the impact of data augmentation on the performance of multivariate time series models, focusing on datasets from the UCR archive. Despite the limited size of these datasets, we achieved classification accuracy improvements in 10 out of 13 datasets using the Rocket and InceptionTime models. This highlights the essential role of sufficient data in training effective models, paralleling the advancements seen in computer vision. Our work delves into adapting and applying existing methods in innovative ways to the domain of multivariate time series classification. Our comprehensive exploration of these techniques sets a new standard for addressing data scarcity in time series analysis, emphasizing that diverse augmentation strategies are crucial for unlocking the potential of both traditional and deep learning models. Moreover, by meticulously analyzing and applying a variety of augmentation techniques, we demonstrate that strategic data enrichment can enhance model accuracy. This not only establishes a benchmark for future research in time series analysis but also underscores the importance of adopting varied augmentation approaches to improve model performance in the face of limited data availability.", "authors": ["Romain Ilbert", "Thai V. Hoang", "Zonghua Zhang"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-06-10", "url": "https://arxiv.org/abs/2406.06518", "pdf_url": "https://arxiv.org/pdf/2406.06518v1", "arxiv_id": "2406.06518", "doi": "10.1109/ICDEW61823.2024.00023", "citation_count": 6, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.2113}
{"id": "d145f690dbcc10b525fae7ab6f401a71f43264b6463ed14707728996f0d4d0c2", "sources": ["arxiv", "semantic_scholar"], "title": "UniTST: Effectively Modeling Inter-Series and Intra-Series Dependencies for Multivariate Time Series Forecasting", "abstract": "Transformer-based models have emerged as powerful tools for multivariate time series forecasting (MTSF). However, existing Transformer models often fall short of capturing both intricate dependencies across variate and temporal dimensions in MTS data. Some recent models are proposed to separately capture variate and temporal dependencies through either two sequential or parallel attention mechanisms. However, these methods cannot directly and explicitly learn the intricate inter-series and intra-series dependencies. In this work, we first demonstrate that these dependencies are very important as they usually exist in real-world data. To directly model these dependencies, we propose a transformer-based model UniTST containing a unified attention mechanism on the flattened patch tokens. Additionally, we add a dispatcher module which reduces the complexity and makes the model feasible for a potentially large number of variates. Although our proposed model employs a simple architecture, it offers compelling performance as shown in our extensive experiments on several datasets for time series forecasting.", "authors": ["Juncheng Liu", "Chenghao Liu", "Gerald Woo", "Yiwei Wang", "Bryan Hooi", "Caiming Xiong", "Doyen Sahoo"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-06-07", "url": "https://arxiv.org/abs/2406.04975", "pdf_url": "https://arxiv.org/pdf/2406.04975v1", "arxiv_id": "2406.04975", "doi": "10.48550/arXiv.2406.04975", "citation_count": 17, "influential_citation_count": 4, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.3495}
{"id": "69b3001203c069684792eef7eb8ca482a7d5a158b6f615d44c0719fae649732e", "sources": ["arxiv", "semantic_scholar"], "title": "FPN-fusion: Enhanced Linear Complexity Time Series Forecasting Model", "abstract": "This study presents a novel time series prediction model, FPN-fusion, designed with linear computational complexity, demonstrating superior predictive performance compared to DLiner without increasing parameter count or computational demands. Our model introduces two key innovations: first, a Feature Pyramid Network (FPN) is employed to effectively capture time series data characteristics, bypassing the traditional decomposition into trend and seasonal components. Second, a multi-level fusion structure is developed to integrate deep and shallow features seamlessly. Empirically, FPN-fusion outperforms DLiner in 31 out of 32 test cases on eight open-source datasets, with an average reduction of 16.8% in mean squared error (MSE) and 11.8% in mean absolute error (MAE). Additionally, compared to the transformer-based PatchTST, FPN-fusion achieves 10 best MSE and 15 best MAE results, using only 8% of PatchTST's total computational load in the 32 test projects.", "authors": ["Chu Li", "Pingjia Xiao", "Qiping Yuan"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-06-06", "url": "https://arxiv.org/abs/2406.06603", "pdf_url": "https://arxiv.org/pdf/2406.06603v1", "arxiv_id": "2406.06603", "doi": "10.48550/arXiv.2406.06603", "citation_count": 2, "influential_citation_count": 0, "has_code": true, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1193}
{"id": "618acebdc52313f49b6bafec0dfe15874c6435fe5010ff6d7c4b85d820b30a13", "sources": ["arxiv", "semantic_scholar"], "title": "Time-Series Foundation Models for Forecasting Soil Moisture Levels in Smart Agriculture", "abstract": "The recent surge in foundation models for natural language processing and computer vision has fueled innovation across various domains. Inspired by this progress, we explore the potential of foundation models for time-series forecasting in smart agriculture, a field often plagued by limited data availability. Specifically, this work presents a novel application of $\\texttt{TimeGPT}$, a state-of-the-art (SOTA) time-series foundation model, to predict soil water potential ($ψ_\\mathrm{soil}$), a key indicator of field water status that is typically used for irrigation advice. Traditionally, this task relies on a wide array of input variables. We explore $ψ_\\mathrm{soil}$'s ability to forecast $ψ_\\mathrm{soil}$ in: ($i$) a zero-shot setting, ($ii$) a fine-tuned setting relying solely on historic $ψ_\\mathrm{soil}$ measurements, and ($iii$) a fine-tuned setting where we also add exogenous variables to the model. We compare $\\texttt{TimeGPT}$'s performance to established SOTA baseline models for forecasting $ψ_\\mathrm{soil}$. Our results demonstrate that $\\texttt{TimeGPT}$ achieves competitive forecasting accuracy using only historical $ψ_\\mathrm{soil}$ data, highlighting its remarkable potential for agricultural applications. This research paves the way for foundation time-series models for sustainable development in agriculture by enabling forecasting tasks that were traditionally reliant on extensive data collection and domain expertise.", "authors": ["Boje Deforce", "Bart Baesens", "Estefanía Serral Asensio"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-05-29", "url": "https://arxiv.org/abs/2405.18913", "pdf_url": "https://arxiv.org/pdf/2405.18913v3", "arxiv_id": "2405.18913", "doi": null, "citation_count": 9, "influential_citation_count": 2, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.25}
{"id": "6ba0e7d51be45d3a70bcc2b7d700f7466d08ef9c5d97c39493eca48b035d1ae8", "sources": ["arxiv", "semantic_scholar"], "title": "Are Self-Attentions Effective for Time Series Forecasting?", "abstract": "Time series forecasting is crucial for applications across multiple domains and various scenarios. Although Transformer models have dramatically advanced the landscape of forecasting, their effectiveness remains debated. Recent findings have indicated that simpler linear models might outperform complex Transformer-based approaches, highlighting the potential for more streamlined architectures. In this paper, we shift the focus from evaluating the overall Transformer architecture to specifically examining the effectiveness of self-attention for time series forecasting. To this end, we introduce a new architecture, Cross-Attention-only Time Series transformer (CATS), that rethinks the traditional Transformer framework by eliminating self-attention and leveraging cross-attention mechanisms instead. By establishing future horizon-dependent parameters as queries and enhanced parameter sharing, our model not only improves long-term forecasting accuracy but also reduces the number of parameters and memory usage. Extensive experiment across various datasets demonstrates that our model achieves superior performance with the lowest mean squared error and uses fewer parameters compared to existing models. The implementation of our model is available at: https://github.com/dongbeank/CATS.", "authors": ["Dongbin Kim", "Jinseong Park", "Jaewook Lee", "Hoki Kim"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-05-27", "url": "https://arxiv.org/abs/2405.16877", "pdf_url": "https://arxiv.org/pdf/2405.16877v3", "arxiv_id": "2405.16877", "doi": "10.48550/arXiv.2405.16877", "citation_count": 34, "influential_citation_count": 2, "has_code": true, "code_url": "https://github.com/dongbeank/CATS", "venue": "Neural Information Processing Systems", "quality_score": 0.386}
{"id": "d867ed1a4abccd38e113ae43fd56cb4a9b392c33473239e8203026a9f62112ce", "sources": ["arxiv", "semantic_scholar"], "title": "Dominant Shuffle: A Simple Yet Powerful Data Augmentation for Time-series Prediction", "abstract": "Recent studies have suggested frequency-domain Data augmentation (DA) is effec tive for time series prediction. Existing frequency-domain augmentations disturb the original data with various full-spectrum noises, leading to excess domain gap between augmented and original data. Although impressive performance has been achieved in certain cases, frequency-domain DA has yet to be generalized to time series prediction datasets. In this paper, we found that frequency-domain augmentations can be significantly improved by two modifications that limit the perturbations. First, we found that limiting the perturbation to only dominant frequencies significantly outperforms full-spectrum perturbations. Dominant fre quencies represent the main periodicity and trends of the signal and are more important than other frequencies. Second, we found that simply shuffling the dominant frequency components is superior over sophisticated designed random perturbations. Shuffle rearranges the original components (magnitudes and phases) and limits the external noise. With these two modifications, we proposed dominant shuffle, a simple yet effective data augmentation for time series prediction. Our method is very simple yet powerful and can be implemented with just a few lines of code. Extensive experiments with eight datasets and six popular time series models demonstrate that our method consistently improves the baseline performance under various settings and significantly outperforms other DA methods. Code can be accessed at https://kaizhao.net/time-series.", "authors": ["Kai Zhao", "Zuojie He", "Alex Hung", "Dan Zeng"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-05-26", "url": "https://arxiv.org/abs/2405.16456", "pdf_url": "https://arxiv.org/pdf/2405.16456v1", "arxiv_id": "2405.16456", "doi": "10.48550/arXiv.2405.16456", "citation_count": 5, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1945}
{"id": "eb06ce3a05b7ac1e90eaa7bb5d9f2cb777d82c9776d28ac2d6a90d23c386610f", "sources": ["arxiv", "semantic_scholar"], "title": "Scaling Law for Time Series Forecasting", "abstract": "Scaling law that rewards large datasets, complex models and enhanced data granularity has been observed in various fields of deep learning. Yet, studies on time series forecasting have cast doubt on scaling behaviors of deep learning methods for time series forecasting: while more training data improves performance, more capable models do not always outperform less capable models, and longer input horizons may hurt performance for some models. We propose a theory for scaling law for time series forecasting that can explain these seemingly abnormal behaviors. We take into account the impact of dataset size and model complexity, as well as time series data granularity, particularly focusing on the look-back horizon, an aspect that has been unexplored in previous theories. Furthermore, we empirically evaluate various models using a diverse set of time series forecasting datasets, which (1) verifies the validity of scaling law on dataset size and model complexity within the realm of time series forecasting, and (2) validates our theoretical framework, particularly regarding the influence of look back horizon. We hope our findings may inspire new models targeting time series forecasting datasets of limited size, as well as large foundational datasets and models for time series forecasting in future work. Code for our experiments has been made public at https://github.com/JingzheShi/ScalingLawForTimeSeriesForecasting.", "authors": ["Jingzhe Shi", "Qinwei Ma", "Huan Ma", "Lei Li"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-05-24", "url": "https://arxiv.org/abs/2405.15124", "pdf_url": "https://arxiv.org/pdf/2405.15124v4", "arxiv_id": "2405.15124", "doi": "10.48550/arXiv.2405.15124", "citation_count": 30, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/JingzheShi/ScalingLawForTimeSeriesForecasting", "venue": "Neural Information Processing Systems", "quality_score": 0.3728}
{"id": "55e9931183a3afd594cf7d2beabd985368fd32762529f9343ce6cf5e174d75de", "sources": ["arxiv", "semantic_scholar"], "title": "Towards a General Time Series Forecasting Model with Unified Representation and Adaptive Transfer", "abstract": "With the growing availability of multi-domain time series data, there is an increasing demand for general forecasting models pre-trained on multi-source datasets to support diverse downstream prediction scenarios. Existing time series foundation models primarily focus on scaling up pre-training datasets and model sizes to enhance generalization performance. In this paper, we take a different approach by addressing two critical aspects of general forecasting models: (1) how to derive unified representations from heterogeneous multi-domain time series data, and (2) how to effectively capture domain-specific features to enable adaptive transfer across various downstream scenarios. To address the first aspect, we propose Decomposed Frequency Learning as the pre-training task, which leverages frequency-based masking and reconstruction to decompose coupled semantic information in time series, resulting in unified representations across domains. For the second aspect, we introduce the Time Series Register, which captures domain-specific representations during pre-training and enhances adaptive transferability to downstream tasks. Our model achieves the state-of-the-art forecasting performance on seven real-world benchmarks, demonstrating remarkable few-shot and zero-shot capabilities.", "authors": ["Yihang Wang", "Yuying Qiu", "Peng Chen", "Kai Zhao", "Yang Shu", "Zhongwen Rao", "Lujia Pan", "Bin Yang", "Chenjuan Guo"], "categories": ["cs.LG", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2024-05-24", "url": "https://arxiv.org/abs/2405.17478", "pdf_url": "https://arxiv.org/pdf/2405.17478v3", "arxiv_id": "2405.17478", "doi": null, "citation_count": 34, "influential_citation_count": 2, "has_code": false, "code_url": null, "venue": "International Conference on Machine Learning", "quality_score": 0.386}
{"id": "d173725fd8e6d5ed81f912dbe1487738f779d7af662b8e0bd2fdcd79a6830dcb", "sources": ["arxiv", "semantic_scholar"], "title": "Time-FFM: Towards LM-Empowered Federated Foundation Model for Time Series Forecasting", "abstract": "Unlike natural language processing and computer vision, the development of Foundation Models (FMs) for time series forecasting is blocked due to data scarcity. While recent efforts are focused on building such FMs by unlocking the potential of language models (LMs) for time series analysis, dedicated parameters for various downstream forecasting tasks need training, which hinders the common knowledge sharing across domains. Moreover, data owners may hesitate to share the access to local data due to privacy concerns and copyright protection, which makes it impossible to simply construct a FM on cross-domain training instances. To address these issues, we propose Time-FFM, a Federated Foundation Model for Time series forecasting by leveraging pretrained LMs. Specifically, we begin by transforming time series into the modality of text tokens. To bootstrap LMs for time series reasoning, we propose a prompt adaption module to determine domain-customized prompts dynamically instead of artificially. Given the data heterogeneity across domains, we design a personalized federated training strategy by learning global encoders and local prediction heads. Our comprehensive experiments indicate that Time-FFM outperforms state-of-the-arts and promises effective few-shot and zero-shot forecaster.", "authors": ["Qingxiang Liu", "Xu Liu", "Chenghao Liu", "Qingsong Wen", "Yuxuan Liang"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-05-23", "url": "https://arxiv.org/abs/2405.14252", "pdf_url": "https://arxiv.org/pdf/2405.14252v4", "arxiv_id": "2405.14252", "doi": "10.48550/arXiv.2405.14252", "citation_count": 33, "influential_citation_count": 5, "has_code": false, "code_url": null, "venue": "Neural Information Processing Systems", "quality_score": 0.3891}
{"id": "345de470b4fe2d304ed7fc2da09cd6598d823c40bd4ce72b56c39a1b912f6d3f", "sources": ["arxiv", "semantic_scholar"], "title": "Scaling-laws for Large Time-series Models", "abstract": "Scaling laws for large language models (LLMs) have provided useful guidance in training ever larger models for predictable performance gains. Time series forecasting shares a similar sequential structure to language, and is amenable to large-scale transformer architectures. Here we show that foundational decoder-only time series transformer models exhibit analogous scaling-behavior to LLMs, with architectural details (aspect ratio and number of heads) having a minimal effect over broad ranges. We assemble a large corpus of heterogenous time series data on which to train, and establish for the first time power-law scaling with parameter count, dataset size, and training compute, spanning five orders of magnitude.", "authors": ["Thomas D. P. Edwards", "James Alvey", "Justin Alsing", "Nam H. Nguyen", "Benjamin D. Wandelt"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-05-22", "url": "https://arxiv.org/abs/2405.13867", "pdf_url": "https://arxiv.org/pdf/2405.13867v2", "arxiv_id": "2405.13867", "doi": "10.48550/arXiv.2405.13867", "citation_count": 18, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.3197}
{"id": "4d8f8ff164337f0a6feb1dc5dca87d864631a5668771f9fec6cc40cd54d3aa99", "sources": ["arxiv", "semantic_scholar"], "title": "Interpretable Multivariate Time Series Forecasting Using Neural Fourier Transform", "abstract": "Multivariate time series forecasting is a pivotal task in several domains, including financial planning, medical diagnostics, and climate science. This paper presents the Neural Fourier Transform (NFT) algorithm, which combines multi-dimensional Fourier transforms with Temporal Convolutional Network layers to improve both the accuracy and interpretability of forecasts. The Neural Fourier Transform is empirically validated on fourteen diverse datasets, showing superior performance across multiple forecasting horizons and lookbacks, setting new benchmarks in the field. This work advances multivariate time series forecasting by providing a model that is both interpretable and highly predictive, making it a valuable tool for both practitioners and researchers. The code for this study is publicly available.", "authors": ["Noam Koren", "Kira Radinsky"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-05-22", "url": "https://arxiv.org/abs/2405.13812", "pdf_url": "https://arxiv.org/pdf/2405.13812v1", "arxiv_id": "2405.13812", "doi": "10.48550/arXiv.2405.13812", "citation_count": 4, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1747}
{"id": "754d28a2a6230ed6da468cda812e8246fb241f9a207cdf485686c6edf9d2c431", "sources": ["arxiv", "semantic_scholar"], "title": "Leveraging 2D Information for Long-term Time Series Forecasting with Vanilla Transformers", "abstract": "Time series prediction is crucial for understanding and forecasting complex dynamics in various domains, ranging from finance and economics to climate and healthcare. Based on Transformer architecture, one approach involves encoding multiple variables from the same timestamp into a single temporal token to model global dependencies. In contrast, another approach embeds the time points of individual series into separate variate tokens. The former method faces challenges in learning variate-centric representations, while the latter risks missing essential temporal information critical for accurate forecasting. In our work, we introduce GridTST, a model that combines the benefits of two approaches using innovative multi-directional attentions based on a vanilla Transformer. We regard the input time series data as a grid, where the $x$-axis represents the time steps and the $y$-axis represents the variates. A vertical slicing of this grid combines the variates at each time step into a \\textit{time token}, while a horizontal slicing embeds the individual series across all time steps into a \\textit{variate token}. Correspondingly, a \\textit{horizontal attention mechanism} focuses on time tokens to comprehend the correlations between data at various time steps, while a \\textit{vertical}, variate-aware \\textit{attention} is employed to grasp multivariate correlations. This combination enables efficient processing of information across both time and variate dimensions, thereby enhancing the model's analytical strength. % We also integrate the patch technique, segmenting time tokens into subseries-level patches, ensuring that local semantic information is retained in the embedding. The GridTST model consistently delivers state-of-the-art performance across various real-world datasets.", "authors": ["Xin Cheng", "Xiuying Chen", "Shuqi Li", "Di Luo", "Xun Wang", "Dongyan Zhao", "Rui Yan"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-05-22", "url": "https://arxiv.org/abs/2405.13810", "pdf_url": "https://arxiv.org/pdf/2405.13810v1", "arxiv_id": "2405.13810", "doi": "10.48550/arXiv.2405.13810", "citation_count": 5, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1945}
{"id": "1e59a88c41aff535377e9ae4d5e7b182bcb232c54a1ed9a506d35d3ca807a5a3", "sources": ["arxiv", "semantic_scholar"], "title": "Enhancing Transformer-based models for Long Sequence Time Series Forecasting via Structured Matrix", "abstract": "Recently, Transformer-based models for long sequence time series forecasting have demonstrated promising results. The self-attention mechanism as the core component of these Transformer-based models exhibits great potential in capturing various dependencies among data points. Despite these advancements, it has been a subject of concern to improve the efficiency of the self-attention mechanism. Unfortunately, current specific optimization methods are facing the challenges in applicability and scalability for the future design of long sequence time series forecasting models. Hence, in this article, we propose a novel architectural framework that enhances Transformer-based models through the integration of Surrogate Attention Blocks (SAB) and Surrogate Feed-Forward Neural Network Blocks (SFB). The framework reduces both time and space complexity by the replacement of the self-attention and feed-forward layers with SAB and SFB while maintaining their expressive power and architectural advantages. The equivalence of this substitution is fully demonstrated. The extensive experiments on 10 Transformer-based models across five distinct time series tasks demonstrate an average performance improvement of 12.4%, alongside 61.3% reduction in parameter counts.", "authors": ["Zhicheng Zhang", "Yong Wang", "Shaoqi Tan", "Bowei Xia", "Yujie Luo"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-05-21", "url": "https://arxiv.org/abs/2405.12462", "pdf_url": "https://arxiv.org/pdf/2405.12462v4", "arxiv_id": "2405.12462", "doi": null, "citation_count": 6, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.2113}
{"id": "e32d9771ba82753759b88a8bc2dfd4ec5ff4d19e4207dbcd84c7e59fcbb41fe5", "sources": ["arxiv", "semantic_scholar"], "title": "Low-Rank Adaptation of Time Series Foundational Models for Out-of-Domain Modality Forecasting", "abstract": "Low-Rank Adaptation (LoRA) is a widely used technique for fine-tuning large pre-trained or foundational models across different modalities and tasks. However, its application to time series data, particularly within foundational models, remains underexplored. This paper examines the impact of LoRA on contemporary time series foundational models: Lag-Llama, MOIRAI, and Chronos. We demonstrate LoRA's fine-tuning potential for forecasting the vital signs of sepsis patients in intensive care units (ICUs), emphasizing the models' adaptability to previously unseen, out-of-domain modalities. Integrating LoRA aims to enhance forecasting performance while reducing inefficiencies associated with fine-tuning large models on limited domain-specific data. Our experiments show that LoRA fine-tuning of time series foundational models significantly improves forecasting, achieving results comparable to state-of-the-art models trained from scratch on similar modalities. We conduct comprehensive ablation studies to demonstrate the trade-offs between the number of tunable parameters and forecasting performance and assess the impact of varying LoRA matrix ranks on model performance.", "authors": ["Divij Gupta", "Anubhav Bhatti", "Suraj Parmar", "Chen Dan", "Yuwei Liu", "Bingjie Shen", "San Lee"], "categories": ["cs.LG", "cs.AI", "eess.SP"], "fields_of_study": ["Computer Science", "Engineering"], "published_date": "2024-05-16", "url": "https://arxiv.org/abs/2405.10216", "pdf_url": "https://arxiv.org/pdf/2405.10216v1", "arxiv_id": "2405.10216", "doi": "10.1145/3678957.3685724", "citation_count": 22, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "International Conference on Multimodal Interaction", "quality_score": 0.3404}
{"id": "7e9ce160019703e644304761e7ad3a20ecb25167bad795725f081e2351588371", "sources": ["arxiv", "semantic_scholar"], "title": "DGCformer: Deep Graph Clustering Transformer for Multivariate Time Series Forecasting", "abstract": "Multivariate time series forecasting tasks are usually conducted in a channel-dependent (CD) way since it can incorporate more variable-relevant information. However, it may also involve a lot of irrelevant variables, and this even leads to worse performance than the channel-independent (CI) strategy. This paper combines the strengths of both strategies and proposes the Deep Graph Clustering Transformer (DGCformer) for multivariate time series forecasting. Specifically, it first groups these relevant variables by a graph convolutional network integrated with an autoencoder, and a former-latter masked self-attention mechanism is then considered with the CD strategy being applied to each group of variables while the CI one for different groups. Extensive experimental results on eight datasets demonstrate the superiority of our method against state-of-the-art models, and our code will be publicly available upon acceptance.", "authors": ["Qinshuo Liu", "Yanwen Fang", "Pengtao Jiang", "Guodong Li"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-05-14", "url": "https://arxiv.org/abs/2405.08440", "pdf_url": "https://arxiv.org/pdf/2405.08440v1", "arxiv_id": "2405.08440", "doi": "10.48550/arXiv.2405.08440", "citation_count": 5, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1945}
{"id": "88f0a3bf1c251bcdbf95da1647b8a23b9bce673f19d2ef9db2ea2614d1e72433", "sources": ["arxiv", "semantic_scholar"], "title": "Forecasting with Hyper-Trees", "abstract": "We introduce Hyper-Trees as a novel framework for modeling time series data using gradient boosted trees. Unlike conventional tree-based approaches that forecast time series directly, Hyper-Trees learn the parameters of a target time series model, such as ARIMA or Exponential Smoothing, as functions of features. These parameters are then used by the target model to generate the final forecasts. Our framework combines the effectiveness of decision trees on tabular data with classical forecasting models, thereby inducing a time series inductive bias into tree-based models. To resolve the scaling limitations of boosted trees when estimating a high-dimensional set of target model parameters, we combine decision trees and neural networks within a unified framework. In this hybrid approach, the trees generate informative representations from the input features, which a shallow network then uses as input to learn the parameters of a time series model. With our research, we explore the effectiveness of Hyper-Trees across a range of forecasting tasks and extend tree-based modeling beyond its conventional use in time series analysis.", "authors": ["Alexander März", "Kashif Rasul"], "categories": ["cs.LG", "stat.ME"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2024-05-13", "url": "https://arxiv.org/abs/2405.07836", "pdf_url": "https://arxiv.org/pdf/2405.07836v5", "arxiv_id": "2405.07836", "doi": "10.48550/arXiv.2405.07836", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.0}
{"id": "29b0b31264d4818d97dac174ba427fb63cce8b04d7ec61ea3e7c35c087e91902", "sources": ["arxiv", "semantic_scholar"], "title": "ReCycle: Fast and Efficient Long Time Series Forecasting with Residual Cyclic Transformers", "abstract": "Transformers have recently gained prominence in long time series forecasting by elevating accuracies in a variety of use cases. Regrettably, in the race for better predictive performance the overhead of model architectures has grown onerous, leading to models with computational demand infeasible for most practical applications. To bridge the gap between high method complexity and realistic computational resources, we introduce the Residual Cyclic Transformer, ReCycle. ReCycle utilizes primary cycle compression to address the computational complexity of the attention mechanism in long time series. By learning residuals from refined smoothing average techniques, ReCycle surpasses state-of-the-art accuracy in a variety of application use cases. The reliable and explainable fallback behavior ensured by simple, yet robust, smoothing average techniques additionally lowers the barrier for user acceptance. At the same time, our approach reduces the run time and energy consumption by more than an order of magnitude, making both training and inference feasible on low-performance, low-power and edge computing devices. Code is available at https://github.com/Helmholtz-AI-Energy/ReCycle", "authors": ["Arvid Weyrauch", "Thomas Steens", "Oskar Taubert", "Benedikt Hanke", "Aslan Eqbal", "Ewa Götz", "Achim Streit", "Markus Götz", "Charlotte Debus"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-05-06", "url": "https://arxiv.org/abs/2405.03429", "pdf_url": "https://arxiv.org/pdf/2405.03429v1", "arxiv_id": "2405.03429", "doi": "10.1109/CAI59869.2024.00212", "citation_count": 2, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/Helmholtz-AI-Energy/ReCycle", "venue": "Conference on Algebraic Informatics", "quality_score": 0.1193}
{"id": "f67594f4d374c1c58aa4002aa6f4293bf71804b65cad6a7750bff2fc30b53c2f", "sources": ["arxiv", "semantic_scholar"], "title": "CVTN: Cross Variable and Temporal Integration for Time Series Forecasting", "abstract": "In multivariate time series forecasting, the Transformer architecture encounters two significant challenges: effectively mining features from historical sequences and avoiding overfitting during the learning of temporal dependencies. To tackle these challenges, this paper deconstructs time series forecasting into the learning of historical sequences and prediction sequences, introducing the Cross-Variable and Time Network (CVTN). This unique method divides multivariate time series forecasting into two phases: cross-variable learning for effectively mining fea tures from historical sequences, and cross-time learning to capture the temporal dependencies of prediction sequences. Separating these two phases helps avoid the impact of overfitting in cross-time learning on cross-variable learning. Exten sive experiments on various real-world datasets have confirmed its state-of-the-art (SOTA) performance. CVTN emphasizes three key dimensions in time series fore casting: the short-term and long-term nature of time series (locality and longevity), feature mining from both historical and prediction sequences, and the integration of cross-variable and cross-time learning. This approach not only advances the current state of time series forecasting but also provides a more comprehensive framework for future research in this field.", "authors": ["Han Zhou", "Yuntian Chen"], "categories": ["cs.LG", "cs.AI", "stat.AP"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2024-04-29", "url": "https://arxiv.org/abs/2404.18730", "pdf_url": "https://arxiv.org/pdf/2404.18730v1", "arxiv_id": "2404.18730", "doi": "10.48550/arXiv.2404.18730", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.0}
{"id": "d436f61df850c5fd192b04fe4bcd6ed2b3ec824eb5f92aae104681c4edbcd696", "sources": ["arxiv", "semantic_scholar"], "title": "SST: Multi-Scale Hybrid Mamba-Transformer Experts for Time Series Forecasting", "abstract": "Time series forecasting has made significant advances, including with Transformer-based models. The attention mechanism in Transformer effectively captures temporal dependencies by attending to all past inputs simultaneously. However, its quadratic complexity with respect to sequence length limits the scalability for long-range modeling. Recent state space models (SSMs) such as Mamba offer a promising alternative by achieving linear complexity without attention. Yet, Mamba compresses historical information into a fixed-size latent state, potentially causing information loss and limiting representational effectiveness. This raises a key research question: Can we design a hybrid Mamba-Transformer architecture that is both effective and efficient for time series forecasting? To address it, we adapt a hybrid Mamba-Transformer architecture Mambaformer, originally proposed for language modeling, to the time series domain. Preliminary experiments reveal that naively stacking Mamba and Transformer layers in Mambaformer is suboptimal for time series forecasting, due to an information interference problem. To mitigate this issue, we introduce a new time series decomposition strategy that separates time series into long-range patterns and short-range variations. Then we show that Mamba excels at capturing long-term structures, while Transformer is more effective at modeling short-term dynamics. Building on this insight, we propose State Space Transformer (SST), a multi-scale hybrid model with expert modules: a Mamba expert for long-range patterns and a Transformer expert for short-term variations. SST also employs a multi-scale patching mechanism to adaptively adjust time series resolution: low resolution for long-term patterns and high resolution for short-term variations. Experiments show that SST obtains SOTA performance with linear scalability. The code is at https://github.com/XiongxiaoXu/SST.", "authors": ["Xiongxiao Xu", "Canyu Chen", "Yueqing Liang", "Baixiang Huang", "Guangji Bai", "Liang Zhao", "Kai Shu"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-04-23", "url": "https://arxiv.org/abs/2404.14757", "pdf_url": "https://arxiv.org/pdf/2404.14757v3", "arxiv_id": "2404.14757", "doi": "10.1145/3746252.3761394", "citation_count": 33, "influential_citation_count": 1, "has_code": true, "code_url": "https://github.com/XiongxiaoXu/SST", "venue": "International Conference on Information and Knowledge Management", "quality_score": 0.3829}
{"id": "b72a099b81d00446158e21e8457fc13ebe1ee14216c567f092b276a09a34b778", "sources": ["arxiv", "semantic_scholar"], "title": "TimeGPT in Load Forecasting: A Large Time Series Model Perspective", "abstract": "Machine learning models have made significant progress in load forecasting, but their forecast accuracy is limited in cases where historical load data is scarce. Inspired by the outstanding performance of large language models (LLMs) in computer vision and natural language processing, this paper aims to discuss the potential of large time series models in load forecasting with scarce historical data. Specifically, the large time series model is constructed as a time series generative pre-trained transformer (TimeGPT), which is trained on massive and diverse time series datasets consisting of 100 billion data points (e.g., finance, transportation, banking, web traffic, weather, energy, healthcare, etc.). Then, the scarce historical load data is used to fine-tune the TimeGPT, which helps it to adapt to the data distribution and characteristics associated with load forecasting. Simulation results show that TimeGPT outperforms the benchmarks (e.g., popular machine learning models and statistical models) for load forecasting on several real datasets with scarce training samples, particularly for short look-ahead times. However, it cannot be guaranteed that TimeGPT is always superior to benchmarks for load forecasting with scarce data, since the performance of TimeGPT may be affected by the distribution differences between the load data and the training data. In practical applications, we can divide the historical data into a training set and a validation set, and then use the validation set loss to decide whether TimeGPT is the best choice for a specific dataset.", "authors": ["Wenlong Liao", "Fernando Porte-Agel", "Jiannong Fang", "Christian Rehtanz", "Shouxiang Wang", "Dechang Yang", "Zhe Yang"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-04-07", "url": "https://arxiv.org/abs/2404.04885", "pdf_url": "https://arxiv.org/pdf/2404.04885v2", "arxiv_id": "2404.04885", "doi": "10.1016/j.apenergy.2024.124973", "citation_count": 69, "influential_citation_count": 2, "has_code": false, "code_url": null, "venue": "Applied Energy", "quality_score": 0.4613}
{"id": "9ff4d15ff40b4096ddf81c5eb1bccd7e87dd6a07b84e7484f914571b15b8b834", "sources": ["arxiv", "semantic_scholar"], "title": "An Analysis of Linear Time Series Forecasting Models", "abstract": "Despite their simplicity, linear models perform well at time series forecasting, even when pitted against deeper and more expensive models. A number of variations to the linear model have been proposed, often including some form of feature normalisation that improves model generalisation. In this paper we analyse the sets of functions expressible using these linear model architectures. In so doing we show that several popular variants of linear models for time series forecasting are equivalent and functionally indistinguishable from standard, unconstrained linear regression. We characterise the model classes for each linear variant. We demonstrate that each model can be reinterpreted as unconstrained linear regression over a suitably augmented feature set, and therefore admit closed-form solutions when using a mean-squared loss function. We provide experimental evidence that the models under inspection learn nearly identical solutions, and finally demonstrate that the simpler closed form solutions are superior forecasters across 72% of test settings.", "authors": ["William Toner", "Luke Darlow"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-03-21", "url": "https://arxiv.org/abs/2403.14587", "pdf_url": "https://arxiv.org/pdf/2403.14587v2", "arxiv_id": "2403.14587", "doi": "10.48550/arXiv.2403.14587", "citation_count": 55, "influential_citation_count": 3, "has_code": false, "code_url": null, "venue": "International Conference on Machine Learning", "quality_score": 0.437}
{"id": "abf9349e148c01cde4f627decfed6a93ba1582fadd4d839b514e1083ff3a7428", "sources": ["arxiv", "semantic_scholar"], "title": "Leveraging Non-Decimated Wavelet Packet Features and Transformer Models for Time Series Forecasting", "abstract": "This article combines wavelet analysis techniques with machine learning methods for univariate time series forecasting, focusing on three main contributions. Firstly, we consider the use of Daubechies wavelets with different numbers of vanishing moments as input features to both non-temporal and temporal forecasting methods, by selecting these numbers during the cross-validation phase. Secondly, we compare the use of both the non-decimated wavelet transform and the non-decimated wavelet packet transform for computing these features, the latter providing a much larger set of potentially useful coefficient vectors. The wavelet coefficients are computed using a shifted version of the typical pyramidal algorithm to ensure no leakage of future information into these inputs. Thirdly, we evaluate the use of these wavelet features on a significantly wider set of forecasting methods than previous studies, including both temporal and non-temporal models, and both statistical and deep learning-based methods. The latter include state-of-the-art transformer-based neural network architectures. Our experiments suggest significant benefit in replacing higher-order lagged features with wavelet features across all examined non-temporal methods for one-step-forward forecasting, and modest benefit when used as inputs for temporal deep learning-based models for long-horizon forecasting.", "authors": ["Guy P Nason", "James L. Wei"], "categories": ["stat.ME", "cs.LG"], "fields_of_study": ["Mathematics", "Computer Science"], "published_date": "2024-03-13", "url": "https://arxiv.org/abs/2403.08630", "pdf_url": "https://arxiv.org/pdf/2403.08630v1", "arxiv_id": "2403.08630", "doi": "10.48550/arXiv.2403.08630", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.0753}
{"id": "5d126ccda7afcd777b528ec98f202aec52da1ea20dee165811685a6dd229188c", "sources": ["arxiv", "semantic_scholar"], "title": "Chronos: Learning the Language of Time Series", "abstract": "We introduce Chronos, a simple yet effective framework for pretrained probabilistic time series models. Chronos tokenizes time series values using scaling and quantization into a fixed vocabulary and trains existing transformer-based language model architectures on these tokenized time series via the cross-entropy loss. We pretrained Chronos models based on the T5 family (ranging from 20M to 710M parameters) on a large collection of publicly available datasets, complemented by a synthetic dataset that we generated via Gaussian processes to improve generalization. In a comprehensive benchmark consisting of 42 datasets, and comprising both classical local models and deep learning methods, we show that Chronos models: (a) significantly outperform other methods on datasets that were part of the training corpus; and (b) have comparable and occasionally superior zero-shot performance on new datasets, relative to methods that were trained specifically on them. Our results demonstrate that Chronos models can leverage time series data from diverse domains to improve zero-shot accuracy on unseen forecasting tasks, positioning pretrained models as a viable tool to greatly simplify forecasting pipelines.", "authors": ["Abdul Fatir Ansari", "Lorenzo Stella", "Caner Turkmen", "Xiyuan Zhang", "Pedro Mercado", "Huibin Shen", "Oleksandr Shchur", "Syama Sundar Rangapuram", "Sebastian Pineda Arango", "Shubham Kapoor", "Jasper Zschiegner", "Danielle C. Maddix", "Hao Wang", "Michael W. Mahoney", "Kari Torkkola", "Andrew Gordon Wilson", "Michael Bohlke-Schneider", "Yuyang Wang"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-03-12", "url": "https://arxiv.org/abs/2403.07815", "pdf_url": "https://arxiv.org/pdf/2403.07815v3", "arxiv_id": "2403.07815", "doi": "10.48550/arXiv.2403.07815", "citation_count": 790, "influential_citation_count": 162, "has_code": true, "code_url": "https://github.com/amazon-science/chronos-forecasting", "venue": null, "quality_score": 1.0}
{"id": "5716178a6ce3a64b49b758de7ebe44d8aa43a51bebbb3ab5aee6710e73c37e47", "sources": ["arxiv", "semantic_scholar"], "title": "A Segmentation Foundation Model for Diverse-type Tumors", "abstract": "Large pre-trained models with their numerous model parameters and extensive training datasets have shown excellent performance in various tasks. Many publicly available medical image datasets do not have a sufficient amount of data so there are few large-scale models in medical imaging. We propose a large-scale Tumor Segmentation Foundation Model (TSFM) with 1.6 billion parameters using Resblock-backbone and Transformer-bottleneck,which has good transfer ability for downstream tasks. To make TSFM exhibit good performance in tumor segmentation, we make full use of the strong spatial correlation between tumors and organs in the medical image, innovatively fuse 7 tumor datasets and 3 multi-organ datasets to build a 3D medical dataset pool, including 2779 cases with totally 300k medical images, whose size currently exceeds many other single publicly available datasets. TSFM is the pre-trained model for medical image segmentation, which also can be transferred to multiple downstream tasks for fine-tuning learning. The average performance of our pre-trained model is 2% higher than that of nnU-Net across various tumor types. In the transfer learning task, TSFM only needs 5% training epochs of nnU-Net to achieve similar performance and can surpass nnU-Net by 2% on average with 10% training epoch. Pre-trained TSFM and its code will be released soon.", "authors": ["Jianhao Xie", "Ziang Zhang", "Guibo Luo", "Yuesheng Zhu"], "categories": ["eess.IV", "cs.CV"], "fields_of_study": ["Engineering", "Computer Science"], "published_date": "2024-03-11", "url": "https://arxiv.org/abs/2403.06396", "pdf_url": "https://arxiv.org/pdf/2403.06396v1", "arxiv_id": "2403.06396", "doi": "10.48550/arXiv.2403.06396", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.0}
{"id": "bdcbf3804284d88d30e757e9d134ac1648c85bc50d9ff126c0a052d46b85a02b", "sources": ["arxiv", "semantic_scholar"], "title": "Considering Nonstationary within Multivariate Time Series with Variational Hierarchical Transformer for Forecasting", "abstract": "The forecasting of Multivariate Time Series (MTS) has long been an important but challenging task. Due to the non-stationary problem across long-distance time steps, previous studies primarily adopt stationarization method to attenuate the non-stationary problem of the original series for better predictability. However, existing methods always adopt the stationarized series, which ignores the inherent non-stationarity, and has difficulty in modeling MTS with complex distributions due to the lack of stochasticity. To tackle these problems, we first develop a powerful hierarchical probabilistic generative module to consider the non-stationarity and stochastic characteristics within MTS, and then combine it with transformer for a well-defined variational generative dynamic model named Hierarchical Time series Variational Transformer (HTV-Trans), which recovers the intrinsic non-stationary information into temporal dependencies. Being a powerful probabilistic model, HTV-Trans is utilized to learn expressive representations of MTS and applied to forecasting tasks. Extensive experiments on diverse datasets show the efficiency of HTV-Trans on MTS forecasting tasks", "authors": ["Muyao Wang", "Wenchao Chen", "Bo Chen"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-03-08", "url": "https://arxiv.org/abs/2403.05406", "pdf_url": "https://arxiv.org/pdf/2403.05406v1", "arxiv_id": "2403.05406", "doi": "10.1609/aaai.v38i14.29483", "citation_count": 13, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "AAAI Conference on Artificial Intelligence", "quality_score": 0.2865}
{"id": "ae4058f9da33e2ffda62744456155df5926cad3a2a1aa0279c63a619dab2ba48", "sources": ["arxiv", "semantic_scholar"], "title": "Hyperparameter Tuning MLPs for Probabilistic Time Series Forecasting", "abstract": "Time series forecasting attempts to predict future events by analyzing past trends and patterns. Although well researched, certain critical aspects pertaining to the use of deep learning in time series forecasting remain ambiguous. Our research primarily focuses on examining the impact of specific hyperparameters related to time series, such as context length and validation strategy, on the performance of the state-of-the-art MLP model in time series forecasting. We have conducted a comprehensive series of experiments involving 4800 configurations per dataset across 20 time series forecasting datasets, and our findings demonstrate the importance of tuning these parameters. Furthermore, in this work, we introduce the largest metadataset for timeseries forecasting to date, named TSBench, comprising 97200 evaluations, which is a twentyfold increase compared to previous works in the field. Finally, we demonstrate the utility of the created metadataset on multi-fidelity hyperparameter optimization tasks.", "authors": ["Kiran Madhusudhanan", "Shayan Jawed", "Lars Schmidt-Thieme"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-03-07", "url": "https://arxiv.org/abs/2403.04477", "pdf_url": "https://arxiv.org/pdf/2403.04477v1", "arxiv_id": "2403.04477", "doi": "10.48550/arXiv.2403.04477", "citation_count": 2, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "Pacific-Asia Conference on Knowledge Discovery and Data Mining", "quality_score": 0.1193}
{"id": "af35a3d8b8f415505debc67c5e8d58fc1133ed6c6883794c38ca0ea3729bfe6a", "sources": ["arxiv", "semantic_scholar"], "title": "EXPRTS: Exploring and Probing the Robustness of Time Series Forecasting Models", "abstract": "When deploying time series forecasting models based on machine learning to real world settings, one often encounter situations where the data distribution drifts. Such drifts expose the forecasting models to out-of-distribution (OOD) data, and machine learning models lack robustness in these settings. Robustness can be improved by using deep generative models or genetic algorithms to augment time series datasets, but these approaches lack interpretability and are computationally expensive. In this work, we develop an interpretable and simple framework for generating time series. Our method combines time-series decompositions with analytic functions, and is able to generate time series with characteristics matching both in- and out-of-distribution data. This approach allows users to generate new time series in an interpretable fashion, which can be used to augment the dataset and improve forecasting robustness. We demonstrate our framework through EXPRTS, a visual analytics tool designed for univariate time series forecasting models and datasets. Different visualizations of the data distribution, forecasting errors and single time series instances enable users to explore time series datasets, apply transformations, and evaluate forecasting model robustness across diverse scenarios. We show how our framework can generate meaningful OOD time series that improve model robustness, and we validate EXPRTS effectiveness and usability through three use-cases and a user study.", "authors": ["Håkon Hanisch Kjærnli", "Lluis Mas-Ribas", "Hans Jakob Håland", "Vegard Sjåvik", "Aida Ashrafi", "Helge Langseth", "Odd Erik Gundersen"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-03-06", "url": "https://arxiv.org/abs/2403.03508", "pdf_url": "https://arxiv.org/pdf/2403.03508v3", "arxiv_id": "2403.03508", "doi": null, "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.0}
{"id": "4bf779b0c582bc27cc351db1034b08dbfa6bfdea90fbf881953d106cafdf3385", "sources": ["arxiv", "semantic_scholar"], "title": "Towards Foundation Time Series Model: To Synthesize Or Not To Synthesize?", "abstract": "The industry is rich in cases when we are required to make forecasting for large amounts of time series at once. However, we might be in a situation where we can not afford to train a separate model for each of them. Such issue in time series modeling remains without due attention. The remedy for this setting is the establishment of a foundation model. Such a model is expected to work in zero-shot and few-shot regimes. However, what should we take as a training dataset for such kind of model? Witnessing the benefits from the enrichment of NLP datasets with artificially-generated data, we might want to adopt their experience for time series. In contrast to natural language, the process of generation of synthetic time series data is even more favorable because it provides full control of series patterns, time horizons, and number of samples. In this work, we consider the essential question if it is advantageous to train a foundation model on synthetic data or it is better to utilize only a limited number of real-life examples. Our experiments are conducted only for regular time series and speak in favor of leveraging solely the real time series. Moreover, the choice of the proper source dataset strongly influences the performance during inference. When provided access even to a limited quantity of short time series data, employing it within a supervised framework yields more favorable results than training on a larger volume of synthetic data. The code for our experiments is publicly available on Github \\url{https://github.com/sb-ai-lab/synthesize_or_not}.", "authors": ["Kseniia Kuvshinova", "Olga Tsymboi", "Alina Kostromina", "Dmitry Simakov", "Elizaveta Kovtun"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-03-04", "url": "https://arxiv.org/abs/2403.02534", "pdf_url": "https://arxiv.org/pdf/2403.02534v1", "arxiv_id": "2403.02534", "doi": "10.48550/arXiv.2403.02534", "citation_count": 5, "influential_citation_count": 1, "has_code": true, "code_url": "https://github.com/sb-ai-lab/synthesize_or_not}", "venue": "arXiv.org", "quality_score": 0.1945}
{"id": "27023931b8355d581ce204c92407291ae9f0c5a79ee0a0edb3a1a225c516f85f", "sources": ["arxiv", "semantic_scholar"], "title": "TimeXer: Empowering Transformers for Time Series Forecasting with Exogenous Variables", "abstract": "Deep models have demonstrated remarkable performance in time series forecasting. However, due to the partially-observed nature of real-world applications, solely focusing on the target of interest, so-called endogenous variables, is usually insufficient to guarantee accurate forecasting. Notably, a system is often recorded into multiple variables, where the exogenous variables can provide valuable external information for endogenous variables. Thus, unlike well-established multivariate or univariate forecasting paradigms that either treat all the variables equally or ignore exogenous information, this paper focuses on a more practical setting: time series forecasting with exogenous variables. We propose a novel approach, TimeXer, to ingest external information to enhance the forecasting of endogenous variables. With deftly designed embedding layers, TimeXer empowers the canonical Transformer with the ability to reconcile endogenous and exogenous information, where patch-wise self-attention and variate-wise cross-attention are used simultaneously. Moreover, global endogenous tokens are learned to effectively bridge the causal information underlying exogenous series into endogenous temporal patches. Experimentally, TimeXer achieves consistent state-of-the-art performance on twelve real-world forecasting benchmarks and exhibits notable generality and scalability. Code is available at this repository: https://github.com/thuml/TimeXer.", "authors": ["Yuxuan Wang", "Haixu Wu", "Jiaxiang Dong", "Guo Qin", "Haoran Zhang", "Yong Liu", "Yunzhong Qiu", "Jianmin Wang", "Mingsheng Long"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-02-29", "url": "https://arxiv.org/abs/2402.19072", "pdf_url": "https://arxiv.org/pdf/2402.19072v4", "arxiv_id": "2402.19072", "doi": "10.48550/arXiv.2402.19072", "citation_count": 446, "influential_citation_count": 42, "has_code": true, "code_url": "https://github.com/thuml/TimeXer", "venue": "Neural Information Processing Systems", "quality_score": 0.8167}
{"id": "15460c2838d16ca96dbf6c2bd5c40337fcdd69190ec948b9bf76ee802df75d4c", "sources": ["arxiv", "semantic_scholar"], "title": "Generative Pretrained Hierarchical Transformer for Time Series Forecasting", "abstract": "Recent efforts have been dedicated to enhancing time series forecasting accuracy by introducing advanced network architectures and self-supervised pretraining strategies. Nevertheless, existing approaches still exhibit two critical drawbacks. Firstly, these methods often rely on a single dataset for training, limiting the model's generalizability due to the restricted scale of the training data. Secondly, the one-step generation schema is widely followed, which necessitates a customized forecasting head and overlooks the temporal dependencies in the output series, and also leads to increased training costs under different horizon length settings. To address these issues, we propose a novel generative pretrained hierarchical transformer architecture for forecasting, named \\textbf{GPHT}. There are two aspects of key designs in GPHT. On the one hand, we advocate for constructing a mixed dataset under the channel-independent assumption for pretraining our model, comprising various datasets from diverse data scenarios. This approach significantly expands the scale of training data, allowing our model to uncover commonalities in time series data and facilitating improved transfer to specific datasets. On the other hand, GPHT employs an auto-regressive forecasting approach, effectively modeling temporal dependencies in the output series. Importantly, no customized forecasting head is required, enabling \\textit{a single model to forecast at arbitrary horizon settings.} We conduct sufficient experiments on eight datasets with mainstream self-supervised pretraining models and supervised models. The results demonstrated that GPHT surpasses the baseline models across various fine-tuning and zero/few-shot learning settings in the traditional long-term forecasting task. We make our codes publicly available\\footnote{https://github.com/icantnamemyself/GPHT}.", "authors": ["Zhiding Liu", "Jiqian Yang", "Mingyue Cheng", "Yucong Luo", "Zhi Li"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-02-26", "url": "https://arxiv.org/abs/2402.16516", "pdf_url": "https://arxiv.org/pdf/2402.16516v2", "arxiv_id": "2402.16516", "doi": "10.1145/3637528.3671855", "citation_count": 44, "influential_citation_count": 3, "has_code": true, "code_url": "https://github.com/icantnamemyself/GPHT}", "venue": "Knowledge Discovery and Data Mining", "quality_score": 0.4133}
{"id": "0fc4dce11589999fc8ddff727be586887e89d1cb3bbc102f127aacadb950c074", "sources": ["arxiv", "semantic_scholar"], "title": "Time Series Forecasting with LLMs: Understanding and Enhancing Model Capabilities", "abstract": "Large language models (LLMs) have been applied in many fields and have developed rapidly in recent years. As a classic machine learning task, time series forecasting has recently been boosted by LLMs. Recent works treat large language models as \\emph{zero-shot} time series reasoners without further fine-tuning, which achieves remarkable performance. However, there are some unexplored research problems when applying LLMs for time series forecasting under the zero-shot setting. For instance, the LLMs' preferences for the input time series are less understood. In this paper, by comparing LLMs with traditional time series forecasting models, we observe many interesting properties of LLMs in the context of time series forecasting. First, our study shows that LLMs perform well in predicting time series with clear patterns and trends, but face challenges with datasets lacking periodicity. This observation can be explained by the ability of LLMs to recognize the underlying period within datasets, which is supported by our experiments. In addition, the input strategy is investigated, and it is found that incorporating external knowledge and adopting natural language paraphrases substantially improve the predictive performance of LLMs for time series. Overall, our study contributes insight into LLMs' advantages and limitations in time series forecasting under different conditions.", "authors": ["Hua Tang", "Chong Zhang", "Mingyu Jin", "Qinkai Yu", "Zhenting Wang", "Xiaobo Jin", "Yongfeng Zhang", "Mengnan Du"], "categories": ["cs.CL"], "fields_of_study": ["Computer Science"], "published_date": "2024-02-16", "url": "https://arxiv.org/abs/2402.10835", "pdf_url": "https://arxiv.org/pdf/2402.10835v5", "arxiv_id": "2402.10835", "doi": "10.1145/3715073.3715083", "citation_count": 88, "influential_citation_count": 3, "has_code": false, "code_url": null, "venue": "SIGKDD Explorations", "quality_score": 0.4873}
{"id": "aa6c56fe181e93d3cb30b2590a9bee6b6c259a0888a5e69e0b979eeafae35de6", "sources": ["arxiv", "semantic_scholar"], "title": "SAMformer: Unlocking the Potential of Transformers in Time Series Forecasting with Sharpness-Aware Minimization and Channel-Wise Attention", "abstract": "Transformer-based architectures achieved breakthrough performance in natural language processing and computer vision, yet they remain inferior to simpler linear baselines in multivariate long-term forecasting. To better understand this phenomenon, we start by studying a toy linear forecasting problem for which we show that transformers are incapable of converging to their true solution despite their high expressive power. We further identify the attention of transformers as being responsible for this low generalization capacity. Building upon this insight, we propose a shallow lightweight transformer model that successfully escapes bad local minima when optimized with sharpness-aware optimization. We empirically demonstrate that this result extends to all commonly used real-world multivariate time series datasets. In particular, SAMformer surpasses current state-of-the-art methods and is on par with the biggest foundation model MOIRAI while having significantly fewer parameters. The code is available at https://github.com/romilbert/samformer.", "authors": ["Romain Ilbert", "Ambroise Odonnat", "Vasilii Feofanov", "Aladin Virmaux", "Giuseppe Paolo", "Themis Palpanas", "Ievgen Redko"], "categories": ["cs.LG", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2024-02-15", "url": "https://arxiv.org/abs/2402.10198", "pdf_url": "https://arxiv.org/pdf/2402.10198v3", "arxiv_id": "2402.10198", "doi": null, "citation_count": 66, "influential_citation_count": 10, "has_code": true, "code_url": "https://github.com/romilbert/samformer", "venue": "International Conference on Machine Learning", "quality_score": 0.5207}
{"id": "8b11da3748880a213a94a90b734327b68c4b0c38b51ee1e90075f841ae25c52b", "sources": ["arxiv", "semantic_scholar"], "title": "Only the Curve Shape Matters: Training Foundation Models for Zero-Shot Multivariate Time Series Forecasting through Next Curve Shape Prediction", "abstract": "We present General Time Transformer (GTT), an encoder-only style foundation model for zero-shot multivariate time series forecasting. GTT is pretrained on a large dataset of 200M high-quality time series samples spanning diverse domains. In our proposed framework, the task of multivariate time series forecasting is formulated as a channel-wise next curve shape prediction problem, where each time series sample is represented as a sequence of non-overlapping curve shapes with a unified numerical magnitude. GTT is trained to predict the next curve shape based on a window of past curve shapes in a channel-wise manner. Experimental results demonstrate that GTT exhibits superior zero-shot multivariate forecasting capabilities on unseen time series datasets, even surpassing state-of-the-art supervised baselines. Additionally, we investigate the impact of varying GTT model parameters and training dataset scales, observing that the scaling law also holds in the context of zero-shot multivariate time series forecasting.", "authors": ["Cheng Feng", "Long Huang", "Denis Krompass"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-02-12", "url": "https://arxiv.org/abs/2402.07570", "pdf_url": "https://arxiv.org/pdf/2402.07570v2", "arxiv_id": "2402.07570", "doi": "10.48550/arXiv.2402.07570", "citation_count": 11, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.2698}
{"id": "1982ae6d69a3a835ed8fea68e6c5a4c5e5fe01c953002327cd358e0e9bffe2d9", "sources": ["arxiv", "semantic_scholar"], "title": "Sparse-VQ Transformer: An FFN-Free Framework with Vector Quantization for Enhanced Time Series Forecasting", "abstract": "Time series analysis is vital for numerous applications, and transformers have become increasingly prominent in this domain. Leading methods customize the transformer architecture from NLP and CV, utilizing a patching technique to convert continuous signals into segments. Yet, time series data are uniquely challenging due to significant distribution shifts and intrinsic noise levels. To address these two challenges,we introduce the Sparse Vector Quantized FFN-Free Transformer (Sparse-VQ). Our methodology capitalizes on a sparse vector quantization technique coupled with Reverse Instance Normalization (RevIN) to reduce noise impact and capture sufficient statistics for forecasting, serving as an alternative to the Feed-Forward layer (FFN) in the transformer architecture. Our FFN-free approach trims the parameter count, enhancing computational efficiency and reducing overfitting. Through evaluations across ten benchmark datasets, including the newly introduced CAISO dataset, Sparse-VQ surpasses leading models with a 7.84% and 4.17% decrease in MAE for univariate and multivariate time series forecasting, respectively. Moreover, it can be seamlessly integrated with existing transformer-based models to elevate their performance.", "authors": ["Yanjun Zhao", "Tian Zhou", "Chao Chen", "Liang Sun", "Yi Qian", "Rong Jin"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-02-08", "url": "https://arxiv.org/abs/2402.05830", "pdf_url": "https://arxiv.org/pdf/2402.05830v1", "arxiv_id": "2402.05830", "doi": "10.1109/ICASSP49660.2025.10889546", "citation_count": 8, "influential_citation_count": 2, "has_code": false, "code_url": null, "venue": "IEEE International Conference on Acoustics, Speech, and Signal Processing", "quality_score": 0.2386}
{"id": "d2593a8e62d9c790583647c5efa053cba7c54e42b0aa19a00776498a347c1b1b", "sources": ["arxiv", "semantic_scholar"], "title": "Attention as Robust Representation for Time Series Forecasting", "abstract": "Time series forecasting is essential for many practical applications, with the adoption of transformer-based models on the rise due to their impressive performance in NLP and CV. Transformers' key feature, the attention mechanism, dynamically fusing embeddings to enhance data representation, often relegating attention weights to a byproduct role. Yet, time series data, characterized by noise and non-stationarity, poses significant forecasting challenges. Our approach elevates attention weights as the primary representation for time series, capitalizing on the temporal relationships among data points to improve forecasting accuracy. Our study shows that an attention map, structured using global landmarks and local windows, acts as a robust kernel representation for data points, withstanding noise and shifts in distribution. Our method outperforms state-of-the-art models, reducing mean squared error (MSE) in multivariate time series forecasting by a notable 3.6% without altering the core neural network architecture. It serves as a versatile component that can readily replace recent patching based embedding schemes in transformer-based models, boosting their performance.", "authors": ["PeiSong Niu", "Tian Zhou", "Xue Wang", "Liang Sun", "Rong Jin"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-02-08", "url": "https://arxiv.org/abs/2402.05370", "pdf_url": "https://arxiv.org/pdf/2402.05370v1", "arxiv_id": "2402.05370", "doi": "10.48550/arXiv.2402.05370", "citation_count": 15, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.301}
{"id": "ede9b7357939b16955022db56fac6c2061684a593f3e59a8d54e60e3726bc045", "sources": ["arxiv", "semantic_scholar"], "title": "MOMENT: A Family of Open Time-series Foundation Models", "abstract": "We introduce MOMENT, a family of open-source foundation models for general-purpose time series analysis. Pre-training large models on time series data is challenging due to (1) the absence of a large and cohesive public time series repository, and (2) diverse time series characteristics which make multi-dataset training onerous. Additionally, (3) experimental benchmarks to evaluate these models, especially in scenarios with limited resources, time, and supervision, are still in their nascent stages. To address these challenges, we compile a large and diverse collection of public time series, called the Time series Pile, and systematically tackle time series-specific challenges to unlock large-scale multi-dataset pre-training. Finally, we build on recent work to design a benchmark to evaluate time series foundation models on diverse tasks and datasets in limited supervision settings. Experiments on this benchmark demonstrate the effectiveness of our pre-trained models with minimal data and task-specific fine-tuning. Finally, we present several interesting empirical observations about large pre-trained time series models. Pre-trained models (AutonLab/MOMENT-1-large) and Time Series Pile (AutonLab/Timeseries-PILE) are available on Huggingface.", "authors": ["Mononito Goswami", "Konrad Szafer", "Arjun Choudhry", "Yifu Cai", "Shuo Li", "Artur Dubrawski"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-02-06", "url": "https://arxiv.org/abs/2402.03885", "pdf_url": "https://arxiv.org/pdf/2402.03885v3", "arxiv_id": "2402.03885", "doi": "10.48550/arXiv.2402.03885", "citation_count": 503, "influential_citation_count": 96, "has_code": true, "code_url": null, "venue": "International Conference on Machine Learning", "quality_score": 0.9934}
{"id": "93d9e4eb970b9aab0ddb42495e92b2f2cf52ab1566940f5d95182c9ef9babc6c", "sources": ["arxiv", "semantic_scholar"], "title": "A Survey on Transformer Compression", "abstract": "Transformer plays a vital role in the realms of natural language processing (NLP) and computer vision (CV), specially for constructing large language models (LLM) and large vision models (LVM). Model compression methods reduce the memory and computational cost of Transformer, which is a necessary step to implement large language/vision models on practical devices. Given the unique architecture of Transformer, featuring alternative attention and feedforward neural network (FFN) modules, specific compression techniques are usually required. The efficiency of these compression methods is also paramount, as retraining large models on the entire training dataset is usually impractical. This survey provides a comprehensive review of recent compression methods, with a specific focus on their application to Transformer-based models. The compression methods are primarily categorized into pruning, quantization, knowledge distillation, and efficient architecture design (Mamba, RetNet, RWKV, etc.). In each category, we discuss compression methods for both language and vision tasks, highlighting common underlying principles. Finally, we delve into the relation between various compression methods, and discuss further directions in this domain.", "authors": ["Yehui Tang", "Yunhe Wang", "Jianyuan Guo", "Zhijun Tu", "Kai Han", "Hailin Hu", "Dacheng Tao"], "categories": ["cs.LG", "cs.CL", "cs.CV"], "fields_of_study": ["Computer Science"], "published_date": "2024-02-05", "url": "https://arxiv.org/abs/2402.05964", "pdf_url": "https://arxiv.org/pdf/2402.05964v2", "arxiv_id": "2402.05964", "doi": "10.48550/arXiv.2402.05964", "citation_count": 77, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.473}
{"id": "9e290e885f5534c137bb8af12519294b2e97480206c8b06b3d8fa4509edcfaaa", "sources": ["arxiv", "semantic_scholar"], "title": "Unified Training of Universal Time Series Forecasting Transformers", "abstract": "Deep learning for time series forecasting has traditionally operated within a one-model-per-dataset framework, limiting its potential to leverage the game-changing impact of large pre-trained models. The concept of universal forecasting, emerging from pre-training on a vast collection of time series datasets, envisions a single Large Time Series Model capable of addressing diverse downstream forecasting tasks. However, constructing such a model poses unique challenges specific to time series data: i) cross-frequency learning, ii) accommodating an arbitrary number of variates for multivariate time series, and iii) addressing the varying distributional properties inherent in large-scale data. To address these challenges, we present novel enhancements to the conventional time series Transformer architecture, resulting in our proposed Masked Encoder-based Universal Time Series Forecasting Transformer (Moirai). Trained on our newly introduced Large-scale Open Time Series Archive (LOTSA) featuring over 27B observations across nine domains, Moirai achieves competitive or superior performance as a zero-shot forecaster when compared to full-shot models. Code, data, and model weights can be found at https://github.com/SalesforceAIResearch/uni2ts.", "authors": ["Gerald Woo", "Chenghao Liu", "Akshat Kumar", "Caiming Xiong", "Silvio Savarese", "Doyen Sahoo"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-02-04", "url": "https://arxiv.org/abs/2402.02592", "pdf_url": "https://arxiv.org/pdf/2402.02592v2", "arxiv_id": "2402.02592", "doi": "10.48550/arXiv.2402.02592", "citation_count": 606, "influential_citation_count": 94, "has_code": true, "code_url": "https://github.com/SalesforceAIResearch/uni2ts", "venue": "International Conference on Machine Learning", "quality_score": 0.9889}
{"id": "165bed7e5bec1caf878b24375a6443ebe6cf3ce5aa609b11dc6f3af158b42f9a", "sources": ["arxiv", "semantic_scholar"], "title": "Pathformer: Multi-scale Transformers with Adaptive Pathways for Time Series Forecasting", "abstract": "Transformers for time series forecasting mainly model time series from limited or fixed scales, making it challenging to capture different characteristics spanning various scales. We propose Pathformer, a multi-scale Transformer with adaptive pathways. It integrates both temporal resolution and temporal distance for multi-scale modeling. Multi-scale division divides the time series into different temporal resolutions using patches of various sizes. Based on the division of each scale, dual attention is performed over these patches to capture global correlations and local details as temporal dependencies. We further enrich the multi-scale Transformer with adaptive pathways, which adaptively adjust the multi-scale modeling process based on the varying temporal dynamics of the input, improving the accuracy and generalization of Pathformer. Extensive experiments on eleven real-world datasets demonstrate that Pathformer not only achieves state-of-the-art performance by surpassing all current models but also exhibits stronger generalization abilities under various transfer scenarios. The code is made available at https://github.com/decisionintelligence/pathformer.", "authors": ["Peng Chen", "Yingying Zhang", "Yunyao Cheng", "Yang Shu", "Yihang Wang", "Qingsong Wen", "Bin Yang", "Chenjuan Guo"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-02-04", "url": "https://arxiv.org/abs/2402.05956", "pdf_url": "https://arxiv.org/pdf/2402.05956v5", "arxiv_id": "2402.05956", "doi": "10.48550/arXiv.2402.05956", "citation_count": 207, "influential_citation_count": 15, "has_code": true, "code_url": "https://github.com/decisionintelligence/pathformer", "venue": "International Conference on Learning Representations", "quality_score": 0.6021}
{"id": "67713c42e015e9325af2a2767437ef221b8f5798bf965651868799a4c06d185b", "sources": ["arxiv", "semantic_scholar"], "title": "AutoTimes: Autoregressive Time Series Forecasters via Large Language Models", "abstract": "Foundation models of time series have not been fully developed due to the limited availability of time series corpora and the underexploration of scalable pre-training. Based on the similar sequential formulation of time series and natural language, increasing research demonstrates the feasibility of leveraging large language models (LLM) for time series. Nevertheless, the inherent autoregressive property and decoder-only architecture of LLMs have not been fully considered, resulting in insufficient utilization of LLM abilities. To fully revitalize the general-purpose token transition and multi-step generation capability of large language models, we propose AutoTimes to repurpose LLMs as autoregressive time series forecasters, which projects time series into the embedding space of language tokens and autoregressively generates future predictions with arbitrary lengths. Compatible with any decoder-only LLMs, the consequent forecaster exhibits the flexibility of the lookback length and scalability with larger LLMs. Further, we formulate time series as prompts, extending the context for prediction beyond the lookback window, termed in-context forecasting. By introducing LLM-embedded textual timestamps, AutoTimes can utilize chronological information to align multivariate time series. Empirically, AutoTimes achieves state-of-the-art with 0.1% trainable parameters and over $5\\times$ training/inference speedup compared to advanced LLM-based forecasters. Code is available at this repository: https://github.com/thuml/AutoTimes.", "authors": ["Yong Liu", "Guo Qin", "Xiangdong Huang", "Jianmin Wang", "Mingsheng Long"], "categories": ["cs.LG", "cs.CL"], "fields_of_study": ["Computer Science"], "published_date": "2024-02-04", "url": "https://arxiv.org/abs/2402.02370", "pdf_url": "https://arxiv.org/pdf/2402.02370v4", "arxiv_id": "2402.02370", "doi": "10.48550/arXiv.2402.02370", "citation_count": 102, "influential_citation_count": 9, "has_code": true, "code_url": "https://github.com/thuml/AutoTimes", "venue": "Neural Information Processing Systems", "quality_score": 0.5032}
{"id": "8ebe9cd4a9b65d72cc89bd502ce7857c190617b309eb352cf623c1711043b6a9", "sources": ["arxiv", "semantic_scholar"], "title": "Large Language Models for Time Series: A Survey", "abstract": "Large Language Models (LLMs) have seen significant use in domains such as natural language processing and computer vision. Going beyond text, image and graphics, LLMs present a significant potential for analysis of time series data, benefiting domains such as climate, IoT, healthcare, traffic, audio and finance. This survey paper provides an in-depth exploration and a detailed taxonomy of the various methodologies employed to harness the power of LLMs for time series analysis. We address the inherent challenge of bridging the gap between LLMs' original text data training and the numerical nature of time series data, and explore strategies for transferring and distilling knowledge from LLMs to numerical time series analysis. We detail various methodologies, including (1) direct prompting of LLMs, (2) time series quantization, (3) aligning techniques, (4) utilization of the vision modality as a bridging mechanism, and (5) the combination of LLMs with tools. Additionally, this survey offers a comprehensive overview of the existing multimodal time series and text datasets and delves into the challenges and future opportunities of this emerging field. We maintain an up-to-date Github repository which includes all the papers and datasets discussed in the survey.", "authors": ["Xiyuan Zhang", "Ranak Roy Chowdhury", "Rajesh K. Gupta", "Jingbo Shang"], "categories": ["cs.LG", "cs.AI", "cs.CL"], "fields_of_study": ["Computer Science"], "published_date": "2024-02-02", "url": "https://arxiv.org/abs/2402.01801", "pdf_url": "https://arxiv.org/pdf/2402.01801v3", "arxiv_id": "2402.01801", "doi": "10.48550/arXiv.2402.01801", "citation_count": 166, "influential_citation_count": 4, "has_code": true, "code_url": "https://github.com/xiyuanzh/awesome-llm-time-series", "venue": "International Joint Conference on Artificial Intelligence", "quality_score": 0.5557}
{"id": "40adeef83fda47d6379c595dacdb93fae047d25478c9cbb6b42234aadbdcda8e", "sources": ["arxiv", "semantic_scholar"], "title": "IN-Flow: Instance Normalization Flow for Non-stationary Time Series Forecasting", "abstract": "Due to the non-stationarity of time series, the distribution shift problem largely hinders the performance of time series forecasting. Existing solutions either rely on using certain statistics to specify the shift, or developing specific mechanisms for certain network architectures. However, the former would fail for the unknown shift beyond simple statistics, while the latter has limited compatibility on different forecasting models. To overcome these problems, we first propose a decoupled formulation for time series forecasting, with no reliance on fixed statistics and no restriction on forecasting architectures. This formulation regards the removing-shift procedure as a special transformation between a raw distribution and a desired target distribution and separates it from the forecasting. Such a formulation is further formalized into a bi-level optimization problem, to enable the joint learning of the transformation (outer loop) and forecasting (inner loop). Moreover, the special requirements of expressiveness and bi-direction for the transformation motivate us to propose instance normalization flow (IN-Flow), a novel invertible network for time series transformation. Different from the classic \"normalizing flow\" models, IN-Flow does not aim for normalizing input to the prior distribution (e.g., Gaussian distribution) for generation, but creatively transforms time series distribution by stacking normalization layers and flow-based invertible networks, which is thus named \"normalization\" flow. Finally, we have conducted extensive experiments on both synthetic data and real-world data, which demonstrate the superiority of our method.", "authors": ["Wei Fan", "Shun Zheng", "Pengyang Wang", "Rui Xie", "Kun Yi", "Qi Zhang", "Jiang Bian", "Yanjie Fu"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-01-30", "url": "https://arxiv.org/abs/2401.16777", "pdf_url": "https://arxiv.org/pdf/2401.16777v2", "arxiv_id": "2401.16777", "doi": "10.1145/3690624.3709260", "citation_count": 12, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "Knowledge Discovery and Data Mining", "quality_score": 0.2785}
{"id": "a40a4bceeabee93d9305703e46832dc07695f3bf6649c77ced71de98ad0d3755", "sources": ["arxiv", "semantic_scholar"], "title": "A Survey of Deep Learning and Foundation Models for Time Series Forecasting", "abstract": "Deep Learning has been successfully applied to many application domains, yet its advantages have been slow to emerge for time series forecasting. For example, in the well-known Makridakis (M) Competitions, hybrids of traditional statistical or machine learning techniques have only recently become the top performers. With the recent architectural advances in deep learning being applied to time series forecasting (e.g., encoder-decoders with attention, transformers, and graph neural networks), deep learning has begun to show significant advantages. Still, in the area of pandemic prediction, there remain challenges for deep learning models: the time series is not long enough for effective training, unawareness of accumulated scientific knowledge, and interpretability of the model. To this end, the development of foundation models (large deep learning models with extensive pre-training) allows models to understand patterns and acquire knowledge that can be applied to new related problems before extensive training data becomes available. Furthermore, there is a vast amount of knowledge available that deep learning models can tap into, including Knowledge Graphs and Large Language Models fine-tuned with scientific domain knowledge. There is ongoing research examining how to utilize or inject such knowledge into deep learning models. In this survey, several state-of-the-art modeling techniques are reviewed, and suggestions for further work are provided.", "authors": ["John A. Miller", "Mohammed Aldosari", "Farah Saeed", "Nasid Habib Barna", "Subas Rana", "I. Budak Arpinar", "Ninghao Liu"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-01-25", "url": "https://arxiv.org/abs/2401.13912", "pdf_url": "https://arxiv.org/pdf/2401.13912v1", "arxiv_id": "2401.13912", "doi": "10.48550/arXiv.2401.13912", "citation_count": 73, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.4673}
{"id": "2ce359f4b1208fc1537c82b5779e9b69613d1b5deb5d2cb8c8e7242433b6c7f0", "sources": ["arxiv", "semantic_scholar"], "title": "Transformers with Attentive Federated Aggregation for Time Series Stock Forecasting", "abstract": "Recent innovations in transformers have shown their superior performance in natural language processing (NLP) and computer vision (CV). The ability to capture long-range dependencies and interactions in sequential data has also triggered a great interest in time series modeling, leading to the widespread use of transformers in many time series applications. However, being the most common and crucial application, the adaptation of transformers to time series forecasting has remained limited, with both promising and inconsistent results. In contrast to the challenges in NLP and CV, time series problems not only add the complexity of order or temporal dependence among input sequences but also consider trend, level, and seasonality information that much of this data is valuable for decision making. The conventional training scheme has shown deficiencies regarding model overfitting, data scarcity, and privacy issues when working with transformers for a forecasting task. In this work, we propose attentive federated transformers for time series stock forecasting with better performance while preserving the privacy of participating enterprises. Empirical results on various stock data from the Yahoo! Finance website indicate the superiority of our proposed scheme in dealing with the above challenges and data heterogeneity in federated learning.", "authors": ["Chu Myaet Thwal", "Ye Lin Tun", "Kitae Kim", "Seong-Bae Park", "Choong Seon Hong"], "categories": ["q-fin.ST", "cs.AI", "cs.CE", "cs.DC", "cs.LG"], "fields_of_study": ["Computer Science", "Economics"], "published_date": "2024-01-22", "url": "https://arxiv.org/abs/2402.06638", "pdf_url": "https://arxiv.org/pdf/2402.06638v1", "arxiv_id": "2402.06638", "doi": "10.1109/ICOIN56518.2023.10048928", "citation_count": 8, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "International Conference on Information Networking", "quality_score": 0.2386}
{"id": "b03f7a363827c2fd6e4ab323afb1cfe17b3b3f9f14333f7819e47b4fd05daaf5", "sources": ["arxiv", "semantic_scholar"], "title": "MSHyper: Multi-Scale Hypergraph Transformer for Long-Range Time Series Forecasting", "abstract": "Demystifying interactions between temporal patterns of different scales is fundamental to precise long-range time series forecasting. However, previous works lack the ability to model high-order interactions. To promote more comprehensive pattern interaction modeling for long-range time series forecasting, we propose a Multi-Scale Hypergraph Transformer (MSHyper) framework. Specifically, a multi-scale hypergraph is introduced to provide foundations for modeling high-order pattern interactions. Then by treating hyperedges as nodes, we also build a hyperedge graph to enhance hypergraph modeling. In addition, a tri-stage message passing mechanism is introduced to aggregate pattern information and learn the interaction strength between temporal patterns of different scales. Extensive experiments on five real-world datasets demonstrate that MSHyper achieves state-of-the-art (SOTA) performance across various settings.", "authors": ["Zongjiang Shang", "Ling Chen", "Binqing Wu", "Dongliang Cui"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-01-17", "url": "https://arxiv.org/abs/2401.09261", "pdf_url": "https://arxiv.org/pdf/2401.09261v2", "arxiv_id": "2401.09261", "doi": "10.48550/arXiv.2401.09261", "citation_count": 15, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.301}
{"id": "f4cb4ec59971001cd876c9ee310b2a4754a8df788182f6a1ebbf2d7a65c22757", "sources": ["arxiv", "semantic_scholar"], "title": "Domain Adaptation for Time series Transformers using One-step fine-tuning", "abstract": "The recent breakthrough of Transformers in deep learning has drawn significant attention of the time series community due to their ability to capture long-range dependencies. However, like other deep learning models, Transformers face limitations in time series prediction, including insufficient temporal understanding, generalization challenges, and data shift issues for the domains with limited data. Additionally, addressing the issue of catastrophic forgetting, where models forget previously learned information when exposed to new data, is another critical aspect that requires attention in enhancing the robustness of Transformers for time series tasks. To address these limitations, in this paper, we pre-train the time series Transformer model on a source domain with sufficient data and fine-tune it on the target domain with limited data. We introduce the \\emph{One-step fine-tuning} approach, adding some percentage of source domain data to the target domains, providing the model with diverse time series instances. We then fine-tune the pre-trained model using a gradual unfreezing technique. This helps enhance the model's performance in time series prediction for domains with limited data. Extensive experimental results on two real-world datasets show that our approach improves over the state-of-the-art baselines by 4.35% and 11.54% for indoor temperature and wind power prediction, respectively.", "authors": ["Subina Khanal", "Seshu Tirupathi", "Giulio Zizzo", "Ambrish Rawat", "Torben Bach Pedersen"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-01-12", "url": "https://arxiv.org/abs/2401.06524", "pdf_url": "https://arxiv.org/pdf/2401.06524v1", "arxiv_id": "2401.06524", "doi": "10.48550/arXiv.2401.06524", "citation_count": 9, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.25}
{"id": "f231d0e305ec74c32bfdaf4128941f62568d19f145e898e6d84ee047d2cfae2f", "sources": ["arxiv", "semantic_scholar"], "title": "HiMTM: Hierarchical Multi-Scale Masked Time Series Modeling with Self-Distillation for Long-Term Forecasting", "abstract": "Time series forecasting is a critical and challenging task in practical application. Recent advancements in pre-trained foundation models for time series forecasting have gained significant interest. However, current methods often overlook the multi-scale nature of time series, which is essential for accurate forecasting. To address this, we propose HiMTM, a hierarchical multi-scale masked time series modeling with self-distillation for long-term forecasting. HiMTM integrates four key components: (1) hierarchical multi-scale transformer (HMT) to capture temporal information at different scales; (2) decoupled encoder-decoder (DED) that directs the encoder towards feature extraction while the decoder focuses on pretext tasks; (3) hierarchical self-distillation (HSD) for multi-stage feature-level supervision signals during pre-training; and (4) cross-scale attention fine-tuning (CSA-FT) to capture dependencies between different scales for downstream tasks. These components collectively enhance multi-scale feature extraction in masked time series modeling, improving forecasting accuracy. Extensive experiments on seven mainstream datasets show that HiMTM surpasses state-of-the-art self-supervised and end-to-end learning methods by a considerable margin of 3.16-68.54\\%. Additionally, HiMTM outperforms the latest robust self-supervised learning method, PatchTST, in cross-domain forecasting by a significant margin of 2.3\\%. The effectiveness of HiMTM is further demonstrated through its application in natural gas demand forecasting.", "authors": ["Shubao Zhao", "Ming Jin", "Zhaoxiang Hou", "Chengyi Yang", "Zengxiang Li", "Qingsong Wen", "Yi Wang"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2024-01-10", "url": "https://arxiv.org/abs/2401.05012", "pdf_url": "https://arxiv.org/pdf/2401.05012v2", "arxiv_id": "2401.05012", "doi": "10.1145/3627673.3679741", "citation_count": 13, "influential_citation_count": 2, "has_code": false, "code_url": null, "venue": "International Conference on Information and Knowledge Management", "quality_score": 0.2865}
{"id": "ba29a73e80741a0d9fcd58dfaf7c21603855b59852efa74aeeddd89ff2acec8f", "sources": ["arxiv", "semantic_scholar"], "title": "The Rise of Diffusion Models in Time-Series Forecasting", "abstract": "This survey delves into the application of diffusion models in time-series forecasting. Diffusion models are demonstrating state-of-the-art results in various fields of generative AI. The paper includes comprehensive background information on diffusion models, detailing their conditioning methods and reviewing their use in time-series forecasting. The analysis covers 11 specific time-series implementations, the intuition and theory behind them, the effectiveness on different datasets, and a comparison among each other. Key contributions of this work are the thorough exploration of diffusion models' applications in time-series forecasting and a chronologically ordered overview of these models. Additionally, the paper offers an insightful discussion on the current state-of-the-art in this domain and outlines potential future research directions. This serves as a valuable resource for researchers in AI and time-series analysis, offering a clear view of the latest advancements and future potential of diffusion models.", "authors": ["Caspar Meijer", "Lydia Y. Chen"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-01-05", "url": "https://arxiv.org/abs/2401.03006", "pdf_url": "https://arxiv.org/pdf/2401.03006v2", "arxiv_id": "2401.03006", "doi": "10.48550/arXiv.2401.03006", "citation_count": 32, "influential_citation_count": 3, "has_code": true, "code_url": "https://github.com/Capsar/The-Rise-of-Diffusion-Models-in-Time-Series-Forecasting", "venue": "arXiv.org", "quality_score": 0.3796}
{"id": "b8888f3130a55ba0b868a86c09930dcf184d4c79f3c422dfef890efc3aed947d", "sources": ["arxiv", "semantic_scholar"], "title": "Explainable Adaptive Tree-based Model Selection for Time Series Forecasting", "abstract": "Tree-based models have been successfully applied to a wide variety of tasks, including time series forecasting. They are increasingly in demand and widely accepted because of their comparatively high level of interpretability. However, many of them suffer from the overfitting problem, which limits their application in real-world decision-making. This problem becomes even more severe in online-forecasting settings where time series observations are incrementally acquired, and the distributions from which they are drawn may keep changing over time. In this context, we propose a novel method for the online selection of tree-based models using the TreeSHAP explainability method in the task of time series forecasting. We start with an arbitrary set of different tree-based models. Then, we outline a performance-based ranking with a coherent design to make TreeSHAP able to specialize the tree-based forecasters across different regions in the input time series. In this framework, adequate model selection is performed online, adaptively following drift detection in the time series. In addition, explainability is supported on three levels, namely online input importance, model selection, and model output explanation. An extensive empirical study on various real-world datasets demonstrates that our method achieves excellent or on-par results in comparison to the state-of-the-art approaches as well as several baselines.", "authors": ["Matthias Jakobs", "Amal Saadallah"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2024-01-02", "url": "https://arxiv.org/abs/2401.01124", "pdf_url": "https://arxiv.org/pdf/2401.01124v1", "arxiv_id": "2401.01124", "doi": "10.1109/ICDM58522.2023.00027", "citation_count": 9, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "Industrial Conference on Data Mining", "quality_score": 0.25}
{"id": "62d451bc54ea85ace1f60b790d9dccc93f669c28a50b2d8e6b8cffef6498eb3b", "sources": ["arxiv", "semantic_scholar"], "title": "auto-sktime: Automated Time Series Forecasting", "abstract": "In today's data-driven landscape, time series forecasting is pivotal in decision-making across various sectors. Yet, the proliferation of more diverse time series data, coupled with the expanding landscape of available forecasting methods, poses significant challenges for forecasters. To meet the growing demand for efficient forecasting, we introduce auto-sktime, a novel framework for automated time series forecasting. The proposed framework uses the power of automated machine learning (AutoML) techniques to automate the creation of the entire forecasting pipeline. The framework employs Bayesian optimization, to automatically construct pipelines from statistical, machine learning (ML) and deep neural network (DNN) models. Furthermore, we propose three essential improvements to adapt AutoML to time series data. First, pipeline templates to account for the different supported forecasting models. Second, a novel warm-starting technique to start the optimization from prior optimization runs. Third, we adapt multi-fidelity optimizations to make them applicable to a search space containing statistical, ML and DNN models. Experimental results on 64 diverse real-world time series datasets demonstrate the effectiveness and efficiency of the framework, outperforming traditional methods while requiring minimal human involvement.", "authors": ["Marc-André Zöller", "Marius Lindauer", "Marco F. Huber"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2023-12-13", "url": "https://arxiv.org/abs/2312.08528", "pdf_url": "https://arxiv.org/pdf/2312.08528v3", "arxiv_id": "2312.08528", "doi": "10.48550/arXiv.2312.08528", "citation_count": 2, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "Learning and Intelligent Optimization", "quality_score": 0.1193}
{"id": "610e9de6df4d4e850bff3a8a0339edd96820861057f6c2bc8a87cb8c9a762bc6", "sources": ["arxiv", "semantic_scholar"], "title": "Sparse Transformer with Local and Seasonal Adaptation for Multivariate Time Series Forecasting", "abstract": "Transformers have achieved remarkable performance in multivariate time series(MTS) forecasting due to their capability to capture long-term dependencies. However, the canonical attention mechanism has two key limitations: (1) its quadratic time complexity limits the sequence length, and (2) it generates future values from the entire historical sequence. To address this, we propose a Dozer Attention mechanism consisting of three sparse components: (1) Local, each query exclusively attends to keys within a localized window of neighboring time steps. (2) Stride, enables each query to attend to keys at predefined intervals. (3) Vary, allows queries to selectively attend to keys from a subset of the historical sequence. Notably, the size of this subset dynamically expands as forecasting horizons extend. Those three components are designed to capture essential attributes of MTS data, including locality, seasonality, and global temporal dependencies. Additionally, we present the Dozerformer Framework, incorporating the Dozer Attention mechanism for the MTS forecasting task. We evaluated the proposed Dozerformer framework with recent state-of-the-art methods on nine benchmark datasets and confirmed its superior performance. The experimental results indicate that excluding a subset of historical time steps from the time series forecasting process does not compromise accuracy while significantly improving efficiency. Code is available at https://github.com/GRYGY1215/Dozerformer.", "authors": ["Yifan Zhang", "Rui Wu", "Sergiu M. Dascalu", "Frederick C. Harris"], "categories": ["cs.LG", "cs.CL"], "fields_of_study": ["Computer Science", "Medicine"], "published_date": "2023-12-11", "url": "https://arxiv.org/abs/2312.06874", "pdf_url": "https://arxiv.org/pdf/2312.06874v2", "arxiv_id": "2312.06874", "doi": "10.1038/s41598-024-66886-1", "citation_count": 16, "influential_citation_count": 1, "has_code": true, "code_url": "https://github.com/GRYGY1215/Dozerformer", "venue": "Scientific Reports", "quality_score": 0.3076}
{"id": "376b926c83ffa939a428dae9529df39253dbb6d77e1b72821257d99944716b2a", "sources": ["arxiv", "semantic_scholar"], "title": "MultiResFormer: Transformer with Adaptive Multi-Resolution Modeling for General Time Series Forecasting", "abstract": "Transformer-based models have greatly pushed the boundaries of time series forecasting recently. Existing methods typically encode time series data into $\\textit{patches}$ using one or a fixed set of patch lengths. This, however, could result in a lack of ability to capture the variety of intricate temporal dependencies present in real-world multi-periodic time series. In this paper, we propose MultiResFormer, which dynamically models temporal variations by adaptively choosing optimal patch lengths. Concretely, at the beginning of each layer, time series data is encoded into several parallel branches, each using a detected periodicity, before going through the transformer encoder block. We conduct extensive evaluations on long- and short-term forecasting datasets comparing MultiResFormer with state-of-the-art baselines. MultiResFormer outperforms patch-based Transformer baselines on long-term forecasting tasks and also consistently outperforms CNN baselines by a large margin, while using much fewer parameters than these baselines.", "authors": ["Linfeng Du", "Ji Xin", "Alex Labach", "Saba Zuberi", "Maksims Volkovs", "Rahul G. Krishnan"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2023-11-30", "url": "https://arxiv.org/abs/2311.18780", "pdf_url": "https://arxiv.org/pdf/2311.18780v2", "arxiv_id": "2311.18780", "doi": "10.48550/arXiv.2311.18780", "citation_count": 3, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1505}
{"id": "e8f239181e3c0a30b6af03094a5a7b35ac65d8813daaa1683fb7de1a9d8b8576", "sources": ["arxiv", "semantic_scholar"], "title": "TimelyGPT: Extrapolatable Transformer Pre-training for Long-term Time-Series Forecasting in Healthcare", "abstract": "Large-scale pre-trained models (PTMs) such as BERT and GPT have recently achieved great success in Natural Language Processing and Computer Vision domains. However, the development of PTMs on healthcare time-series data is lagging behind.This underscores the limitations of the existing transformer-based architectures, particularly their scalability to handle large-scale time series and ability to capture long-term temporal dependencies. In this study, we present Timely Generative Pre-trained Transformer (TimelyGPT). TimelyGPT employs an extrapolatable position (xPos) embedding to encode trend and periodic patterns into time-series representations. It also integrates recurrent attention and temporal convolution modules to effectively capture global-local temporal dependencies. We evaluated TimelyGPT on two large-scale healthcare time series datasets corresponding to continuous biosignals and irregularly-sampled time series, respectively. Our experiments show that during pre-training, TimelyGPT excels in learning time-series representations from continuously monitored biosignals and irregularly-sampled time series data commonly observed in longitudinal electronic health records (EHRs). In forecasting continuous biosignals, TimelyGPT achieves accurate extrapolation up to 6,000 timesteps of body temperature during the sleep stage transition, given a short look-up window (i.e., prompt) containing only 2,000 timesteps. For irregularly-sampled time series, TimelyGPT with a proposed time-specific inference demonstrates high top recall scores in predicting future diagnoses using early diagnostic records, effectively handling irregular intervals between clinical records. Together, we envision TimelyGPT to be useful in a broad spectrum of health domains, including long-term patient health state forecasting and patient risk trajectory prediction.", "authors": ["Ziyang Song", "Qincheng Lu", "Hao Xu", "He Zhu", "David L. Buckeridge", "Yue Li"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science", "Medicine"], "published_date": "2023-11-29", "url": "https://arxiv.org/abs/2312.00817", "pdf_url": "https://arxiv.org/pdf/2312.00817v3", "arxiv_id": "2312.00817", "doi": "10.1007/s13755-025-00384-0", "citation_count": 12, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "Health Information Science and Systems", "quality_score": 0.2785}
{"id": "96b37b726e9e0199097d755f169f3a2803aba2a2b75a4624b1a8b42459ecf948", "sources": ["arxiv", "semantic_scholar"], "title": "A projected nonlinear state-space model for forecasting time series signals", "abstract": "Learning and forecasting stochastic time series is essential in various scientific fields. However, despite the proposals of nonlinear filters and deep-learning methods, it remains challenging to capture nonlinear dynamics from a few noisy samples and predict future trajectories with uncertainty estimates while maintaining computational efficiency. Here, we propose a fast algorithm to learn and forecast nonlinear dynamics from noisy time series data. A key feature of the proposed model is kernel functions applied to projected lines, enabling fast and efficient capture of nonlinearities in the latent dynamics. Through empirical case studies and benchmarking, the model demonstrates its effectiveness in learning and forecasting complex nonlinear dynamics, offering a valuable tool for researchers and practitioners in time series analysis.", "authors": ["Christian Donner", "Anuj Mishra", "Hideaki Shimazaki"], "categories": ["stat.ME", "stat.ML"], "fields_of_study": ["Mathematics"], "published_date": "2023-11-22", "url": "https://arxiv.org/abs/2311.13247", "pdf_url": "https://arxiv.org/pdf/2311.13247v2", "arxiv_id": "2311.13247", "doi": "10.1016/j.ijforecast.2025.01.002", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "International Journal of Forecasting", "quality_score": 0.0753}
{"id": "9f836b99b1e7c6b492909855acf77b96473e0cec4fe0b7217d98bf92befd1d9d", "sources": ["arxiv", "semantic_scholar"], "title": "Multi-resolution Time-Series Transformer for Long-term Forecasting", "abstract": "The performance of transformers for time-series forecasting has improved significantly. Recent architectures learn complex temporal patterns by segmenting a time-series into patches and using the patches as tokens. The patch size controls the ability of transformers to learn the temporal patterns at different frequencies: shorter patches are effective for learning localized, high-frequency patterns, whereas mining long-term seasonalities and trends requires longer patches. Inspired by this observation, we propose a novel framework, Multi-resolution Time-Series Transformer (MTST), which consists of a multi-branch architecture for simultaneous modeling of diverse temporal patterns at different resolutions. In contrast to many existing time-series transformers, we employ relative positional encoding, which is better suited for extracting periodic components at different scales. Extensive experiments on several real-world datasets demonstrate the effectiveness of MTST in comparison to state-of-the-art forecasting techniques.", "authors": ["Yitian Zhang", "Liheng Ma", "Soumyasundar Pal", "Yingxue Zhang", "Mark Coates"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2023-11-07", "url": "https://arxiv.org/abs/2311.04147", "pdf_url": "https://arxiv.org/pdf/2311.04147v2", "arxiv_id": "2311.04147", "doi": "10.48550/arXiv.2311.04147", "citation_count": 102, "influential_citation_count": 9, "has_code": false, "code_url": null, "venue": "International Conference on Artificial Intelligence and Statistics", "quality_score": 0.5032}
{"id": "58cc9768a694fb6718a1d8b830a3f7aa0d67a13985e06bfdee40d54e588aef92", "sources": ["arxiv", "semantic_scholar"], "title": "Deep Double Descent for Time Series Forecasting: Avoiding Undertrained Models", "abstract": "Deep learning models, particularly Transformers, have achieved impressive results in various domains, including time series forecasting. While existing time series literature primarily focuses on model architecture modifications and data augmentation techniques, this paper explores the training schema of deep learning models for time series; how models are trained regardless of their architecture. We perform extensive experiments to investigate the occurrence of deep double descent in several Transformer models trained on public time series data sets. We demonstrate epoch-wise deep double descent and that overfitting can be reverted using more epochs. Leveraging these findings, we achieve state-of-the-art results for long sequence time series forecasting in nearly 70% of the 72 benchmarks tested. This suggests that many models in the literature may possess untapped potential. Additionally, we introduce a taxonomy for classifying training schema modifications, covering data augmentation, model inputs, model targets, time series per model, and computational budget.", "authors": ["Valentino Assandri", "Sam Heshmati", "Burhaneddin Yaman", "Anton Iakovlev", "Ariel Emiliano Repetur"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2023-11-02", "url": "https://arxiv.org/abs/2311.01442", "pdf_url": "https://arxiv.org/pdf/2311.01442v3", "arxiv_id": "2311.01442", "doi": "10.48550/arXiv.2311.01442", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.0}
{"id": "1c5d25a1cb3a1e68689d486be8688a2521e55c3b9db2d082d25580fff3bea12e", "sources": ["arxiv", "semantic_scholar"], "title": "A Systematic Review for Transformer-based Long-term Series Forecasting", "abstract": "The emergence of deep learning has yielded noteworthy advancements in time series forecasting (TSF). Transformer architectures, in particular, have witnessed broad utilization and adoption in TSF tasks. Transformers have proven to be the most successful solution to extract the semantic correlations among the elements within a long sequence. Various variants have enabled transformer architecture to effectively handle long-term time series forecasting (LTSF) tasks. In this article, we first present a comprehensive overview of transformer architectures and their subsequent enhancements developed to address various LTSF tasks. Then, we summarize the publicly available LTSF datasets and relevant evaluation metrics. Furthermore, we provide valuable insights into the best practices and techniques for effectively training transformers in the context of time-series analysis. Lastly, we propose potential research directions in this rapidly evolving field.", "authors": ["Liyilei Su", "Xumin Zuo", "Rui Li", "Xin Wang", "Heng Zhao", "Bingding Huang"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2023-10-31", "url": "https://arxiv.org/abs/2310.20218", "pdf_url": "https://arxiv.org/pdf/2310.20218v1", "arxiv_id": "2310.20218", "doi": "10.1007/s10462-024-11044-2", "citation_count": 144, "influential_citation_count": 2, "has_code": false, "code_url": null, "venue": "Artificial Intelligence Review", "quality_score": 0.5403}
{"id": "7b911b439808d2c27d17249d20895b66f65910b1bafd3541c56dc97a8a93e445", "sources": ["arxiv", "semantic_scholar"], "title": "Graph Deep Learning for Time Series Forecasting", "abstract": "Graph deep learning methods have become popular tools to process collections of correlated time series. Unlike traditional multivariate forecasting methods, graph-based predictors leverage pairwise relationships by conditioning forecasts on graphs spanning the time series collection. The conditioning takes the form of architectural inductive biases on the forecasting architecture, resulting in a family of models called spatiotemporal graph neural networks. These biases allow for training global forecasting models on large collections of time series while localizing predictions w.r.t. each element in the set (nodes) by accounting for correlations among them (edges). Recent advances in graph neural networks and deep learning for time series forecasting make the adoption of such processing framework appealing and timely. However, most studies focus on refining existing architectures by exploiting modern deep-learning practices. Conversely, foundational and methodological aspects have not been subject to systematic investigation. To fill this void, this tutorial paper aims to introduce a comprehensive methodological framework formalizing the forecasting problem and providing design principles for graph-based predictors, as well as methods to assess their performance. In addition, together with an overview of the field, we provide design guidelines and best practices, as well as an in-depth discussion of open challenges and future directions.", "authors": ["Andrea Cini", "Ivan Marisca", "Daniele Zambon", "Cesare Alippi"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2023-10-24", "url": "https://arxiv.org/abs/2310.15978", "pdf_url": "https://arxiv.org/pdf/2310.15978v2", "arxiv_id": "2310.15978", "doi": "10.1145/3742784", "citation_count": 50, "influential_citation_count": 2, "has_code": false, "code_url": null, "venue": "ACM Computing Surveys", "quality_score": 0.4269}
{"id": "ecf1963e798d36b1acd6f2d79a4946e1c919aa0c183e4945d7cca91d22909f6b", "sources": ["arxiv", "semantic_scholar"], "title": "Mid-Long Term Daily Electricity Consumption Forecasting Based on Piecewise Linear Regression and Dilated Causal CNN", "abstract": "Daily electricity consumption forecasting is a classical problem. Existing forecasting algorithms tend to have decreased accuracy on special dates like holidays. This study decomposes the daily electricity consumption series into three components: trend, seasonal, and residual, and constructs a two-stage prediction method using piecewise linear regression as a filter and Dilated Causal CNN as a predictor. The specific steps involve setting breakpoints on the time axis and fitting the piecewise linear regression model with one-hot encoded information such as month, weekday, and holidays. For the challenging prediction of the Spring Festival, distance is introduced as a variable using a third-degree polynomial form in the model. The residual sequence obtained in the previous step is modeled using Dilated Causal CNN, and the final prediction of daily electricity consumption is the sum of the two-stage predictions. Experimental results demonstrate that this method achieves higher accuracy compared to existing approaches.", "authors": ["Zhou Lan", "Ben Liu", "Yi Feng", "Danhuang Dong", "Peng Zhang"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2023-10-23", "url": "https://arxiv.org/abs/2310.15204", "pdf_url": "https://arxiv.org/pdf/2310.15204v1", "arxiv_id": "2310.15204", "doi": "10.48550/arXiv.2310.15204", "citation_count": 2, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1193}
{"id": "ef6a4073129353d0ef86dacf7ba1e3b2c4e97b62bf3add7f58c15b5ace9e82dd", "sources": ["arxiv", "semantic_scholar"], "title": "Pyramidal Hidden Markov Model For Multivariate Time Series Forecasting", "abstract": "The Hidden Markov Model (HMM) can predict the future value of a time series based on its current and previous values, making it a powerful algorithm for handling various types of time series. Numerous studies have explored the improvement of HMM using advanced techniques, leading to the development of several variations of HMM. Despite these studies indicating the increased competitiveness of HMM compared to other advanced algorithms, few have recognized the significance and impact of incorporating multistep stochastic states into its performance. In this work, we propose a Pyramidal Hidden Markov Model (PHMM) that can capture multiple multistep stochastic states. Initially, a multistep HMM is designed for extracting short multistep stochastic states. Next, a novel time series forecasting structure is proposed based on PHMM, which utilizes pyramid-like stacking to adaptively identify long multistep stochastic states. By employing these two schemes, our model can effectively handle non-stationary and noisy data, while also establishing long-term dependencies for more accurate and comprehensive forecasting. The experimental results on diverse multivariate time series datasets convincingly demonstrate the superior performance of our proposed PHMM compared to its competitive peers in time series forecasting.", "authors": ["YeXin Huang"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2023-10-22", "url": "https://arxiv.org/abs/2310.14341", "pdf_url": "https://arxiv.org/pdf/2310.14341v2", "arxiv_id": "2310.14341", "doi": "10.48550/arXiv.2310.14341", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.0}
{"id": "1dd83ae107551b7389905f055021d77a26dbf12876796b7dbcde1079ab977f54", "sources": ["arxiv", "semantic_scholar"], "title": "A decoder-only foundation model for time-series forecasting", "abstract": "Motivated by recent advances in large language models for Natural Language Processing (NLP), we design a time-series foundation model for forecasting whose out-of-the-box zero-shot performance on a variety of public datasets comes close to the accuracy of state-of-the-art supervised forecasting models for each individual dataset. Our model is based on pretraining a patched-decoder style attention model on a large time-series corpus, and can work well across different forecasting history lengths, prediction lengths and temporal granularities.", "authors": ["Abhimanyu Das", "Weihao Kong", "Rajat Sen", "Yichen Zhou"], "categories": ["cs.CL", "cs.AI", "cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2023-10-14", "url": "https://arxiv.org/abs/2310.10688", "pdf_url": "https://arxiv.org/pdf/2310.10688v4", "arxiv_id": "2310.10688", "doi": "10.48550/arXiv.2310.10688", "citation_count": 717, "influential_citation_count": 119, "has_code": false, "code_url": null, "venue": "International Conference on Machine Learning", "quality_score": 1.0}
{"id": "a6539f836a7fb2ee427dbeac1b48de66e2e94d18727b33e91adb1f756c9cad08", "sources": ["arxiv", "semantic_scholar"], "title": "Lag-Llama: Towards Foundation Models for Probabilistic Time Series Forecasting", "abstract": "Over the past years, foundation models have caused a paradigm shift in machine learning due to their unprecedented capabilities for zero-shot and few-shot generalization. However, despite the success of foundation models in modalities such as natural language processing and computer vision, the development of foundation models for time series forecasting has lagged behind. We present Lag-Llama, a general-purpose foundation model for univariate probabilistic time series forecasting based on a decoder-only transformer architecture that uses lags as covariates. Lag-Llama is pretrained on a large corpus of diverse time series data from several domains, and demonstrates strong zero-shot generalization capabilities compared to a wide range of forecasting models on downstream datasets across domains. Moreover, when fine-tuned on relatively small fractions of such previously unseen datasets, Lag-Llama achieves state-of-the-art performance, outperforming prior deep learning approaches, emerging as the best general-purpose model on average. Lag-Llama serves as a strong contender to the current state-of-art in time series forecasting and paves the way for future advancements in foundation models tailored to time series data.", "authors": ["Kashif Rasul", "Arjun Ashok", "Andrew Robert Williams", "Hena Ghonia", "Rishika Bhagwatkar", "Arian Khorasani", "Mohammad Javad Darvishi Bayazi", "George Adamopoulos", "Roland Riachi", "Nadhir Hassen", "Marin Biloš", "Sahil Garg", "Anderson Schneider", "Nicolas Chapados", "Alexandre Drouin", "Valentina Zantedeschi", "Yuriy Nevmyvaka", "Irina Rish"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2023-10-12", "url": "https://arxiv.org/abs/2310.08278", "pdf_url": "https://arxiv.org/pdf/2310.08278v3", "arxiv_id": "2310.08278", "doi": null, "citation_count": 125, "influential_citation_count": 13, "has_code": true, "code_url": "https://github.com/time-series-foundation-models/lag-llama", "venue": null, "quality_score": 0.5731}
{"id": "8d899ecb105f65ca425558d5022bd2d25990595ed41c5c070fa52d35c53cfdf9", "sources": ["arxiv", "semantic_scholar"], "title": "Counterfactual Explanations for Time Series Forecasting", "abstract": "Among recent developments in time series forecasting methods, deep forecasting models have gained popularity as they can utilize hidden feature patterns in time series to improve forecasting performance. Nevertheless, the majority of current deep forecasting models are opaque, hence making it challenging to interpret the results. While counterfactual explanations have been extensively employed as a post-hoc approach for explaining classification models, their application to forecasting models still remains underexplored. In this paper, we formulate the novel problem of counterfactual generation for time series forecasting, and propose an algorithm, called ForecastCF, that solves the problem by applying gradient-based perturbations to the original time series. ForecastCF guides the perturbations by applying constraints to the forecasted values to obtain desired prediction outcomes. We experimentally evaluate ForecastCF using four state-of-the-art deep model architectures and compare to two baselines. Our results show that ForecastCF outperforms the baseline in terms of counterfactual validity and data manifold closeness. Overall, our findings suggest that ForecastCF can generate meaningful and relevant counterfactual explanations for various forecasting tasks.", "authors": ["Zhendong Wang", "Ioanna Miliou", "Isak Samsten", "Panagiotis Papapetrou"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2023-10-12", "url": "https://arxiv.org/abs/2310.08137", "pdf_url": "https://arxiv.org/pdf/2310.08137v1", "arxiv_id": "2310.08137", "doi": "10.1109/ICDM58522.2023.00180", "citation_count": 18, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "Industrial Conference on Data Mining", "quality_score": 0.3197}
{"id": "7e252ad66346c3ae08e75cf6ddf3874d90eda72986f3621ddfe8d0f3a1b91662", "sources": ["arxiv", "semantic_scholar"], "title": "iTransformer: Inverted Transformers Are Effective for Time Series Forecasting", "abstract": "The recent boom of linear forecasting models questions the ongoing passion for architectural modifications of Transformer-based forecasters. These forecasters leverage Transformers to model the global dependencies over temporal tokens of time series, with each token formed by multiple variates of the same timestamp. However, Transformers are challenged in forecasting series with larger lookback windows due to performance degradation and computation explosion. Besides, the embedding for each temporal token fuses multiple variates that represent potential delayed events and distinct physical measurements, which may fail in learning variate-centric representations and result in meaningless attention maps. In this work, we reflect on the competent duties of Transformer components and repurpose the Transformer architecture without any modification to the basic components. We propose iTransformer that simply applies the attention and feed-forward network on the inverted dimensions. Specifically, the time points of individual series are embedded into variate tokens which are utilized by the attention mechanism to capture multivariate correlations; meanwhile, the feed-forward network is applied for each variate token to learn nonlinear representations. The iTransformer model achieves state-of-the-art on challenging real-world datasets, which further empowers the Transformer family with promoted performance, generalization ability across different variates, and better utilization of arbitrary lookback windows, making it a nice alternative as the fundamental backbone of time series forecasting. Code is available at this repository: https://github.com/thuml/iTransformer.", "authors": ["Yong Liu", "Tengge Hu", "Haoran Zhang", "Haixu Wu", "Shiyu Wang", "Lintao Ma", "Mingsheng Long"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2023-10-10", "url": "https://arxiv.org/abs/2310.06625", "pdf_url": "https://arxiv.org/pdf/2310.06625v4", "arxiv_id": "2310.06625", "doi": "10.48550/arXiv.2310.06625", "citation_count": 1882, "influential_citation_count": 251, "has_code": true, "code_url": "https://github.com/thuml/iTransformer", "venue": "International Conference on Learning Representations", "quality_score": 1.0}
{"id": "c851a0cd29f56d00ded2b7711768a90438376cd983d4ff3d58d7fc7e58ae7d1d", "sources": ["arxiv", "semantic_scholar"], "title": "Introducing the Attribution Stability Indicator: a Measure for Time Series XAI Attributions", "abstract": "Given the increasing amount and general complexity of time series data in domains such as finance, weather forecasting, and healthcare, there is a growing need for state-of-the-art performance models that can provide interpretable insights into underlying patterns and relationships. Attribution techniques enable the extraction of explanations from time series models to gain insights but are hard to evaluate for their robustness and trustworthiness. We propose the Attribution Stability Indicator (ASI), a measure to incorporate robustness and trustworthiness as properties of attribution techniques for time series into account. We extend a perturbation analysis with correlations of the original time series to the perturbed instance and the attributions to include wanted properties in the measure. We demonstrate the wanted properties based on an analysis of the attributions in a dimension-reduced space and the ASI scores distribution over three whole time series classification datasets.", "authors": ["Udo Schlegel", "Daniel A. Keim"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2023-10-06", "url": "https://arxiv.org/abs/2310.04178", "pdf_url": "https://arxiv.org/pdf/2310.04178v1", "arxiv_id": "2310.04178", "doi": "10.48550/arXiv.2310.04178", "citation_count": 4, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.1747}
{"id": "07e1cb20e2c1768eb4b72cb9605acadbe576bdfea4eb347fd77e36f4c82ff744", "sources": ["arxiv", "semantic_scholar"], "title": "Toward a Foundation Model for Time Series Data", "abstract": "A foundation model is a machine learning model trained on a large and diverse set of data, typically using self-supervised learning-based pre-training techniques, that can be adapted to various downstream tasks. However, current research on time series pre-training has mostly focused on models pre-trained solely on data from a single domain, resulting in a lack of knowledge about other types of time series. However, current research on time series pre-training has predominantly focused on models trained exclusively on data from a single domain. As a result, these models possess domain-specific knowledge that may not be easily transferable to time series from other domains. In this paper, we aim to develop an effective time series foundation model by leveraging unlabeled samples from multiple domains. To achieve this, we repurposed the publicly available UCR Archive and evaluated four existing self-supervised learning-based pre-training methods, along with a novel method, on the datasets. We tested these methods using four popular neural network architectures for time series to understand how the pre-training methods interact with different network designs. Our experimental results show that pre-training improves downstream classification tasks by enhancing the convergence of the fine-tuning process. Furthermore, we found that the proposed pre-training method, when combined with the Transformer model, outperforms the alternatives.", "authors": ["Chin-Chia Michael Yeh", "Xin Dai", "Huiyuan Chen", "Yan Zheng", "Yujie Fan", "Audrey Der", "Vivian Lai", "Zhongfang Zhuang", "Junpeng Wang", "Liang Wang", "Wei Zhang"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2023-10-05", "url": "https://arxiv.org/abs/2310.03916", "pdf_url": "https://arxiv.org/pdf/2310.03916v1", "arxiv_id": "2310.03916", "doi": "10.1145/3583780.3615155", "citation_count": 44, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "International Conference on Information and Knowledge Management", "quality_score": 0.4133}
{"id": "116ceddb8dd9247fac3662a07987df47b934169300403c0af0c0dc411234b952", "sources": ["arxiv", "semantic_scholar"], "title": "Time-LLM: Time Series Forecasting by Reprogramming Large Language Models", "abstract": "Time series forecasting holds significant importance in many real-world dynamic systems and has been extensively studied. Unlike natural language process (NLP) and computer vision (CV), where a single large model can tackle multiple tasks, models for time series forecasting are often specialized, necessitating distinct designs for different tasks and applications. While pre-trained foundation models have made impressive strides in NLP and CV, their development in time series domains has been constrained by data sparsity. Recent studies have revealed that large language models (LLMs) possess robust pattern recognition and reasoning abilities over complex sequences of tokens. However, the challenge remains in effectively aligning the modalities of time series data and natural language to leverage these capabilities. In this work, we present Time-LLM, a reprogramming framework to repurpose LLMs for general time series forecasting with the backbone language models kept intact. We begin by reprogramming the input time series with text prototypes before feeding it into the frozen LLM to align the two modalities. To augment the LLM's ability to reason with time series data, we propose Prompt-as-Prefix (PaP), which enriches the input context and directs the transformation of reprogrammed input patches. The transformed time series patches from the LLM are finally projected to obtain the forecasts. Our comprehensive evaluations demonstrate that Time-LLM is a powerful time series learner that outperforms state-of-the-art, specialized forecasting models. Moreover, Time-LLM excels in both few-shot and zero-shot learning scenarios.", "authors": ["Ming Jin", "Shiyu Wang", "Lintao Ma", "Zhixuan Chu", "James Y. Zhang", "Xiaoming Shi", "Pin-Yu Chen", "Yuxuan Liang", "Yuan-Fang Li", "Shirui Pan", "Qingsong Wen"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2023-10-03", "url": "https://arxiv.org/abs/2310.01728", "pdf_url": "https://arxiv.org/pdf/2310.01728v2", "arxiv_id": "2310.01728", "doi": "10.48550/arXiv.2310.01728", "citation_count": 999, "influential_citation_count": 128, "has_code": false, "code_url": null, "venue": "International Conference on Learning Representations", "quality_score": 1.0}
{"id": "764b8d3bba7c82e5d3b6e967c9114633d22868ec49bd334f69f80fcf604dcb91", "sources": ["arxiv", "semantic_scholar"], "title": "Modality-aware Transformer for Financial Time series Forecasting", "abstract": "Time series forecasting presents a significant challenge, particularly when its accuracy relies on external data sources rather than solely on historical values. This issue is prevalent in the financial sector, where the future behavior of time series is often intricately linked to information derived from various textual reports and a multitude of economic indicators. In practice, the key challenge lies in constructing a reliable time series forecasting model capable of harnessing data from diverse sources and extracting valuable insights to predict the target time series accurately. In this work, we tackle this challenging problem and introduce a novel multimodal transformer-based model named the \\textit{Modality-aware Transformer}. Our model excels in exploring the power of both categorical text and numerical timeseries to forecast the target time series effectively while providing insights through its neural attention mechanism. To achieve this, we develop feature-level attention layers that encourage the model to focus on the most relevant features within each data modality. By incorporating the proposed feature-level attention, we develop a novel Intra-modal multi-head attention (MHA), Inter-modal MHA and Target-modal MHA in a way that both feature and temporal attentions are incorporated in MHAs. This enables the MHAs to generate temporal attentions with consideration of modality and feature importance which leads to more informative embeddings. The proposed modality-aware structure enables the model to effectively exploit information within each modality as well as foster cross-modal understanding. Our extensive experiments on financial datasets demonstrate that Modality-aware Transformer outperforms existing methods, offering a novel and practical solution to the complex challenges of multi-modal financial time series forecasting.", "authors": ["Hajar Emami", "Xuan-Hong Dang", "Yousaf Shah", "Petros Zerfos"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2023-10-02", "url": "https://arxiv.org/abs/2310.01232", "pdf_url": "https://arxiv.org/pdf/2310.01232v2", "arxiv_id": "2310.01232", "doi": "10.1145/3677052.3698654", "citation_count": 20, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "International Conference on AI in Finance", "quality_score": 0.3306}
{"id": "1228aed58d0055d19b8c959ff70dd01706e62c39d913f12f1e0e3648a1b174a6", "sources": ["arxiv", "semantic_scholar"], "title": "Jais and Jais-chat: Arabic-Centric Foundation and Instruction-Tuned Open Generative Large Language Models", "abstract": "We introduce Jais and Jais-chat, new state-of-the-art Arabic-centric foundation and instruction-tuned open generative large language models (LLMs). The models are based on the GPT-3 decoder-only architecture and are pretrained on a mixture of Arabic and English texts, including source code in various programming languages. With 13 billion parameters, they demonstrate better knowledge and reasoning capabilities in Arabic than any existing open Arabic and multilingual models by a sizable margin, based on extensive evaluation. Moreover, the models are competitive in English compared to English-centric open models of similar size, despite being trained on much less English data. We provide a detailed description of the training, the tuning, the safety alignment, and the evaluation of the models. We release two open versions of the model -- the foundation Jais model, and an instruction-tuned Jais-chat variant -- with the aim of promoting research on Arabic LLMs. Available at https://huggingface.co/inception-mbzuai/jais-13b-chat", "authors": ["Neha Sengupta", "Sunil Kumar Sahu", "Bokang Jia", "Satheesh Katipomu", "Haonan Li", "Fajri Koto", "William Marshall", "Gurpreet Gosal", "Cynthia Liu", "Zhiming Chen", "Osama Mohammed Afzal", "Samta Kamboj", "Onkar Pandit", "Rahul Pal", "Lalit Pradhan", "Zain Muhammad Mujahid", "Massa Baali", "Xudong Han", "Sondos Mahmoud Bsharat", "Alham Fikri Aji", "Zhiqiang Shen", "Zhengzhong Liu", "Natalia Vassilieva", "Joel Hestness", "Andy Hock", "Andrew Feldman", "Jonathan Lee", "Andrew Jackson", "Hector Xuguang Ren", "Preslav Nakov", "Timothy Baldwin", "Eric Xing"], "categories": ["cs.CL", "cs.AI", "cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2023-08-30", "url": "https://arxiv.org/abs/2308.16149", "pdf_url": "https://arxiv.org/pdf/2308.16149v2", "arxiv_id": "2308.16149", "doi": "10.48550/arXiv.2308.16149", "citation_count": 81, "influential_citation_count": 8, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.4785}
{"id": "2ae4c04bf55470eb3fd564e5d30be3697aa53de55782935c8c95df24a3e49459", "sources": ["arxiv", "semantic_scholar"], "title": "Hierarchical Time Series Forecasting with Bayesian Modeling", "abstract": "We encounter time series data in many domains such as finance, physics, business, and weather. One of the main tasks of time series analysis, one that helps to take informed decisions under uncertainty, is forecasting. Time series are often hierarchically structured, e.g., a company sales might be broken down into different regions, and each region into different stores. In some cases the number of series in the hierarchy is too big to fit in a single model to produce forecasts in relevant time, and a decentralized approach is beneficial. One way to do this is to train independent forecasting models for each series and for some summary statistics series implied by the hierarchy (e.g. the sum of all series) and to pass those models to a reconciliation algorithm to improve those forecasts by sharing information between the series. In this work we focus on the reconciliation step, and propose a method to do so from a Bayesian perspective - Bayesian forecast reconciliation. We also define the common case of linear Gaussian reconciliation, where the forecasts are Gaussian and the hierarchy has linear structure, and show that we can compute reconciliation in closed form. We evaluate these methods on synthetic and real data sets, and compare them to other work in this field.", "authors": ["Gal Elgavish"], "categories": ["cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2023-08-28", "url": "https://arxiv.org/abs/2308.14719", "pdf_url": "https://arxiv.org/pdf/2308.14719v1", "arxiv_id": "2308.14719", "doi": "10.48550/arXiv.2308.14719", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.0753}
{"id": "3244572c3f923ab4c9c7cda00c5d20195daaa94101c990adba10b2a69037dd2e", "sources": ["arxiv", "semantic_scholar"], "title": "Multi-scale Transformer Pyramid Networks for Multivariate Time Series Forecasting", "abstract": "Multivariate Time Series (MTS) forecasting involves modeling temporal dependencies within historical records. Transformers have demonstrated remarkable performance in MTS forecasting due to their capability to capture long-term dependencies. However, prior work has been confined to modeling temporal dependencies at either a fixed scale or multiple scales that exponentially increase (most with base 2). This limitation hinders their effectiveness in capturing diverse seasonalities, such as hourly and daily patterns. In this paper, we introduce a dimension invariant embedding technique that captures short-term temporal dependencies and projects MTS data into a higher-dimensional space, while preserving the dimensions of time steps and variables in MTS data. Furthermore, we present a novel Multi-scale Transformer Pyramid Network (MTPNet), specifically designed to effectively capture temporal dependencies at multiple unconstrained scales. The predictions are inferred from multi-scale latent representations obtained from transformers at various scales. Extensive experiments on nine benchmark datasets demonstrate that the proposed MTPNet outperforms recent state-of-the-art methods.", "authors": ["Yifan Zhang", "Rui Wu", "Sergiu M. Dascalu", "Frederick C. Harris"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2023-08-23", "url": "https://arxiv.org/abs/2308.11946", "pdf_url": "https://arxiv.org/pdf/2308.11946v1", "arxiv_id": "2308.11946", "doi": "10.1109/ACCESS.2024.3357693", "citation_count": 38, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "IEEE Access", "quality_score": 0.3978}
{"id": "410010e76d4d522eba748a8f5dba35408df62df61de26d0490ac5991e7c77a18", "sources": ["arxiv", "semantic_scholar"], "title": "PETformer: Long-term Time Series Forecasting via Placeholder-enhanced Transformer", "abstract": "Recently, the superiority of Transformer for long-term time series forecasting (LTSF) tasks has been challenged, particularly since recent work has shown that simple models can outperform numerous Transformer-based approaches. This suggests that a notable gap remains in fully leveraging the potential of Transformer in LTSF tasks. Consequently, this study investigates key issues when applying Transformer to LTSF, encompassing aspects of temporal continuity, information density, and multi-channel relationships. We introduce the Placeholder-enhanced Technique (PET) to enhance the computational efficiency and predictive accuracy of Transformer in LTSF tasks. Furthermore, we delve into the impact of larger patch strategies and channel interaction strategies on Transformer's performance, specifically Long Sub-sequence Division (LSD) and Multi-channel Separation and Interaction (MSI). These strategies collectively constitute a novel model termed PETformer. Extensive experiments have demonstrated that PETformer achieves state-of-the-art performance on eight commonly used public datasets for LTSF, surpassing all existing models. The insights and enhancement methodologies presented in this paper serve as valuable reference points and sources of inspiration for future research endeavors.", "authors": ["Shengsheng Lin", "Weiwei Lin", "Wentai Wu", "Songbo Wang", "Yongxiang Wang"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2023-08-09", "url": "https://arxiv.org/abs/2308.04791", "pdf_url": "https://arxiv.org/pdf/2308.04791v2", "arxiv_id": "2308.04791", "doi": "10.1109/TETCI.2024.3502437", "citation_count": 57, "influential_citation_count": 3, "has_code": false, "code_url": null, "venue": "IEEE Transactions on Emerging Topics in Computational Intelligence", "quality_score": 0.4409}
{"id": "b72a3e913312cdf6bc83983100c5b80812c5b6a57da854e94248ae1a9b9efdff", "sources": ["arxiv", "semantic_scholar"], "title": "Conformal PID Control for Time Series Prediction", "abstract": "We study the problem of uncertainty quantification for time series prediction, with the goal of providing easy-to-use algorithms with formal guarantees. The algorithms we present build upon ideas from conformal prediction and control theory, are able to prospectively model conformal scores in an online setting, and adapt to the presence of systematic errors due to seasonality, trends, and general distribution shifts. Our theory both simplifies and strengthens existing analyses in online conformal prediction. Experiments on 4-week-ahead forecasting of statewide COVID-19 death counts in the U.S. show an improvement in coverage over the ensemble forecaster used in official CDC communications. We also run experiments on predicting electricity demand, market returns, and temperature using autoregressive, Theta, Prophet, and Transformer models. We provide an extendable codebase for testing our methods and for the integration of new algorithms, data sets, and forecasting rules.", "authors": ["Anastasios N. Angelopoulos", "Emmanuel J. Candes", "Ryan J. Tibshirani"], "categories": ["cs.LG", "eess.SY", "stat.ME", "stat.ML"], "fields_of_study": ["Computer Science", "Engineering", "Mathematics"], "published_date": "2023-07-31", "url": "https://arxiv.org/abs/2307.16895", "pdf_url": "https://arxiv.org/pdf/2307.16895v1", "arxiv_id": "2307.16895", "doi": "10.48550/arXiv.2307.16895", "citation_count": 134, "influential_citation_count": 15, "has_code": true, "code_url": "https://github.com/aangelopoulos/conformal-time-series", "venue": "Neural Information Processing Systems", "quality_score": 0.6021}
{"id": "9a22587367d3b8b9ce7a7366db87a9badcebf1f2e714d3c7fcef38ece2c9674e", "sources": ["arxiv", "semantic_scholar"], "title": "Differential Evolution Algorithm based Hyper-Parameters Selection of Transformer Neural Network Model for Load Forecasting", "abstract": "Accurate load forecasting plays a vital role in numerous sectors, but accurately capturing the complex dynamics of dynamic power systems remains a challenge for traditional statistical models. For these reasons, time-series models (ARIMA) and deep-learning models (ANN, LSTM, GRU, etc.) are commonly deployed and often experience higher success. In this paper, we analyze the efficacy of the recently developed Transformer-based Neural Network model in Load forecasting. Transformer models have the potential to improve Load forecasting because of their ability to learn long-range dependencies derived from their Attention Mechanism. We apply several metaheuristics namely Differential Evolution to find the optimal hyperparameters of the Transformer-based Neural Network to produce accurate forecasts. Differential Evolution provides scalable, robust, global solutions to non-differentiable, multi-objective, or constrained optimization problems. Our work compares the proposed Transformer based Neural Network model integrated with different metaheuristic algorithms by their performance in Load forecasting based on numerical metrics such as Mean Squared Error (MSE) and Mean Absolute Percentage Error (MAPE). Our findings demonstrate the potential of metaheuristic-enhanced Transformer-based Neural Network models in Load forecasting accuracy and provide optimal hyperparameters for each model.", "authors": ["Anuvab Sen", "Arul Rhik Mazumder", "Udayon Sen"], "categories": ["cs.NE", "cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2023-07-28", "url": "https://arxiv.org/abs/2307.15299", "pdf_url": "https://arxiv.org/pdf/2307.15299v5", "arxiv_id": "2307.15299", "doi": "10.1109/SSCI52147.2023.10371846", "citation_count": 5, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "IEEE Symposium Series on Computational Intelligence", "quality_score": 0.1945}
{"id": "a4ba9ac0de00a7482f0353c675ab2255f266d91c8de3e936f128ed468b463847", "sources": ["arxiv", "semantic_scholar"], "title": "Predict, Refine, Synthesize: Self-Guiding Diffusion Models for Probabilistic Time Series Forecasting", "abstract": "Diffusion models have achieved state-of-the-art performance in generative modeling tasks across various domains. Prior works on time series diffusion models have primarily focused on developing conditional models tailored to specific forecasting or imputation tasks. In this work, we explore the potential of task-agnostic, unconditional diffusion models for several time series applications. We propose TSDiff, an unconditionally-trained diffusion model for time series. Our proposed self-guidance mechanism enables conditioning TSDiff for downstream tasks during inference, without requiring auxiliary networks or altering the training procedure. We demonstrate the effectiveness of our method on three different time series tasks: forecasting, refinement, and synthetic data generation. First, we show that TSDiff is competitive with several task-specific conditional forecasting methods (predict). Second, we leverage the learned implicit probability density of TSDiff to iteratively refine the predictions of base forecasters with reduced computational overhead over reverse diffusion (refine). Notably, the generative performance of the model remains intact -- downstream forecasters trained on synthetic samples from TSDiff outperform forecasters that are trained on samples from other state-of-the-art generative time series models, occasionally even outperforming models trained on real data (synthesize).", "authors": ["Marcel Kollovieh", "Abdul Fatir Ansari", "Michael Bohlke-Schneider", "Jasper Zschiegner", "Hao Wang", "Yuyang Wang"], "categories": ["cs.LG", "cs.AI", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2023-07-21", "url": "https://arxiv.org/abs/2307.11494", "pdf_url": "https://arxiv.org/pdf/2307.11494v3", "arxiv_id": "2307.11494", "doi": "10.48550/arXiv.2307.11494", "citation_count": 115, "influential_citation_count": 13, "has_code": true, "code_url": "https://github.com/amazon-science/unconditional-time-series-diffusion", "venue": "Neural Information Processing Systems", "quality_score": 0.5731}
{"id": "7695fb5826451ed7f5eb32fcc1ee8712aad13857314e185aedb3e17ccde1954a", "sources": ["arxiv", "semantic_scholar"], "title": "GBT: Two-stage transformer framework for non-stationary time series forecasting", "abstract": "This paper shows that time series forecasting Transformer (TSFT) suffers from severe over-fitting problem caused by improper initialization method of unknown decoder inputs, esp. when handling non-stationary time series. Based on this observation, we propose GBT, a novel two-stage Transformer framework with Good Beginning. It decouples the prediction process of TSFT into two stages, including Auto-Regression stage and Self-Regression stage to tackle the problem of different statistical properties between input and prediction sequences.Prediction results of Auto-Regression stage serve as a Good Beginning, i.e., a better initialization for inputs of Self-Regression stage. We also propose Error Score Modification module to further enhance the forecasting capability of the Self-Regression stage in GBT. Extensive experiments on seven benchmark datasets demonstrate that GBT outperforms SOTA TSFTs (FEDformer, Pyraformer, ETSformer, etc.) and many other forecasting models (SCINet, N-HiTS, etc.) with only canonical attention and convolution while owning less time and space complexity. It is also general enough to couple with these models to strengthen their forecasting capability. The source code is available at: https://github.com/OrigamiSL/GBT", "authors": ["Li Shen", "Yuning Wei", "Yangzhu Wang"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science", "Medicine"], "published_date": "2023-07-17", "url": "https://arxiv.org/abs/2307.08302", "pdf_url": "https://arxiv.org/pdf/2307.08302v1", "arxiv_id": "2307.08302", "doi": "10.1016/j.neunet.2023.06.044", "citation_count": 40, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/OrigamiSL/GBT", "venue": "Neural Networks", "quality_score": 0.4032}
{"id": "bea6844a6d9a15b007cf04ae71b169a1031b188e6c06233edd4993bda94ed92e", "sources": ["arxiv", "semantic_scholar"], "title": "Autoregressive with Slack Time Series Model for Forecasting a Partially-Observed Dynamical Time Series", "abstract": "This study delves into the domain of dynamical systems, specifically the forecasting of dynamical time series defined through an evolution function. Traditional approaches in this area predict the future behavior of dynamical systems by inferring the evolution function. However, these methods may confront obstacles due to the presence of missing variables, which are usually attributed to challenges in measurement and a partial understanding of the system of interest. To overcome this obstacle, we introduce the autoregressive with slack time series (ARS) model, that simultaneously estimates the evolution function and imputes missing variables as a slack time series. Assuming time-invariance and linearity in the (underlying) entire dynamical time series, our experiments demonstrate the ARS model's capability to forecast future time series. From a theoretical perspective, we prove that a 2-dimensional time-invariant and linear system can be reconstructed by utilizing observations from a single, partially observed dimension of the system.", "authors": ["Akifumi Okuno", "Yuya Morishita", "Yoh-ichi Mototake"], "categories": ["stat.ME", "cs.LG", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2023-06-28", "url": "https://arxiv.org/abs/2306.16593", "pdf_url": "https://arxiv.org/pdf/2306.16593v2", "arxiv_id": "2306.16593", "doi": "10.1109/ACCESS.2024.3365724", "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "IEEE Access", "quality_score": 0.0}
{"id": "71141085c83ee6d1988ec92b950334a2c7de835db1456c128feaf2304215dc97", "sources": ["arxiv", "semantic_scholar"], "title": "What Constitutes Good Contrastive Learning in Time-Series Forecasting?", "abstract": "In recent years, the introduction of self-supervised contrastive learning (SSCL) has demonstrated remarkable improvements in representation learning across various domains, including natural language processing and computer vision. By leveraging the inherent benefits of self-supervision, SSCL enables the pre-training of representation models using vast amounts of unlabeled data. Despite these advances, there remains a significant gap in understanding the impact of different SSCL strategies on time series forecasting performance, as well as the specific benefits that SSCL can bring. This paper aims to address these gaps by conducting a comprehensive analysis of the effectiveness of various training variables, including different SSCL algorithms, learning strategies, model architectures, and their interplay. Additionally, to gain deeper insights into the improvements brought about by SSCL in the context of time-series forecasting, a qualitative analysis of the empirical receptive field is performed. Through our experiments, we demonstrate that the end-to-end training of a Transformer model using the Mean Squared Error (MSE) loss and SSCL emerges as the most effective approach in time series forecasting. Notably, the incorporation of the contrastive objective enables the model to prioritize more pertinent information for forecasting, such as scale and periodic relationships. These findings contribute to a better understanding of the benefits of SSCL in time series forecasting and provide valuable insights for future research in this area. Our codes are available at https://github.com/chiyuzhang94/contrastive_learning_time-series_e2e.", "authors": ["Chiyu Zhang", "Qi Yan", "Lili Meng", "Tristan Sylvain"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2023-06-21", "url": "https://arxiv.org/abs/2306.12086", "pdf_url": "https://arxiv.org/pdf/2306.12086v2", "arxiv_id": "2306.12086", "doi": "10.48550/arXiv.2306.12086", "citation_count": 2, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/chiyuzhang94/contrastive_learning_time-series_e2e", "venue": "arXiv.org", "quality_score": 0.1193}
{"id": "2f9187d087c02b0c2028da8a5a47049cedb4c09bec01983a702ac8f0384c2d1e", "sources": ["arxiv", "semantic_scholar"], "title": "Transformer Training Strategies for Forecasting Multiple Load Time Series", "abstract": "In the smart grid of the future, accurate load forecasts on the level of individual clients can help to balance supply and demand locally and to prevent grid outages. While the number of monitored clients will increase with the ongoing smart meter rollout, the amount of data per client will always be limited. We evaluate whether a Transformer load forecasting model benefits from a transfer learning strategy, where a global univariate model is trained on the load time series from multiple clients. In experiments with two datasets containing load time series from several hundred clients, we find that the global training strategy is superior to the multivariate and local training strategies used in related work. On average, the global training strategy results in 21.8% and 12.8% lower forecasting errors than the two other strategies, measured across forecasting horizons from one day to one month into the future. A comparison to linear models, multi-layer perceptrons and LSTMs shows that Transformers are effective for load forecasting when they are trained with the global training strategy.", "authors": ["Matthias Hertel", "Maximilian Beichter", "Benedikt Heidrich", "Oliver Neumann", "Benjamin Schäfer", "Ralf Mikut", "Veit Hagenmeyer"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2023-06-19", "url": "https://arxiv.org/abs/2306.10891", "pdf_url": "https://arxiv.org/pdf/2306.10891v3", "arxiv_id": "2306.10891", "doi": "10.1186/s42162-023-00278-z", "citation_count": 32, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "Energy Informatics", "quality_score": 0.3796}
{"id": "563eb3ec192c7e004e21a06757933e841d0df8ff293748fcebec379d02d0386a", "sources": ["arxiv", "semantic_scholar"], "title": "TSMixer: Lightweight MLP-Mixer Model for Multivariate Time Series Forecasting", "abstract": "Transformers have gained popularity in time series forecasting for their ability to capture long-sequence interactions. However, their high memory and computing requirements pose a critical bottleneck for long-term forecasting. To address this, we propose TSMixer, a lightweight neural architecture exclusively composed of multi-layer perceptron (MLP) modules for multivariate forecasting and representation learning on patched time series. Inspired by MLP-Mixer's success in computer vision, we adapt it for time series, addressing challenges and introducing validated components for enhanced accuracy. This includes a novel design paradigm of attaching online reconciliation heads to the MLP-Mixer backbone, for explicitly modeling the time-series properties such as hierarchy and channel-correlations. We also propose a novel Hybrid channel modeling and infusion of a simple gating approach to effectively handle noisy channel interactions and generalization across diverse datasets. By incorporating these lightweight components, we significantly enhance the learning capability of simple MLP structures, outperforming complex Transformer models with minimal computing usage. Moreover, TSMixer's modular design enables compatibility with both supervised and masked self-supervised learning methods, making it a promising building block for time-series Foundation Models. TSMixer outperforms state-of-the-art MLP and Transformer models in forecasting by a considerable margin of 8-60%. It also outperforms the latest strong benchmarks of Patch-Transformer models (by 1-2%) with a significant reduction in memory and runtime (2-3X). The source code of our model is officially released as PatchTSMixer in the HuggingFace. Model: https://huggingface.co/docs/transformers/main/en/model_doc/patchtsmixer Examples: https://github.com/ibm/tsfm/#notebooks-links", "authors": ["Vijay Ekambaram", "Arindam Jati", "Nam Nguyen", "Phanwadee Sinthong", "Jayant Kalagnanam"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2023-06-14", "url": "https://arxiv.org/abs/2306.09364", "pdf_url": "https://arxiv.org/pdf/2306.09364v4", "arxiv_id": "2306.09364", "doi": "10.1145/3580305.3599533", "citation_count": 376, "influential_citation_count": 13, "has_code": true, "code_url": "https://github.com/ibm/tsfm/#notebooks-links", "venue": "Knowledge Discovery and Data Mining", "quality_score": 0.6441}
{"id": "50080bd7438f712ab80d6fba01af53154de285daa2eaecb7020d4abb30de1d30", "sources": ["arxiv", "semantic_scholar"], "title": "Self-Interpretable Time Series Prediction with Counterfactual Explanations", "abstract": "Interpretable time series prediction is crucial for safety-critical areas such as healthcare and autonomous driving. Most existing methods focus on interpreting predictions by assigning important scores to segments of time series. In this paper, we take a different and more challenging route and aim at developing a self-interpretable model, dubbed Counterfactual Time Series (CounTS), which generates counterfactual and actionable explanations for time series predictions. Specifically, we formalize the problem of time series counterfactual explanations, establish associated evaluation protocols, and propose a variational Bayesian deep learning model equipped with counterfactual inference capability of time series abduction, action, and prediction. Compared with state-of-the-art baselines, our self-interpretable model can generate better counterfactual explanations while maintaining comparable prediction accuracy.", "authors": ["Jingquan Yan", "Hao Wang"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2023-06-09", "url": "https://arxiv.org/abs/2306.06024", "pdf_url": "https://arxiv.org/pdf/2306.06024v3", "arxiv_id": "2306.06024", "doi": "10.48550/arXiv.2306.06024", "citation_count": 29, "influential_citation_count": 2, "has_code": true, "code_url": "https://github.com/Wang-ML-Lab/self-interpretable-time-series", "venue": "International Conference on Machine Learning", "quality_score": 0.3693}
{"id": "fcd8c4cf17213c1b8269af8a85774a6da7c9a0729f78784ec5bd1d5a2415c7be", "sources": ["arxiv", "semantic_scholar"], "title": "Client: Cross-variable Linear Integrated Enhanced Transformer for Multivariate Long-Term Time Series Forecasting", "abstract": "Long-term time series forecasting (LTSF) is a crucial aspect of modern society, playing a pivotal role in facilitating long-term planning and developing early warning systems. While many Transformer-based models have recently been introduced for LTSF, a doubt have been raised regarding the effectiveness of attention modules in capturing cross-time dependencies. In this study, we design a mask-series experiment to validate this assumption and subsequently propose the \"Cross-variable Linear Integrated ENhanced Transformer for Multivariate Long-Term Time Series Forecasting\" (Client), an advanced model that outperforms both traditional Transformer-based models and linear models. Client employs linear modules to learn trend information and attention modules to capture cross-variable dependencies. Meanwhile, it simplifies the embedding and position encoding layers and replaces the decoder module with a projection layer. Essentially, Client incorporates non-linearity and cross-variable dependencies, which sets it apart from conventional linear models and Transformer-based models. Extensive experiments with nine real-world datasets have confirmed the SOTA performance of Client with the least computation time and memory consumption compared with the previous Transformer-based models. Our code is available at https://github.com/daxin007/Client.", "authors": ["Jiaxin Gao", "Wenbo Hu", "Yuntian Chen"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2023-05-30", "url": "https://arxiv.org/abs/2305.18838", "pdf_url": "https://arxiv.org/pdf/2305.18838v1", "arxiv_id": "2305.18838", "doi": "10.48550/arXiv.2305.18838", "citation_count": 15, "influential_citation_count": 1, "has_code": true, "code_url": "https://github.com/daxin007/Client", "venue": "arXiv.org", "quality_score": 0.301}
{"id": "1581b6128b3ad6d0ae0b6d4ac21563dbe7ed5ea3a98e2e3e63365a8572f57d41", "sources": ["arxiv", "semantic_scholar"], "title": "Adaptive Sparsity Level during Training for Efficient Time Series Forecasting with Transformers", "abstract": "Efficient time series forecasting has become critical for real-world applications, particularly with deep neural networks (DNNs). Efficiency in DNNs can be achieved through sparse connectivity and reducing the model size. However, finding the sparsity level automatically during training remains challenging due to the heterogeneity in the loss-sparsity tradeoffs across the datasets. In this paper, we propose \\enquote{\\textbf{P}runing with \\textbf{A}daptive \\textbf{S}parsity \\textbf{L}evel} (\\textbf{PALS}), to automatically seek a decent balance between loss and sparsity, all without the need for a predefined sparsity level. PALS draws inspiration from sparse training and during-training methods. It introduces the novel \"expand\" mechanism in training sparse neural networks, allowing the model to dynamically shrink, expand, or remain stable to find a proper sparsity level. In this paper, we focus on achieving efficiency in transformers known for their excellent time series forecasting performance but high computational cost. Nevertheless, PALS can be applied directly to any DNN. To this aim, we demonstrate its effectiveness also on the DLinear model. Experimental results on six benchmark datasets and five state-of-the-art (SOTA) transformer variants show that PALS substantially reduces model size while maintaining comparable performance to the dense model. More interestingly, PALS even outperforms the dense model, in \\textcolor{blue}{12} and \\textcolor{blue}{14} cases out of 30 cases in terms of MSE and MAE loss, respectively, while reducing \\textcolor{blue}{65\\%} parameter count and \\textcolor{blue}{63\\%} FLOPs on average. Our code and supplementary material are available on Github\\footnote{\\tiny \\url{https://github.com/zahraatashgahi/PALS}}.", "authors": ["Zahra Atashgahi", "Mykola Pechenizkiy", "Raymond Veldhuis", "Decebal Constantin Mocanu"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2023-05-28", "url": "https://arxiv.org/abs/2305.18382", "pdf_url": "https://arxiv.org/pdf/2305.18382v2", "arxiv_id": "2305.18382", "doi": "10.48550/arXiv.2305.18382", "citation_count": 2, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/zahraatashgahi/PALS}}", "venue": null, "quality_score": 0.1193}
{"id": "f6fb9350adc71c514555c39dc66d20d724b2f5d3df265590d9ec6eba30942772", "sources": ["arxiv", "semantic_scholar"], "title": "Stecformer: Spatio-temporal Encoding Cascaded Transformer for Multivariate Long-term Time Series Forecasting", "abstract": "Multivariate long-term time series forecasting is of great application across many domains, such as energy consumption and weather forecasting. With the development of transformer-based methods, the performance of multivariate long-term time series forecasting has been significantly improved, however, the study of spatial features extracting in transformer-based model is rare and the consistency of different prediction periods is unsatisfactory due to the large span. In this work, we propose a complete solution to address these problems in terms of feature extraction and target prediction. For extraction, we design an efficient spatio-temporal encoding extractor including a semi-adaptive graph to acquire sufficient spatio-temporal information. For prediction, we propose a Cascaded Decoding Predictor (CDP) to strengthen the correlation between different intervals, which can also be utilized as a generic component to improve the performance of transformer-based methods. The proposed method, termed as Spatio-temporal Encoding Cascaded Transformer (Stecformer), achieving a notable gap over the baseline model and is comparable with the state-of-the-art performance of transformer-based methods on five benchmark datasets. We hope our attempt will serve as a regular configuration in multivariate long-term time series forecasting in the future.", "authors": ["Zheng Sun", "Yi Wei", "Wenxiao Jia", "Long Yu"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2023-05-25", "url": "https://arxiv.org/abs/2305.16370", "pdf_url": "https://arxiv.org/pdf/2305.16370v1", "arxiv_id": "2305.16370", "doi": "10.48550/arXiv.2305.16370", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.0753}
{"id": "431a0f2686c808d482bb0503c01b8761b06dea21f90fc5926002dba174f9c263", "sources": ["arxiv", "semantic_scholar"], "title": "CARD: Channel Aligned Robust Blend Transformer for Time Series Forecasting", "abstract": "Recent studies have demonstrated the great power of Transformer models for time series forecasting. One of the key elements that lead to the transformer's success is the channel-independent (CI) strategy to improve the training robustness. However, the ignorance of the correlation among different channels in CI would limit the model's forecasting capacity. In this work, we design a special Transformer, i.e., Channel Aligned Robust Blend Transformer (CARD for short), that addresses key shortcomings of CI type Transformer in time series forecasting. First, CARD introduces a channel-aligned attention structure that allows it to capture both temporal correlations among signals and dynamical dependence among multiple variables over time. Second, in order to efficiently utilize the multi-scale knowledge, we design a token blend module to generate tokens with different resolutions. Third, we introduce a robust loss function for time series forecasting to alleviate the potential overfitting issue. This new loss function weights the importance of forecasting over a finite horizon based on prediction uncertainties. Our evaluation of multiple long-term and short-term forecasting datasets demonstrates that CARD significantly outperforms state-of-the-art time series forecasting methods. The code is available at the following repository:https://github.com/wxie9/CARD", "authors": ["Wang Xue", "Tian Zhou", "Qingsong Wen", "Jinyang Gao", "Bolin Ding", "Rong Jin"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2023-05-20", "url": "https://arxiv.org/abs/2305.12095", "pdf_url": "https://arxiv.org/pdf/2305.12095v5", "arxiv_id": "2305.12095", "doi": null, "citation_count": 99, "influential_citation_count": 7, "has_code": true, "code_url": "https://github.com/wxie9/CARD", "venue": "International Conference on Learning Representations", "quality_score": 0.5}
{"id": "ca63b54038a9b2ecae9eab76daa293c4435ebdc672de3cadccfb173255985930", "sources": ["arxiv", "semantic_scholar"], "title": "Revisiting Long-term Time Series Forecasting: An Investigation on Linear Mapping", "abstract": "Introduction: Long-term time series forecasting (LTSF) has gained significant attention in recent years. While various specialized designs exist for capturing temporal dependency, recent studies have shown that even a single linear layer can achieve competitive performance. This paper investigates the intrinsic effectiveness of recent LTSF approaches and reveals the critical role of affine mapping. Materials and methods: We conduct comprehensive experiments on both simulated and real-world datasets to analyze the components of state-of-the-art models. A theoretical analysis is provided to explain the working mechanisms of affine mapping in periodic signal forecasting. We evaluate the impact of reversible normalization and input horizon extension on model robustness. Results: We find that (1) affine mapping dominates forecasting performance across commonly utilized benchmarks, with models learning similar transition matrices from input to output; (2) affine mapping effectively captures periodic patterns but struggles with non-periodic signals or time series with varying periods across channels; (3) reversible normalization significantly enhances trend forecasting by transforming non-periodic trends into periodic-like patterns; (4) increasing input horizon improves performance on multi-channel data with different periods. Code is available at: \\url{https://github.com/plumprc/RTSF}. Conclusions: Our findings provide theoretical and experimental insights into the working mechanisms of LTSF models, highlighting both the strengths and limitations of linear approaches. The results suggest that future model development should focus on handling cross-channel period variations and non-periodic components.", "authors": ["Zhe Li", "Shiyi Qi", "Yiduo Li", "Zenglin Xu"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2023-05-18", "url": "https://arxiv.org/abs/2305.10721", "pdf_url": "https://arxiv.org/pdf/2305.10721v2", "arxiv_id": "2305.10721", "doi": "10.20935/AcadAI8236", "citation_count": 4, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/plumprc/RTSF}", "venue": "Li, Zhe, Shiyi Qi, Yiduo Li, and Zenglin Xu. Revisiting Long-Term Time Series Forecasting: an Investigation on Affine Mapping. Academia AI and Applications 2, no. 2 (2026)", "quality_score": 0.1747}
{"id": "dc8922bd676a8e0c82a023a405e6576b149ea5250dcfa79088ad3448649cfb6c", "sources": ["arxiv", "semantic_scholar"], "title": "pTSE: A Multi-model Ensemble Method for Probabilistic Time Series Forecasting", "abstract": "Various probabilistic time series forecasting models have sprung up and shown remarkably good performance. However, the choice of model highly relies on the characteristics of the input time series and the fixed distribution that the model is based on. Due to the fact that the probability distributions cannot be averaged over different models straightforwardly, the current time series model ensemble methods cannot be directly applied to improve the robustness and accuracy of forecasting. To address this issue, we propose pTSE, a multi-model distribution ensemble method for probabilistic forecasting based on Hidden Markov Model (HMM). pTSE only takes off-the-shelf outputs from member models without requiring further information about each model. Besides, we provide a complete theoretical analysis of pTSE to prove that the empirical distribution of time series subject to an HMM will converge to the stationary distribution almost surely. Experiments on benchmarks show the superiority of pTSE overall member models and competitive ensemble methods.", "authors": ["Yunyi Zhou", "Zhixuan Chu", "Yijia Ruan", "Ge Jin", "Yuchen Huang", "Sheng Li"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2023-05-16", "url": "https://arxiv.org/abs/2305.11304", "pdf_url": "https://arxiv.org/pdf/2305.11304v2", "arxiv_id": "2305.11304", "doi": "10.48550/arXiv.2305.11304", "citation_count": 18, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "International Joint Conference on Artificial Intelligence", "quality_score": 0.3197}
{"id": "e8e0d457e2bfb0aba171662d3ac04320392b8e70790b6b4db90b2adc30981e26", "sources": ["arxiv", "semantic_scholar"], "title": "Differential Convolutional Fuzzy Time Series Forecasting", "abstract": "Fuzzy time series forecasting (FTSF) is a typical forecasting method with wide application. Traditional FTSF is regarded as an expert system which leads to loss of the ability to recognize undefined features. The mentioned is the main reason for poor forecasting with FTSF. To solve the problem, the proposed model Differential Fuzzy Convolutional Neural Network (DFCNN) utilizes a convolution neural network to re-implement FTSF with learnable ability. DFCNN is capable of recognizing potential information and improving forecasting accuracy. Thanks to the learnable ability of the neural network, the length of fuzzy rules established in FTSF is expended to an arbitrary length that the expert is not able to handle by the expert system. At the same time, FTSF usually cannot achieve satisfactory performance of non-stationary time series due to the trend of non-stationary time series. The trend of non-stationary time series causes the fuzzy set established by FTSF to be invalid and causes the forecasting to fail. DFCNN utilizes the Difference algorithm to weaken the non-stationary of time series so that DFCNN can forecast the non-stationary time series with a low error that FTSF cannot forecast in satisfactory performance. After the mass of experiments, DFCNN has an excellent prediction effect, which is ahead of the existing FTSF and common time series forecasting algorithms. Finally, DFCNN provides further ideas for improving FTSF and holds continued research value.", "authors": ["Tianxiang Zhan", "Yuanpeng He", "Yong Deng", "Zhen Li"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2023-05-15", "url": "https://arxiv.org/abs/2305.08890", "pdf_url": "https://arxiv.org/pdf/2305.08890v2", "arxiv_id": "2305.08890", "doi": "10.1109/TFUZZ.2023.3309811", "citation_count": 20, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "IEEE transactions on fuzzy systems", "quality_score": 0.3306}
{"id": "e04e9c65d8e9d0d6c16406cf03679c0f15a658d6a0509bc55d695e99e7b05654", "sources": ["arxiv", "semantic_scholar"], "title": "Financial Time Series Forecasting using CNN and Transformer", "abstract": "Time series forecasting is important across various domains for decision-making. In particular, financial time series such as stock prices can be hard to predict as it is difficult to model short-term and long-term temporal dependencies between data points. Convolutional Neural Networks (CNN) are good at capturing local patterns for modeling short-term dependencies. However, CNNs cannot learn long-term dependencies due to the limited receptive field. Transformers on the other hand are capable of learning global context and long-term dependencies. In this paper, we propose to harness the power of CNNs and Transformers to model both short-term and long-term dependencies within a time series, and forecast if the price would go up, down or remain the same (flat) in the future. In our experiments, we demonstrated the success of the proposed method in comparison to commonly adopted statistical and deep learning methods on forecasting intraday stock price change of S&P 500 constituents.", "authors": ["Zhen Zeng", "Rachneet Kaur", "Suchetha Siddagangappa", "Saba Rahimi", "Tucker Balch", "Manuela Veloso"], "categories": ["cs.LG", "cs.AI", "econ.EM", "q-fin.CP"], "fields_of_study": ["Computer Science", "Economics"], "published_date": "2023-04-11", "url": "https://arxiv.org/abs/2304.04912", "pdf_url": "https://arxiv.org/pdf/2304.04912v1", "arxiv_id": "2304.04912", "doi": "10.48550/arXiv.2304.04912", "citation_count": 62, "influential_citation_count": 2, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.4498}
{"id": "fa6e3f90428ff03efd9ed0a223cd32ea1a52816581bab0367a7ec3d74ed0f5e1", "sources": ["arxiv", "semantic_scholar"], "title": "Clustering Multivariate Time Series using Energy Distance", "abstract": "A novel methodology is proposed for clustering multivariate time series data using energy distance defined in Székely and Rizzo (2013). Specifically, a dissimilarity matrix is formed using the energy distance statistic to measure separation between the finite dimensional distributions for the component time series. Once the pairwise dissimilarity matrix is calculated, a hierarchical clustering method is then applied to obtain the dendrogram. This procedure is completely nonparametric as the dissimilarities between stationary distributions are directly calculated without making any model assumptions. In order to justify this procedure, asymptotic properties of the energy distance estimates are derived for general stationary and ergodic time series. The method is illustrated in a simulation study for various component time series that are either linear or nonlinear. Finally the methodology is applied to two examples; one involves GDP of selected countries and the other is population size of various states in the U.S.A. in the years 1900 -1999.", "authors": ["Richard A. Davis", "Leon Fernandes", "Konstantinos Fokianos"], "categories": ["stat.ME"], "fields_of_study": ["Mathematics"], "published_date": "2023-03-24", "url": "https://arxiv.org/abs/2303.14295", "pdf_url": "https://arxiv.org/pdf/2303.14295v1", "arxiv_id": "2303.14295", "doi": "10.1111/jtsa.12688", "citation_count": 5, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "Journal of Time Series Analysis", "quality_score": 0.1945}
{"id": "fb17a934ef4a720f21c64de84f964c9c139371989fee4cf13301c8b183d67349", "sources": ["arxiv", "semantic_scholar"], "title": "TsSHAP: Robust model agnostic feature-based explainability for time series forecasting", "abstract": "A trustworthy machine learning model should be accurate as well as explainable. Understanding why a model makes a certain decision defines the notion of explainability. While various flavors of explainability have been well-studied in supervised learning paradigms like classification and regression, literature on explainability for time series forecasting is relatively scarce. In this paper, we propose a feature-based explainability algorithm, TsSHAP, that can explain the forecast of any black-box forecasting model. The method is agnostic of the forecasting model and can provide explanations for a forecast in terms of interpretable features defined by the user a prior. The explanations are in terms of the SHAP values obtained by applying the TreeSHAP algorithm on a surrogate model that learns a mapping between the interpretable feature space and the forecast of the black-box model. Moreover, we formalize the notion of local, semi-local, and global explanations in the context of time series forecasting, which can be useful in several scenarios. We validate the efficacy and robustness of TsSHAP through extensive experiments on multiple datasets.", "authors": ["Vikas C. Raykar", "Arindam Jati", "Sumanta Mukherjee", "Nupur Aggarwal", "Kanthi Sarpatwar", "Giridhar Ganapavarapu", "Roman Vaculin"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2023-03-22", "url": "https://arxiv.org/abs/2303.12316", "pdf_url": "https://arxiv.org/pdf/2303.12316v1", "arxiv_id": "2303.12316", "doi": "10.48550/arXiv.2303.12316", "citation_count": 9, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.25}
{"id": "8514901a512a9508d1709251abdd7e9c956e71893baa73c03f2a7b56c5ad6164", "sources": ["arxiv", "semantic_scholar"], "title": "Hybrid Variational Autoencoder for Time Series Forecasting", "abstract": "Variational autoencoders (VAE) are powerful generative models that learn the latent representations of input data as random variables. Recent studies show that VAE can flexibly learn the complex temporal dynamics of time series and achieve more promising forecasting results than deterministic models. However, a major limitation of existing works is that they fail to jointly learn the local patterns (e.g., seasonality and trend) and temporal dynamics of time series for forecasting. Accordingly, we propose a novel hybrid variational autoencoder (HyVAE) to integrate the learning of local patterns and temporal dynamics by variational inference for time series forecasting. Experimental results on four real-world datasets show that the proposed HyVAE achieves better forecasting results than various counterpart methods, as well as two HyVAE variants that only learn the local patterns or temporal dynamics of time series, respectively.", "authors": ["Borui Cai", "Shuiqiao Yang", "Longxiang Gao", "Yong Xiang"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2023-03-13", "url": "https://arxiv.org/abs/2303.07048", "pdf_url": "https://arxiv.org/pdf/2303.07048v1", "arxiv_id": "2303.07048", "doi": "10.1016/j.knosys.2023.111079", "citation_count": 38, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "Knowledge-Based Systems", "quality_score": 0.3978}
{"id": "832805243c191054beaddb97ac79d45f876881cb9ae670769216d13051b6d2db", "sources": ["arxiv", "semantic_scholar"], "title": "Interpretable Water Level Forecaster with Spatiotemporal Causal Attention Mechanisms", "abstract": "Accurate forecasting of river water levels is vital for effectively managing traffic flow and mitigating the risks associated with natural disasters. This task presents challenges due to the intricate factors influencing the flow of a river. Recent advances in machine learning have introduced numerous effective forecasting methods. However, these methods lack interpretability due to their complex structure, resulting in limited reliability. Addressing this issue, this study proposes a deep learning model that quantifies interpretability, with an emphasis on water level forecasting. This model focuses on generating quantitative interpretability measurements, which align with the common knowledge embedded in the input data. This is facilitated by the utilization of a transformer architecture that is purposefully designed with masking, incorporating a multi-layer network that captures spatiotemporal causation. We perform a comparative analysis on the Han River dataset obtained from Seoul, South Korea, from 2016 to 2021. The results illustrate that our approach offers enhanced interpretability consistent with common knowledge, outperforming competing methods and also enhances robustness against distribution shift.", "authors": ["Sungchul Hong", "Yunjin Choi", "Jong-June Jeon"], "categories": ["cs.LG", "stat.ME"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2023-02-28", "url": "https://arxiv.org/abs/2303.00515", "pdf_url": "https://arxiv.org/pdf/2303.00515v8", "arxiv_id": "2303.00515", "doi": "10.1016/j.ijforecast.2024.10.003", "citation_count": 10, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "International Journal of Forecasting", "quality_score": 0.2603}
{"id": "e7fa7b42c339fc2b754d8a1137e504bc043593cd3b0bcc3db9ad5ad24a9c3a4c", "sources": ["arxiv", "semantic_scholar"], "title": "A comparative assessment of deep learning models for day-ahead load forecasting: Investigating key accuracy drivers", "abstract": "Short-term load forecasting (STLF) is vital for the effective and economic operation of power grids and energy markets. However, the non-linearity and non-stationarity of electricity demand as well as its dependency on various external factors renders STLF a challenging task. To that end, several deep learning models have been proposed in the literature for STLF, reporting promising results. In order to evaluate the accuracy of said models in day-ahead forecasting settings, in this paper we focus on the national net aggregated STLF of Portugal and conduct a comparative study considering a set of indicative, well-established deep autoregressive models, namely multi-layer perceptrons (MLP), long short-term memory networks (LSTM), neural basis expansion coefficient analysis (N-BEATS), temporal convolutional networks (TCN), and temporal fusion transformers (TFT). Moreover, we identify factors that significantly affect the demand and investigate their impact on the accuracy of each model. Our results suggest that N-BEATS consistently outperforms the rest of the examined models. MLP follows, providing further evidence towards the use of feed-forward networks over relatively more sophisticated architectures. Finally, certain calendar and weather features like the hour of the day and the temperature are identified as key accuracy drivers, providing insights regarding the forecasting approach that should be used per case.", "authors": ["Sotiris Pelekis", "Ioannis-Konstantinos Seisopoulos", "Evangelos Spiliotis", "Theodosios Pountridis", "Evangelos Karakolis", "Spiros Mouzakitis", "Dimitris Askounis"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2023-02-23", "url": "https://arxiv.org/abs/2302.12168", "pdf_url": "https://arxiv.org/pdf/2302.12168v2", "arxiv_id": "2302.12168", "doi": "10.1016/j.segan.2023.101171", "citation_count": 32, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "Sustainable Energy, Grids and Networks", "quality_score": 0.3796}
{"id": "78ba99ee7cd181cc5885608eebcfa5a8471d703b556ce673013b5d4c49452af3", "sources": ["arxiv", "semantic_scholar"], "title": "Creating Probabilistic Forecasts from Arbitrary Deterministic Forecasts using Conditional Invertible Neural Networks", "abstract": "In various applications, probabilistic forecasts are required to quantify the inherent uncertainty associated with the forecast. However, numerous modern forecasting methods are still designed to create deterministic forecasts. Transforming these deterministic forecasts into probabilistic forecasts is often challenging and based on numerous assumptions that may not hold in real-world situations. Therefore, the present article proposes a novel approach for creating probabilistic forecasts from arbitrary deterministic forecasts. In order to implement this approach, we use a conditional Invertible Neural Network (cINN). More specifically, we apply a cINN to learn the underlying distribution of the data and then combine the uncertainty from this distribution with an arbitrary deterministic forecast to generate accurate probabilistic forecasts. Our approach enables the simple creation of probabilistic forecasts without complicated statistical loss functions or further assumptions. Besides showing the mathematical validity of our approach, we empirically show that our approach noticeably outperforms traditional methods for including uncertainty in deterministic forecasts and generally outperforms state-of-the-art probabilistic forecasting benchmarks.", "authors": ["Kaleb Phipps", "Benedikt Heidrich", "Marian Turowski", "Moritz Wittig", "Ralf Mikut", "Veit Hagenmeyer"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2023-02-03", "url": "https://arxiv.org/abs/2302.01800", "pdf_url": "https://arxiv.org/pdf/2302.01800v1", "arxiv_id": "2302.01800", "doi": "10.48550/arXiv.2302.01800", "citation_count": 1, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.0753}
{"id": "26be3acb351dc33531f7f45757fd42001094c0e86461ffa2909df731d309b270", "sources": ["arxiv", "semantic_scholar"], "title": "Towards Better Long-range Time Series Forecasting using Generative Forecasting", "abstract": "Long-range time series forecasting is usually based on one of two existing forecasting strategies: Direct Forecasting and Iterative Forecasting, where the former provides low bias, high variance forecasts and the latter leads to low variance, high bias forecasts. In this paper, we propose a new forecasting strategy called Generative Forecasting (GenF), which generates synthetic data for the next few time steps and then makes long-range forecasts based on generated and observed data. We theoretically prove that GenF is able to better balance the forecasting variance and bias, leading to a much smaller forecasting error. We implement GenF via three components: (i) a novel conditional Wasserstein Generative Adversarial Network (GAN) based generator for synthetic time series data generation, called CWGAN-TS. (ii) a transformer based predictor, which makes long-range predictions using both generated and observed data. (iii) an information theoretic clustering algorithm to improve the training of both the CWGAN-TS and the transformer based predictor. The experimental results on five public datasets demonstrate that GenF significantly outperforms a diverse range of state-of-the-art benchmarks and classical approaches. Specifically, we find a 5% - 11% improvement in predictive performance (mean absolute error) while having a 15% - 50% reduction in parameters compared to the benchmarks. Lastly, we conduct an ablation study to further explore and demonstrate the effectiveness of the components comprising GenF.", "authors": ["Shiyu Liu", "Rohan Ghosh", "Mehul Motani"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2022-12-09", "url": "https://arxiv.org/abs/2212.06142", "pdf_url": "https://arxiv.org/pdf/2212.06142v1", "arxiv_id": "2212.06142", "doi": "10.48550/arXiv.2212.06142", "citation_count": 4, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1747}
{"id": "574b61df547a697d6a6c104d4d641a1cec03097c99b19aac439b96a4c0c161c7", "sources": ["arxiv", "semantic_scholar"], "title": "A K-variate Time Series Is Worth K Words: Evolution of the Vanilla Transformer Architecture for Long-term Multivariate Time Series Forecasting", "abstract": "Multivariate time series forecasting (MTSF) is a fundamental problem in numerous real-world applications. Recently, Transformer has become the de facto solution for MTSF, especially for the long-term cases. However, except for the one forward operation, the basic configurations in existing MTSF Transformer architectures were barely carefully verified. In this study, we point out that the current tokenization strategy in MTSF Transformer architectures ignores the token uniformity inductive bias of Transformers. Therefore, the vanilla MTSF transformer struggles to capture details in time series and presents inferior performance. Based on this observation, we make a series of evolution on the basic architecture of the vanilla MTSF transformer. We vary the flawed tokenization strategy, along with the decoder structure and embeddings. Surprisingly, the evolved simple transformer architecture is highly effective, which successfully avoids the over-smoothing phenomena in the vanilla MTSF transformer, achieves a more detailed and accurate prediction, and even substantially outperforms the state-of-the-art Transformers that are well-designed for MTSF.", "authors": ["Zanwei Zhou", "Ruizhe Zhong", "Chen Yang", "Yan Wang", "Xiaokang Yang", "Wei Shen"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2022-12-06", "url": "https://arxiv.org/abs/2212.02789", "pdf_url": "https://arxiv.org/pdf/2212.02789v1", "arxiv_id": "2212.02789", "doi": "10.48550/arXiv.2212.02789", "citation_count": 10, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.2603}
{"id": "f02d5d0065bf89f4633140d2c8e46b2fb9ef32c0e2e0b8c7d959e8e43fe0ac69", "sources": ["arxiv", "semantic_scholar"], "title": "Hierarchical Proxy Modeling for Improved HPO in Time Series Forecasting", "abstract": "Selecting the right set of hyperparameters is crucial in time series forecasting. The classical temporal cross-validation framework for hyperparameter optimization (HPO) often leads to poor test performance because of a possible mismatch between validation and test periods. To address this test-validation mismatch, we propose a novel technique, H-Pro to drive HPO via test proxies by exploiting data hierarchies often associated with time series datasets. Since higher-level aggregated time series often show less irregularity and better predictability as compared to the lowest-level time series which can be sparse and intermittent, we optimize the hyperparameters of the lowest-level base-forecaster by leveraging the proxy forecasts for the test period generated from the forecasters at higher levels. H-Pro can be applied on any off-the-shelf machine learning model to perform HPO. We validate the efficacy of our technique with extensive empirical evaluation on five publicly available hierarchical forecasting datasets. Our approach outperforms existing state-of-the-art methods in Tourism, Wiki, and Traffic datasets, and achieves competitive result in Tourism-L dataset, without any model-specific enhancements. Moreover, our method outperforms the winning method of the M5 forecast accuracy competition.", "authors": ["Arindam Jati", "Vijay Ekambaram", "Shaonli Pal", "Brian Quanz", "Wesley M. Gifford", "Pavithra Harsha", "Stuart Siegel", "Sumanta Mukherjee", "Chandra Narayanaswami"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2022-11-28", "url": "https://arxiv.org/abs/2211.15092", "pdf_url": "https://arxiv.org/pdf/2211.15092v2", "arxiv_id": "2211.15092", "doi": "10.1145/3580305.3599378", "citation_count": 5, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "Knowledge Discovery and Data Mining", "quality_score": 0.1945}
{"id": "428922e48c2dd71d47b47d7dd687e20bd6c7b24ac2f40fa5dad123a0d0a97a3a", "sources": ["arxiv", "semantic_scholar"], "title": "A Time Series is Worth 64 Words: Long-term Forecasting with Transformers", "abstract": "We propose an efficient design of Transformer-based models for multivariate time series forecasting and self-supervised representation learning. It is based on two key components: (i) segmentation of time series into subseries-level patches which are served as input tokens to Transformer; (ii) channel-independence where each channel contains a single univariate time series that shares the same embedding and Transformer weights across all the series. Patching design naturally has three-fold benefit: local semantic information is retained in the embedding; computation and memory usage of the attention maps are quadratically reduced given the same look-back window; and the model can attend longer history. Our channel-independent patch time series Transformer (PatchTST) can improve the long-term forecasting accuracy significantly when compared with that of SOTA Transformer-based models. We also apply our model to self-supervised pre-training tasks and attain excellent fine-tuning performance, which outperforms supervised training on large datasets. Transferring of masked pre-trained representation on one dataset to others also produces SOTA forecasting accuracy. Code is available at: https://github.com/yuqinie98/PatchTST.", "authors": ["Yuqi Nie", "Nam H. Nguyen", "Phanwadee Sinthong", "Jayant Kalagnanam"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2022-11-27", "url": "https://arxiv.org/abs/2211.14730", "pdf_url": "https://arxiv.org/pdf/2211.14730v2", "arxiv_id": "2211.14730", "doi": "10.48550/arXiv.2211.14730", "citation_count": 3709, "influential_citation_count": 702, "has_code": true, "code_url": "https://github.com/yuqinie98/PatchTST", "venue": "International Conference on Learning Representations", "quality_score": 1.0}
{"id": "552ac740a46e703e5ebe8b89db46390c013daf90c09c6093ab3bc8f5065604f5", "sources": ["arxiv", "semantic_scholar"], "title": "Probabilistic Decomposition Transformer for Time Series Forecasting", "abstract": "Time series forecasting is crucial for many fields, such as disaster warning, weather prediction, and energy consumption. The Transformer-based models are considered to have revolutionized the field of sequence modeling. However, the complex temporal patterns of the time series hinder the model from mining reliable temporal dependencies. Furthermore, the autoregressive form of the Transformer introduces cumulative errors in the inference step. In this paper, we propose the probabilistic decomposition Transformer model that combines the Transformer with a conditional generative model, which provides hierarchical and interpretable probabilistic forecasts for intricate time series. The Transformer is employed to learn temporal patterns and implement primary probabilistic forecasts, while the conditional generative model is used to achieve non-autoregressive hierarchical probabilistic forecasts by introducing latent space feature representations. In addition, the conditional generative model reconstructs typical features of the series, such as seasonality and trend terms, from probability distributions in the latent space to enable complex pattern separation and provide interpretable forecasts. Extensive experiments on several datasets demonstrate the effectiveness and robustness of the proposed model, indicating that it compares favorably with the state of the art.", "authors": ["Junlong Tong", "Liping Xie", "Wankou Yang", "Kanjian Zhang"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2022-10-31", "url": "https://arxiv.org/abs/2210.17393", "pdf_url": "https://arxiv.org/pdf/2210.17393v1", "arxiv_id": "2210.17393", "doi": "10.48550/arXiv.2210.17393", "citation_count": 11, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "SDM", "quality_score": 0.2698}
{"id": "89458a99f446939427c04c23e8e0d71aa230003d123642e0d54078ad7553b9eb", "sources": ["arxiv", "semantic_scholar"], "title": "Retrieval Based Time Series Forecasting", "abstract": "Time series data appears in a variety of applications such as smart transportation and environmental monitoring. One of the fundamental problems for time series analysis is time series forecasting. Despite the success of recent deep time series forecasting methods, they require sufficient observation of historical values to make accurate forecasting. In other words, the ratio of the output length (or forecasting horizon) to the sum of the input and output lengths should be low enough (e.g., 0.3). As the ratio increases (e.g., to 0.8), the uncertainty for the forecasting accuracy increases significantly. In this paper, we show both theoretically and empirically that the uncertainty could be effectively reduced by retrieving relevant time series as references. In the theoretical analysis, we first quantify the uncertainty and show its connections to the Mean Squared Error (MSE). Then we prove that models with references are easier to learn than models without references since the retrieved references could reduce the uncertainty. To empirically demonstrate the effectiveness of the retrieval based time series forecasting models, we introduce a simple yet effective two-stage method, called ReTime consisting of a relational retrieval and a content synthesis. We also show that ReTime can be easily adapted to the spatial-temporal time series and time series imputation settings. Finally, we evaluate ReTime on real-world datasets to demonstrate its effectiveness.", "authors": ["Baoyu Jing", "Si Zhang", "Yada Zhu", "Bin Peng", "Kaiyu Guan", "Andrew Margenot", "Hanghang Tong"], "categories": ["cs.AI", "cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2022-09-27", "url": "https://arxiv.org/abs/2209.13525", "pdf_url": "https://arxiv.org/pdf/2209.13525v1", "arxiv_id": "2209.13525", "doi": "10.48550/arXiv.2209.13525", "citation_count": 19, "influential_citation_count": 2, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.3253}
{"id": "2062f179768475929ecbb992af7f22f10bd78be864fc6477a4923896f772153a", "sources": ["arxiv", "semantic_scholar"], "title": "PromptCast: A New Prompt-based Learning Paradigm for Time Series Forecasting", "abstract": "This paper presents a new perspective on time series forecasting. In existing time series forecasting methods, the models take a sequence of numerical values as input and yield numerical values as output. The existing SOTA models are largely based on the Transformer architecture, modified with multiple encoding mechanisms to incorporate the context and semantics around the historical data. Inspired by the successes of pre-trained language foundation models, we pose a question about whether these models can also be adapted to solve time-series forecasting. Thus, we propose a new forecasting paradigm: prompt-based time series forecasting (PromptCast). In this novel task, the numerical input and output are transformed into prompts and the forecasting task is framed in a sentence-to-sentence manner, making it possible to directly apply language models for forecasting purposes. To support and facilitate the research of this task, we also present a large-scale dataset (PISA) that includes three real-world forecasting scenarios. We evaluate different SOTA numerical-based forecasting methods and language generation models. The benchmark results with various forecasting settings demonstrate the proposed PromptCast with language generation models is a promising research direction. Additionally, in comparison to conventional numerical-based forecasting, PromptCast shows a much better generalization ability under the zero-shot setting.", "authors": ["Hao Xue", "Flora D. Salim"], "categories": ["stat.ME", "cs.AI", "cs.CL", "cs.LG", "math.ST"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2022-09-20", "url": "https://arxiv.org/abs/2210.08964", "pdf_url": "https://arxiv.org/pdf/2210.08964v5", "arxiv_id": "2210.08964", "doi": "10.1109/TKDE.2023.3342137", "citation_count": 344, "influential_citation_count": 17, "has_code": false, "code_url": null, "venue": "IEEE Transactions on Knowledge and Data Engineering", "quality_score": 0.6345}
{"id": "c5c84d1545cebf0ccc2d6675ccc554735d3e7b4e16d1ea2db7f489658ab0b460", "sources": ["arxiv", "semantic_scholar"], "title": "W-Transformers : A Wavelet-based Transformer Framework for Univariate Time Series Forecasting", "abstract": "Deep learning utilizing transformers has recently achieved a lot of success in many vital areas such as natural language processing, computer vision, anomaly detection, and recommendation systems, among many others. Among several merits of transformers, the ability to capture long-range temporal dependencies and interactions is desirable for time series forecasting, leading to its progress in various time series applications. In this paper, we build a transformer model for non-stationary time series. The problem is challenging yet crucially important. We present a novel framework for univariate time series representation learning based on the wavelet-based transformer encoder architecture and call it W-Transformer. The proposed W-Transformers utilize a maximal overlap discrete wavelet transformation (MODWT) to the time series data and build local transformers on the decomposed datasets to vividly capture the nonstationarity and long-range nonlinear dependencies in the time series. Evaluating our framework on several publicly available benchmark time series datasets from various domains and with diverse characteristics, we demonstrate that it performs, on average, significantly better than the baseline forecasters for short-term and long-term forecasting, even for datasets that consist of only a few hundred training samples.", "authors": ["Lena Sasal", "Tanujit Chakraborty", "Abdenour Hadid"], "categories": ["cs.LG", "econ.EM", "eess.SP", "stat.ML"], "fields_of_study": ["Computer Science", "Physics", "Mathematics", "Economics", "Engineering"], "published_date": "2022-09-08", "url": "https://arxiv.org/abs/2209.03945", "pdf_url": "https://arxiv.org/pdf/2209.03945v1", "arxiv_id": "2209.03945", "doi": "10.1109/ICMLA55696.2022.00111", "citation_count": 43, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "International Conference on Machine Learning and Applications", "quality_score": 0.4109}
{"id": "7e0311792c2cc76c5e7a1169577a28d8531d12076ddcc9ace50c7b7e841eb1e5", "sources": ["arxiv", "semantic_scholar"], "title": "Persistence Initialization: A novel adaptation of the Transformer architecture for Time Series Forecasting", "abstract": "Time series forecasting is an important problem, with many real world applications. Ensembles of deep neural networks have recently achieved impressive forecasting accuracy, but such large ensembles are impractical in many real world settings. Transformer models been successfully applied to a diverse set of challenging problems. We propose a novel adaptation of the original Transformer architecture focusing on the task of time series forecasting, called Persistence Initialization. The model is initialized as a naive persistence model by using a multiplicative gating mechanism combined with a residual skip connection. We use a decoder Transformer with ReZero normalization and Rotary positional encodings, but the adaptation is applicable to any auto-regressive neural network model. We evaluate our proposed architecture on the challenging M4 dataset, achieving competitive performance compared to ensemble based methods. We also compare against existing recently proposed Transformer models for time series forecasting, showing superior performance on the M4 dataset. Extensive ablation studies show that Persistence Initialization leads to better performance and faster convergence. As the size of the model increases, only the models with our proposed adaptation gain in performance. We also perform an additional ablation study to determine the importance of the choice of normalization and positional encoding, and find both the use of Rotary encodings and ReZero normalization to be essential for good forecasting performance.", "authors": ["Espen Haugsdal", "Erlend Aune", "Massimiliano Ruocco"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2022-08-30", "url": "https://arxiv.org/abs/2208.14236", "pdf_url": "https://arxiv.org/pdf/2208.14236v1", "arxiv_id": "2208.14236", "doi": "10.1007/s10489-023-04927-4", "citation_count": 30, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.3728}
{"id": "1c69651f5be0b47aa9cde1e026ce92feec9fe3610f7cad667dafa76e6e17c495", "sources": ["arxiv", "semantic_scholar"], "title": "Identifying and Overcoming Transformation Bias in Forecasting Models", "abstract": "Log and square root transformations of target variable are routinely used in forecasting models to predict future sales. These transformations often lead to better performing models. However, they also introduce a systematic negative bias (under-forecasting). In this paper, we demonstrate the existence of this bias, dive deep into its root cause and introduce two methods to correct for the bias. We conclude that the proposed bias correction methods improve model performance (by up to 50%) and make a case for incorporating bias correction in modeling workflow. We also experiment with `Tweedie' family of cost functions which circumvents the transformation bias issue by modeling directly on sales. We conclude that Tweedie regression gives the best performance so far when modeling on sales making it a strong alternative to working with a transformed target variable.", "authors": ["Sushant More"], "categories": ["cs.LG", "stat.AP"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2022-08-24", "url": "https://arxiv.org/abs/2208.12264", "pdf_url": "https://arxiv.org/pdf/2208.12264v1", "arxiv_id": "2208.12264", "doi": "10.48550/arXiv.2208.12264", "citation_count": 3, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1505}
{"id": "fb63413cd40b2591678843468dbc6a883c046794e9f167c5e071b90a3e3e794e", "sources": ["arxiv", "semantic_scholar"], "title": "Expressing Multivariate Time Series as Graphs with Time Series Attention Transformer", "abstract": "A reliable and efficient representation of multivariate time series is crucial in various downstream machine learning tasks. In multivariate time series forecasting, each variable depends on its historical values and there are inter-dependencies among variables as well. Models have to be designed to capture both intra- and inter-relationships among the time series. To move towards this goal, we propose the Time Series Attention Transformer (TSAT) for multivariate time series representation learning. Using TSAT, we represent both temporal information and inter-dependencies of multivariate time series in terms of edge-enhanced dynamic graphs. The intra-series correlations are represented by nodes in a dynamic graph; a self-attention mechanism is modified to capture the inter-series correlations by using the super-empirical mode decomposition (SMD) module. We applied the embedded dynamic graphs to times series forecasting problems, including two real-world datasets and two benchmark datasets. Extensive experiments show that TSAT clearly outerperforms six state-of-the-art baseline methods in various forecasting horizons. We further visualize the embedded dynamic graphs to illustrate the graph representation power of TSAT. We share our code at https://github.com/RadiantResearch/TSAT.", "authors": ["William T. Ng", "K. Siu", "Albert C. Cheung", "Michael K. Ng"], "categories": ["cs.LG", "cs.AI", "math.DS", "math.RT"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2022-08-19", "url": "https://arxiv.org/abs/2208.09300", "pdf_url": "https://arxiv.org/pdf/2208.09300v1", "arxiv_id": "2208.09300", "doi": "10.48550/arXiv.2208.09300", "citation_count": 14, "influential_citation_count": 1, "has_code": true, "code_url": "https://github.com/RadiantResearch/TSAT", "venue": "arXiv.org", "quality_score": 0.294}
{"id": "fb0543f13f55eaa408964f8f44df41d00f69044e9e04ed7e3774f75dfca07162", "sources": ["arxiv", "semantic_scholar"], "title": "SDWPF: A Dataset for Spatial Dynamic Wind Power Forecasting Challenge at KDD Cup 2022", "abstract": "The variability of wind power supply can present substantial challenges to incorporating wind power into a grid system. Thus, Wind Power Forecasting (WPF) has been widely recognized as one of the most critical issues in wind power integration and operation. There has been an explosion of studies on wind power forecasting problems in the past decades. Nevertheless, how to well handle the WPF problem is still challenging, since high prediction accuracy is always demanded to ensure grid stability and security of supply. We present a unique Spatial Dynamic Wind Power Forecasting dataset: SDWPF, which includes the spatial distribution of wind turbines, as well as the dynamic context factors. Whereas, most of the existing datasets have only a small number of wind turbines without knowing the locations and context information of wind turbines at a fine-grained time scale. By contrast, SDWPF provides the wind power data of 134 wind turbines from a wind farm over half a year with their relative positions and internal statuses. We use this dataset to launch the Baidu KDD Cup 2022 to examine the limit of current WPF solutions. The dataset is released at https://aistudio.baidu.com/aistudio/competition/detail/152/0/datasets.", "authors": ["Jingbo Zhou", "Xinjiang Lu", "Yixiong Xiao", "Jiantao Su", "Junfu Lyu", "Yanjun Ma", "Dejing Dou"], "categories": ["cs.LG", "eess.SP"], "fields_of_study": ["Computer Science", "Engineering", "Medicine"], "published_date": "2022-08-08", "url": "https://arxiv.org/abs/2208.04360", "pdf_url": "https://arxiv.org/pdf/2208.04360v2", "arxiv_id": "2208.04360", "doi": "10.1038/s41597-024-03427-5", "citation_count": 69, "influential_citation_count": 5, "has_code": false, "code_url": null, "venue": "Scientific Data", "quality_score": 0.4613}
{"id": "1fbb06951e330c5ba74b4aaddf58821924e51d05af6468e17f2daa94053cc7b2", "sources": ["arxiv", "semantic_scholar"], "title": "LETS-GZSL: A Latent Embedding Model for Time Series Generalized Zero Shot Learning", "abstract": "One of the recent developments in deep learning is generalized zero-shot learning (GZSL), which aims to recognize objects from both seen and unseen classes, when only the labeled examples from seen classes are provided. Over the past couple of years, GZSL has picked up traction and several models have been proposed to solve this problem. Whereas an extensive amount of research on GZSL has been carried out in fields such as computer vision and natural language processing, no such research has been carried out to deal with time series data. GZSL is used for applications such as detecting abnormalities from ECG and EEG data and identifying unseen classes from sensor, spectrograph and other devices' data. In this regard, we propose a Latent Embedding for Time Series - GZSL (LETS-GZSL) model that can solve the problem of GZSL for time series classification (TSC). We utilize an embedding-based approach and combine it with attribute vectors to predict the final class labels. We report our results on the widely popular UCR archive datasets. Our framework is able to achieve a harmonic mean value of at least 55% on most of the datasets except when the number of unseen classes is greater than 3 or the amount of data is very low (less than 100 training examples).", "authors": ["Sathvik Bhaskarpandit", "Priyanka Gupta", "Manik Gupta"], "categories": ["cs.AI", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2022-07-25", "url": "https://arxiv.org/abs/2207.12007", "pdf_url": "https://arxiv.org/pdf/2207.12007v1", "arxiv_id": "2207.12007", "doi": "10.48550/arXiv.2207.12007", "citation_count": 2, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1193}
{"id": "c4075f00f03e59d9722cff5b7216f92fea7d97f38c4fa98734f1f96473e32a6e", "sources": ["arxiv", "semantic_scholar"], "title": "Respecting Time Series Properties Makes Deep Time Series Forecasting Perfect", "abstract": "How to handle time features shall be the core question of any time series forecasting model. Ironically, it is often ignored or misunderstood by deep-learning based models, even those baselines which are state-of-the-art. This behavior makes their inefficient, untenable and unstable. In this paper, we rigorously analyze three prevalent but deficient/unfounded deep time series forecasting mechanisms or methods from the view of time series properties, including normalization methods, multivariate forecasting and input sequence length. Corresponding corollaries and solutions are given on both empirical and theoretical basis. We thereby propose a novel time series forecasting network, i.e. RTNet, on the basis of aforementioned analysis. It is general enough to be combined with both supervised and self-supervised forecasting format. Thanks to the core idea of respecting time series properties, no matter in which forecasting format, RTNet shows obviously superior forecasting performances compared with dozens of other SOTA time series forecasting baselines in three real-world benchmark datasets. By and large, it even occupies less time complexity and memory usage while acquiring better forecasting accuracy. The source code is available at https://github.com/OrigamiSL/RTNet.", "authors": ["Li Shen", "Yuning Wei", "Yangzhu Wang"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2022-07-22", "url": "https://arxiv.org/abs/2207.10941", "pdf_url": "https://arxiv.org/pdf/2207.10941v1", "arxiv_id": "2207.10941", "doi": "10.48550/arXiv.2207.10941", "citation_count": 8, "influential_citation_count": 0, "has_code": true, "code_url": "https://github.com/OrigamiSL/RTNet", "venue": "arXiv.org", "quality_score": 0.2386}
{"id": "2db2077b3a143418b41a18442b4549ade04ef23cb6a834fbc41c2a13e2bbcffa", "sources": ["arxiv", "semantic_scholar"], "title": "MQRetNN: Multi-Horizon Time Series Forecasting with Retrieval Augmentation", "abstract": "Multi-horizon probabilistic time series forecasting has wide applicability to real-world tasks such as demand forecasting. Recent work in neural time-series forecasting mainly focus on the use of Seq2Seq architectures. For example, MQTransformer - an improvement of MQCNN - has shown the state-of-the-art performance in probabilistic demand forecasting. In this paper, we consider incorporating cross-entity information to enhance model performance by adding a cross-entity attention mechanism along with a retrieval mechanism to select which entities to attend over. We demonstrate how our new neural architecture, MQRetNN, leverages the encoded contexts from a pretrained baseline model on the entire population to improve forecasting accuracy. Using MQCNN as the baseline model (due to computational constraints, we do not use MQTransformer), we first show on a small demand forecasting dataset that it is possible to achieve ~3% improvement in test loss by adding a cross-entity attention mechanism where each entity attends to all others in the population. We then evaluate the model with our proposed retrieval methods - as a means of approximating an attention over a large population - on a large-scale demand forecasting application with over 2 million products and observe ~1% performance gain over the MQCNN baseline.", "authors": ["Sitan Yang", "Carson Eisenach", "Dhruv Madeka"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2022-07-21", "url": "https://arxiv.org/abs/2207.10517", "pdf_url": "https://arxiv.org/pdf/2207.10517v2", "arxiv_id": "2207.10517", "doi": "10.48550/arXiv.2207.10517", "citation_count": 11, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.2698}
{"id": "d95348abaefef939caa6838c6b75efa51e04eca79080dc25f5951f7e9d97c0ff", "sources": ["arxiv", "semantic_scholar"], "title": "Learning Deep Time-index Models for Time Series Forecasting", "abstract": "Deep learning has been actively applied to time series forecasting, leading to a deluge of new methods, belonging to the class of historical-value models. Yet, despite the attractive properties of time-index models, such as being able to model the continuous nature of underlying time series dynamics, little attention has been given to them. Indeed, while naive deep time-index models are far more expressive than the manually predefined function representations of classical time-index models, they are inadequate for forecasting, being unable to generalize to unseen time steps due to the lack of inductive bias. In this paper, we propose DeepTime, a meta-optimization framework to learn deep time-index models which overcome these limitations, yielding an efficient and accurate forecasting model. Extensive experiments on real world datasets in the long sequence time-series forecasting setting demonstrate that our approach achieves competitive results with state-of-the-art methods, and is highly efficient. Code is available at https://github.com/salesforce/DeepTime.", "authors": ["Gerald Woo", "Chenghao Liu", "Doyen Sahoo", "Akshat Kumar", "Steven Hoi"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2022-07-13", "url": "https://arxiv.org/abs/2207.06046", "pdf_url": "https://arxiv.org/pdf/2207.06046v4", "arxiv_id": "2207.06046", "doi": null, "citation_count": 39, "influential_citation_count": 3, "has_code": true, "code_url": "https://github.com/salesforce/DeepTime", "venue": "International Conference on Machine Learning", "quality_score": 0.4005}
{"id": "de0600621139f75dad2718defdf2e77cb3b473c22c758d3d6e716134c78c7d18", "sources": ["arxiv", "semantic_scholar"], "title": "Dateformer: Time-modeling Transformer for Longer-term Series Forecasting", "abstract": "Transformers have demonstrated impressive strength in long-term series forecasting. Existing prediction research mostly focused on mapping past short sub-series (lookback window) to future series (forecast window). The longer training dataset time series will be discarded, once training is completed. Models can merely rely on lookback window information for inference, which impedes models from analyzing time series from a global perspective. And these windows used by Transformers are quite narrow because they must model each time-step therein. Under this point-wise processing style, broadening windows will rapidly exhaust their model capacity. This, for fine-grained time series, leads to a bottleneck in information input and prediction output, which is mortal to long-term series forecasting. To overcome the barrier, we propose a brand-new methodology to utilize Transformer for time series forecasting. Specifically, we split time series into patches by day and reform point-wise to patch-wise processing, which considerably enhances the information input and output of Transformers. To further help models leverage the whole training set's global information during inference, we distill the information, store it in time representations, and replace series with time representations as the main modeling entities. Our designed time-modeling Transformer -- Dateformer yields state-of-the-art accuracy on 7 real-world datasets with a 33.6\\% relative improvement and extends the maximum forecast range to half-year.", "authors": ["Julong Young", "Junhui Chen", "Feihu Huang", "Jian Peng"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2022-07-12", "url": "https://arxiv.org/abs/2207.05397", "pdf_url": "https://arxiv.org/pdf/2207.05397v2", "arxiv_id": "2207.05397", "doi": null, "citation_count": 6, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.2113}
{"id": "4f8e9d737fe95ea493b272581a43586b858b1cb1905501bfa94699564ac95193", "sources": ["arxiv", "semantic_scholar"], "title": "LASSO Principal Component Averaging -- a fully automated approach for point forecast pooling", "abstract": "This paper develops a novel, fully automated forecast averaging scheme, which combines LASSO estimation method with Principal Component Averaging (PCA). LASSO-PCA (LPCA) explores a pool of predictions based on a single model but calibrated to windows of different sizes. It uses information criteria to select tuning parameters and hence reduces the impact of researchers' at hock decisions. The method is applied to average predictions of hourly day-ahead electricity prices over 650 point forecasts obtained with various lengths of calibration windows. It is evaluated on four European and American markets with almost two and a half year of out-of-sample period and compared to other semi- and fully automated methods, such as simple mean, AW/WAW, LASSO and PCA. The results indicate that the LASSO averaging is very efficient in terms of forecast error reduction, whereas PCA method is robust to the selection of the specification parameter. LPCA inherits the advantages of both methods and outperforms other approaches in terms of MAE, remaining insensitive the the choice of a tuning parameter.", "authors": ["Bartosz Uniejewski", "Katarzyna Maciejowska"], "categories": ["stat.AP", "q-fin.RM", "q-fin.ST"], "fields_of_study": ["Mathematics", "Economics"], "published_date": "2022-07-11", "url": "https://arxiv.org/abs/2207.04794", "pdf_url": "https://arxiv.org/pdf/2207.04794v1", "arxiv_id": "2207.04794", "doi": "10.1016/j.ijforecast.2022.09.004", "citation_count": 15, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "International Journal of Forecasting", "quality_score": 0.301}
{"id": "88c45fbedc6872a35333570100c4b05f8661ca73fb66d034e287118ae67cb4b1", "sources": ["arxiv", "semantic_scholar"], "title": "DRAformer: Differentially Reconstructed Attention Transformer for Time-Series Forecasting", "abstract": "Time-series forecasting plays an important role in many real-world scenarios, such as equipment life cycle forecasting, weather forecasting, and traffic flow forecasting. It can be observed from recent research that a variety of transformer-based models have shown remarkable results in time-series forecasting. However, there are still some issues that limit the ability of transformer-based models on time-series forecasting tasks: (i) learning directly on raw data is susceptible to noise due to its complex and unstable feature representation; (ii) the self-attention mechanisms pay insufficient attention to changing features and temporal dependencies. In order to solve these two problems, we propose a transformer-based differentially reconstructed attention model DRAformer. Specifically, DRAformer has the following innovations: (i) learning against differenced sequences, which preserves clear and stable sequence features by differencing and highlights the changing properties of sequences; (ii) the reconstructed attention: integrated distance attention exhibits sequential distance through a learnable Gaussian kernel, distributed difference attention calculates distribution difference by mapping the difference sequence to the adaptive feature space, and the combination of the two effectively focuses on the sequences with prominent associations; (iii) the reconstructed decoder input, which extracts sequence features by integrating variation information and temporal correlations, thereby obtaining a more comprehensive sequence representation. Extensive experiments on four large-scale datasets demonstrate that DRAformer outperforms state-of-the-art baselines.", "authors": ["Benhan Li", "Shengdong Du", "Tianrui Li", "Jie Hu", "Zhen Jia"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2022-06-11", "url": "https://arxiv.org/abs/2206.05495", "pdf_url": "https://arxiv.org/pdf/2206.05495v1", "arxiv_id": "2206.05495", "doi": "10.48550/arXiv.2206.05495", "citation_count": 2, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1193}
{"id": "70da371d78dbfc48011b0b41e4f323198e006b538b8e44ecc97753521d7a379c", "sources": ["arxiv", "semantic_scholar"], "title": "Scaleformer: Iterative Multi-scale Refining Transformers for Time Series Forecasting", "abstract": "The performance of time series forecasting has recently been greatly improved by the introduction of transformers. In this paper, we propose a general multi-scale framework that can be applied to the state-of-the-art transformer-based time series forecasting models (FEDformer, Autoformer, etc.). By iteratively refining a forecasted time series at multiple scales with shared weights, introducing architecture adaptations, and a specially-designed normalization scheme, we are able to achieve significant performance improvements, from 5.5% to 38.5% across datasets and transformer architectures, with minimal additional computational overhead. Via detailed ablation studies, we demonstrate the effectiveness of each of our contributions across the architecture and methodology. Furthermore, our experiments on various public datasets demonstrate that the proposed improvements outperform their corresponding baseline counterparts. Our code is publicly available in https://github.com/BorealisAI/scaleformer.", "authors": ["Amin Shabani", "Amir Abdi", "Lili Meng", "Tristan Sylvain"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2022-06-08", "url": "https://arxiv.org/abs/2206.04038", "pdf_url": "https://arxiv.org/pdf/2206.04038v4", "arxiv_id": "2206.04038", "doi": "10.48550/arXiv.2206.04038", "citation_count": 114, "influential_citation_count": 4, "has_code": true, "code_url": "https://github.com/BorealisAI/scaleformer", "venue": "International Conference on Learning Representations", "quality_score": 0.5152}
{"id": "7f15039e3892bd5846c428664ae5ccf73dee571b962bab39c97f8d456523636c", "sources": ["arxiv", "semantic_scholar"], "title": "Time series aggregation for optimization: One-size-fits-all?", "abstract": "One of the fundamental problems of using optimization models that use different time series as data input, is the trade-off between model accuracy and computational tractability. To overcome the computational intractability of these full optimization models, the dimension of input data and model size is commonly reduced through time series aggregation (TSA) methods. However, traditional TSA methods often apply a one-size-fits-all approach based on the common belief that the clusters that best approximate the input data also lead to the aggregated model that best approximates the full model, while the metric that really matters - the resulting output error in optimization results - is not well addressed. In this paper, we plan to challenge this belief and show that output-error based TSA methods with theoretical underpinnings have unprecedented potential of computational efficiency and accuracy.", "authors": ["Sonja Wogrin"], "categories": ["math.OC"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2022-06-07", "url": "https://arxiv.org/abs/2206.03186", "pdf_url": "https://arxiv.org/pdf/2206.03186v1", "arxiv_id": "2206.03186", "doi": "10.1109/TSG.2023.3242467", "citation_count": 19, "influential_citation_count": 4, "has_code": false, "code_url": null, "venue": "IEEE Transactions on Smart Grid", "quality_score": 0.3495}
{"id": "7eef1ea60ae236fcf6baeb8cee34263e902b271b4f90a7ef270fc569c3ba4b75", "sources": ["arxiv", "semantic_scholar"], "title": "Non-stationary Transformers: Exploring the Stationarity in Time Series Forecasting", "abstract": "Transformers have shown great power in time series forecasting due to their global-range modeling ability. However, their performance can degenerate terribly on non-stationary real-world data in which the joint distribution changes over time. Previous studies primarily adopt stationarization to attenuate the non-stationarity of original series for better predictability. But the stationarized series deprived of inherent non-stationarity can be less instructive for real-world bursty events forecasting. This problem, termed over-stationarization in this paper, leads Transformers to generate indistinguishable temporal attentions for different series and impedes the predictive capability of deep models. To tackle the dilemma between series predictability and model capability, we propose Non-stationary Transformers as a generic framework with two interdependent modules: Series Stationarization and De-stationary Attention. Concretely, Series Stationarization unifies the statistics of each input and converts the output with restored statistics for better predictability. To address the over-stationarization problem, De-stationary Attention is devised to recover the intrinsic non-stationary information into temporal dependencies by approximating distinguishable attentions learned from raw series. Our Non-stationary Transformers framework consistently boosts mainstream Transformers by a large margin, which reduces MSE by 49.43% on Transformer, 47.34% on Informer, and 46.89% on Reformer, making them the state-of-the-art in time series forecasting. Code is available at this repository: https://github.com/thuml/Nonstationary_Transformers.", "authors": ["Yong Liu", "Haixu Wu", "Jianmin Wang", "Mingsheng Long"], "categories": ["cs.LG", "eess.SP"], "fields_of_study": ["Computer Science", "Engineering"], "published_date": "2022-05-28", "url": "https://arxiv.org/abs/2205.14415", "pdf_url": "https://arxiv.org/pdf/2205.14415v4", "arxiv_id": "2205.14415", "doi": "10.52202/068431-0718", "citation_count": 882, "influential_citation_count": 78, "has_code": true, "code_url": "https://github.com/thuml/Nonstationary_Transformers", "venue": "Neural Information Processing Systems", "quality_score": 0.9488}
{"id": "4143f5b84b251c7011f4c95164a2c0702a8acb024b120c2db5003394af7b5029", "sources": ["arxiv", "semantic_scholar"], "title": "Are Transformers Effective for Time Series Forecasting?", "abstract": "Recently, there has been a surge of Transformer-based solutions for the long-term time series forecasting (LTSF) task. Despite the growing performance over the past few years, we question the validity of this line of research in this work. Specifically, Transformers is arguably the most successful solution to extract the semantic correlations among the elements in a long sequence. However, in time series modeling, we are to extract the temporal relations in an ordered set of continuous points. While employing positional encoding and using tokens to embed sub-series in Transformers facilitate preserving some ordering information, the nature of the \\emph{permutation-invariant} self-attention mechanism inevitably results in temporal information loss. To validate our claim, we introduce a set of embarrassingly simple one-layer linear models named LTSF-Linear for comparison. Experimental results on nine real-life datasets show that LTSF-Linear surprisingly outperforms existing sophisticated Transformer-based LTSF models in all cases, and often by a large margin. Moreover, we conduct comprehensive empirical studies to explore the impacts of various design elements of LTSF models on their temporal relation extraction capability. We hope this surprising finding opens up new research directions for the LTSF task. We also advocate revisiting the validity of Transformer-based solutions for other time series analysis tasks (e.g., anomaly detection) in the future. Code is available at: \\url{https://github.com/cure-lab/LTSF-Linear}.", "authors": ["Ailing Zeng", "Muxi Chen", "Lei Zhang", "Qiang Xu"], "categories": ["cs.AI", "cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2022-05-26", "url": "https://arxiv.org/abs/2205.13504", "pdf_url": "https://arxiv.org/pdf/2205.13504v3", "arxiv_id": "2205.13504", "doi": "10.48550/arXiv.2205.13504", "citation_count": 3906, "influential_citation_count": 433, "has_code": true, "code_url": "https://github.com/cure-lab/LTSF-Linear", "venue": "AAAI Conference on Artificial Intelligence", "quality_score": 1.0}
{"id": "baa2464cd55fb43a4dfed3800c5d83157a055d04c30dd2b952fd26e9acfea15f", "sources": ["arxiv", "semantic_scholar"], "title": "Wind energy forecasting with missing values within a fully conditional specification framework", "abstract": "Wind power forecasting is essential to power system operation and electricity markets. As abundant data became available thanks to the deployment of measurement infrastructures and the democratization of meteorological modelling, extensive data-driven approaches have been developed within both point and probabilistic forecasting frameworks. These models usually assume that the dataset at hand is complete and overlook missing value issues that often occur in practice. In contrast to that common approach, we rigorously consider here the wind power forecasting problem in the presence of missing values, by jointly accommodating imputation and forecasting tasks. Our approach allows inferring the joint distribution of input features and target variables at the model estimation stage based on incomplete observations only. We place emphasis on a fully conditional specification method owing to its desirable properties, e.g., being assumption-free when it comes to these joint distributions. Then, at the operational forecasting stage, with available features at hand, one can issue forecasts by implicitly imputing all missing entries. The approach is applicable to both point and probabilistic forecasting, while yielding competitive forecast quality within both simulation and real-world case studies. It confirms that by using a powerful universal imputation method like fully conditional specification, the proposed approach is superior to the common approach, especially in the context of probabilistic forecasting.", "authors": ["Honglin Wen", "Pierre Pinson", "Jie Gu", "Zhijian Jin"], "categories": ["stat.AP", "eess.SY"], "fields_of_study": ["Mathematics", "Computer Science", "Engineering"], "published_date": "2022-03-15", "url": "https://arxiv.org/abs/2203.08252", "pdf_url": "https://arxiv.org/pdf/2203.08252v2", "arxiv_id": "2203.08252", "doi": "10.1016/j.ijforecast.2022.12.006", "citation_count": 16, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "International Journal of Forecasting", "quality_score": 0.3076}
{"id": "67f2326af9e454d376caea9815dea19831f94902344204af948c804ec0874a46", "sources": ["arxiv", "semantic_scholar"], "title": "Robust Probabilistic Time Series Forecasting", "abstract": "Probabilistic time series forecasting has played critical role in decision-making processes due to its capability to quantify uncertainties. Deep forecasting models, however, could be prone to input perturbations, and the notion of such perturbations, together with that of robustness, has not even been completely established in the regime of probabilistic forecasting. In this work, we propose a framework for robust probabilistic time series forecasting. First, we generalize the concept of adversarial input perturbations, based on which we formulate the concept of robustness in terms of bounded Wasserstein deviation. Then we extend the randomized smoothing technique to attain robust probabilistic forecasters with theoretical robustness certificates against certain classes of adversarial perturbations. Lastly, extensive experiments demonstrate that our methods are empirically effective in enhancing the forecast quality under additive adversarial attacks and forecast consistency under supplement of noisy observations.", "authors": ["TaeHo Yoon", "Youngsuk Park", "Ernest K. Ryu", "Yuyang Wang"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2022-02-24", "url": "https://arxiv.org/abs/2202.11910", "pdf_url": "https://arxiv.org/pdf/2202.11910v1", "arxiv_id": "2202.11910", "doi": null, "citation_count": 32, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "International Conference on Artificial Intelligence and Statistics", "quality_score": 0.3796}
{"id": "1546b3bd20732f45c66ec323981307e7b8f59b22a04d747ca03c3843181d4e53", "sources": ["arxiv", "semantic_scholar"], "title": "A Differential Attention Fusion Model Based on Transformer for Time Series Forecasting", "abstract": "Time series forecasting is widely used in the fields of equipment life cycle forecasting, weather forecasting, traffic flow forecasting, and other fields. Recently, some scholars have tried to apply Transformer to time series forecasting because of its powerful parallel training ability. However, the existing Transformer methods do not pay enough attention to the small time segments that play a decisive role in prediction, making it insensitive to small changes that affect the trend of time series, and it is difficult to effectively learn continuous time-dependent features. To solve this problem, we propose a differential attention fusion model based on Transformer, which designs the differential layer, neighbor attention, sliding fusion mechanism, and residual layer on the basis of classical Transformer architecture. Specifically, the differences of adjacent time points are extracted and focused by difference and neighbor attention. The sliding fusion mechanism fuses various features of each time point so that the data can participate in encoding and decoding without losing important information. The residual layer including convolution and LSTM further learns the dependence between time points and enables our model to carry out deeper training. A large number of experiments on three datasets show that the prediction results produced by our method are favorably comparable to the state-of-the-art.", "authors": ["Benhan Li", "Shengdong Du", "Tianrui Li"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2022-02-23", "url": "https://arxiv.org/abs/2202.11402", "pdf_url": "https://arxiv.org/pdf/2202.11402v1", "arxiv_id": "2202.11402", "doi": null, "citation_count": 2, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1505}
{"id": "6d046e6fa6f36939240ceac1cdd7a62fa3bec93556280cc35a5603cd722734f3", "sources": ["arxiv", "semantic_scholar"], "title": "Preformer: Predictive Transformer with Multi-Scale Segment-wise Correlations for Long-Term Time Series Forecasting", "abstract": "Transformer-based methods have shown great potential in long-term time series forecasting. However, most of these methods adopt the standard point-wise self-attention mechanism, which not only becomes intractable for long-term forecasting since its complexity increases quadratically with the length of time series, but also cannot explicitly capture the predictive dependencies from contexts since the corresponding key and value are transformed from the same point. This paper proposes a predictive Transformer-based model called {\\em Preformer}. Preformer introduces a novel efficient {\\em Multi-Scale Segment-Correlation} mechanism that divides time series into segments and utilizes segment-wise correlation-based attention for encoding time series. A multi-scale structure is developed to aggregate dependencies at different temporal scales and facilitate the selection of segment length. Preformer further designs a predictive paradigm for decoding, where the key and value come from two successive segments rather than the same segment. In this way, if a key segment has a high correlation score with the query segment, its successive segment contributes more to the prediction of the query segment. Extensive experiments demonstrate that our Preformer outperforms other Transformer-based methods.", "authors": ["Dazhao Du", "Bing Su", "Zhewei Wei"], "categories": ["cs.LG", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2022-02-23", "url": "https://arxiv.org/abs/2202.11356", "pdf_url": "https://arxiv.org/pdf/2202.11356v1", "arxiv_id": "2202.11356", "doi": "10.1109/ICASSP49357.2023.10096881", "citation_count": 84, "influential_citation_count": 4, "has_code": false, "code_url": null, "venue": "IEEE International Conference on Acoustics, Speech, and Signal Processing", "quality_score": 0.4824}
{"id": "30907e7f748972340c6aabc664927dc85e023cbe4bc7ba8444bbbd12a6630245", "sources": ["arxiv", "semantic_scholar"], "title": "Combating Distribution Shift for Accurate Time Series Forecasting via Hypernetworks", "abstract": "Time series forecasting has widespread applications in urban life ranging from air quality monitoring to traffic analysis. However, accurate time series forecasting is challenging because real-world time series suffer from the distribution shift problem, where their statistical properties change over time. Despite extensive solutions to distribution shifts in domain adaptation or generalization, they fail to function effectively in unknown, constantly-changing distribution shifts, which are common in time series. In this paper, we propose Hyper Time- Series Forecasting (HTSF), a hypernetwork-based framework for accurate time series forecasting under distribution shift. HTSF jointly learns the time-varying distributions and the corresponding forecasting models in an end-to-end fashion. Specifically, HTSF exploits the hyper layers to learn the best characterization of the distribution shifts, generating the model parameters for the main layers to make accurate predictions. We implement HTSF as an extensible framework that can incorporate diverse time series forecasting models such as RNNs and Transformers. Extensive experiments on 9 benchmarks demonstrate that HTSF achieves state-of-the-art performances.", "authors": ["Wenying Duan", "Xiaoxi He", "Lu Zhou", "Lothar Thiele", "Hong Rao"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2022-02-22", "url": "https://arxiv.org/abs/2202.10808", "pdf_url": "https://arxiv.org/pdf/2202.10808v2", "arxiv_id": "2202.10808", "doi": "10.1109/ICPADS56603.2022.00121", "citation_count": 18, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "International Conference on Parallel and Distributed Systems", "quality_score": 0.3197}
{"id": "b19ca04ec305d19faf498a9a79beb5c15b2c4777402be20b6348f6a08f1c87b8", "sources": ["arxiv", "semantic_scholar"], "title": "Multi-Objective Model Selection for Time Series Forecasting", "abstract": "Research on time series forecasting has predominantly focused on developing methods that improve accuracy. However, other criteria such as training time or latency are critical in many real-world applications. We therefore address the question of how to choose an appropriate forecasting model for a given dataset among the plethora of available forecasting methods when accuracy is only one of many criteria. For this, our contributions are two-fold. First, we present a comprehensive benchmark, evaluating 7 classical and 6 deep learning forecasting methods on 44 heterogeneous, publicly available datasets. The benchmark code is open-sourced along with evaluations and forecasts for all methods. These evaluations enable us to answer open questions such as the amount of data required for deep learning models to outperform classical ones. Second, we leverage the benchmark evaluations to learn good defaults that consider multiple objectives such as accuracy and latency. By learning a mapping from forecasting models to performance metrics, we show that our method PARETOSELECT is able to accurately select models from the Pareto front -- alleviating the need to train or evaluate many forecasting models for model selection. To the best of our knowledge, PARETOSELECT constitutes the first method to learn default models in a multi-objective setting.", "authors": ["Oliver Borchert", "David Salinas", "Valentin Flunkert", "Tim Januschowski", "Stephan Günnemann"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2022-02-17", "url": "https://arxiv.org/abs/2202.08485", "pdf_url": "https://arxiv.org/pdf/2202.08485v1", "arxiv_id": "2202.08485", "doi": null, "citation_count": 10, "influential_citation_count": 0, "has_code": true, "code_url": null, "venue": "arXiv.org", "quality_score": 0.2603}
{"id": "2219911ca20986c8cd4e6f7c776d328742bc9bb7824071660c672e376882e1fa", "sources": ["arxiv", "semantic_scholar"], "title": "Transformers in Time Series: A Survey", "abstract": "Transformers have achieved superior performances in many tasks in natural language processing and computer vision, which also triggered great interest in the time series community. Among multiple advantages of Transformers, the ability to capture long-range dependencies and interactions is especially attractive for time series modeling, leading to exciting progress in various time series applications. In this paper, we systematically review Transformer schemes for time series modeling by highlighting their strengths as well as limitations. In particular, we examine the development of time series Transformers in two perspectives. From the perspective of network structure, we summarize the adaptations and modifications that have been made to Transformers in order to accommodate the challenges in time series analysis. From the perspective of applications, we categorize time series Transformers based on common tasks including forecasting, anomaly detection, and classification. Empirically, we perform robust analysis, model size analysis, and seasonal-trend decomposition analysis to study how Transformers perform in time series. Finally, we discuss and suggest future directions to provide useful research guidance. To the best of our knowledge, this paper is the first work to comprehensively and systematically summarize the recent advances of Transformers for modeling time series data. We hope this survey will ignite further research interests in time series Transformers.", "authors": ["Qingsong Wen", "Tian Zhou", "Chaoli Zhang", "Weiqi Chen", "Ziqing Ma", "Junchi Yan", "Liang Sun"], "categories": ["cs.LG", "cs.AI", "eess.SP", "stat.ML"], "fields_of_study": ["Computer Science", "Engineering", "Mathematics"], "published_date": "2022-02-15", "url": "https://arxiv.org/abs/2202.07125", "pdf_url": "https://arxiv.org/pdf/2202.07125v5", "arxiv_id": "2202.07125", "doi": "10.24963/ijcai.2023/759", "citation_count": 1487, "influential_citation_count": 52, "has_code": true, "code_url": "https://github.com/qingsongedu/time-series-transformers-review", "venue": "International Joint Conference on Artificial Intelligence", "quality_score": 0.8621}
{"id": "504e812d964020fbc980240997f1ad356df869050f99dd1243e82bffb79168db", "sources": ["arxiv", "semantic_scholar"], "title": "ETSformer: Exponential Smoothing Transformers for Time-series Forecasting", "abstract": "Transformers have been actively studied for time-series forecasting in recent years. While often showing promising results in various scenarios, traditional Transformers are not designed to fully exploit the characteristics of time-series data and thus suffer some fundamental limitations, e.g., they generally lack of decomposition capability and interpretability, and are neither effective nor efficient for long-term forecasting. In this paper, we propose ETSFormer, a novel time-series Transformer architecture, which exploits the principle of exponential smoothing in improving Transformers for time-series forecasting. In particular, inspired by the classical exponential smoothing methods in time-series forecasting, we propose the novel exponential smoothing attention (ESA) and frequency attention (FA) to replace the self-attention mechanism in vanilla Transformers, thus improving both accuracy and efficiency. Based on these, we redesign the Transformer architecture with modular decomposition blocks such that it can learn to decompose the time-series data into interpretable time-series components such as level, growth and seasonality. Extensive experiments on various time-series benchmarks validate the efficacy and advantages of the proposed method. Code is available at https://github.com/salesforce/ETSformer.", "authors": ["Gerald Woo", "Chenghao Liu", "Doyen Sahoo", "Akshat Kumar", "Steven Hoi"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2022-02-03", "url": "https://arxiv.org/abs/2202.01381", "pdf_url": "https://arxiv.org/pdf/2202.01381v2", "arxiv_id": "2202.01381", "doi": null, "citation_count": 307, "influential_citation_count": 30, "has_code": true, "code_url": "https://github.com/salesforce/ETSformer", "venue": "arXiv.org", "quality_score": 0.7457}
{"id": "5c736395feba1a5fd2ecdb7dd51eebfec1da6e14925423eb9439e89d5524ed7d", "sources": ["arxiv", "semantic_scholar"], "title": "Review of automated time series forecasting pipelines", "abstract": "Time series forecasting is fundamental for various use cases in different domains such as energy systems and economics. Creating a forecasting model for a specific use case requires an iterative and complex design process. The typical design process includes the five sections (1) data pre-processing, (2) feature engineering, (3) hyperparameter optimization, (4) forecasting method selection, and (5) forecast ensembling, which are commonly organized in a pipeline structure. One promising approach to handle the ever-growing demand for time series forecasts is automating this design process. The present paper, thus, analyzes the existing literature on automated time series forecasting pipelines to investigate how to automate the design process of forecasting models. Thereby, we consider both Automated Machine Learning (AutoML) and automated statistical forecasting methods in a single forecasting pipeline. For this purpose, we firstly present and compare the proposed automation methods for each pipeline section. Secondly, we analyze the automation methods regarding their interaction, combination, and coverage of the five pipeline sections. For both, we discuss the literature, identify problems, give recommendations, and suggest future research. This review reveals that the majority of papers only cover two or three of the five pipeline sections. We conclude that future research has to holistically consider the automation of the forecasting pipeline to enable the large-scale application of time series forecasting.", "authors": ["Stefan Meisenbacher", "Marian Turowski", "Kaleb Phipps", "Martin Rätz", "Dirk Müller", "Veit Hagenmeyer", "Ralf Mikut"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2022-02-03", "url": "https://arxiv.org/abs/2202.01712", "pdf_url": "https://arxiv.org/pdf/2202.01712v1", "arxiv_id": "2202.01712", "doi": "10.1002/widm.1475", "citation_count": 87, "influential_citation_count": 5, "has_code": false, "code_url": null, "venue": "WIREs Data Mining and Knowledge Discovery (2022) e1475", "quality_score": 0.4861}
{"id": "d052a242cc46021a754c3985a567e7a6d63a63c5b9fb82de1691800695e7ebd0", "sources": ["arxiv", "semantic_scholar"], "title": "Monte Carlo EM for Deep Time Series Anomaly Detection", "abstract": "Time series data are often corrupted by outliers or other kinds of anomalies. Identifying the anomalous points can be a goal on its own (anomaly detection), or a means to improving performance of other time series tasks (e.g. forecasting). Recent deep-learning-based approaches to anomaly detection and forecasting commonly assume that the proportion of anomalies in the training data is small enough to ignore, and treat the unlabeled data as coming from the nominal data distribution. We present a simple yet effective technique for augmenting existing time series models so that they explicitly account for anomalies in the training data. By augmenting the training data with a latent anomaly indicator variable whose distribution is inferred while training the underlying model using Monte Carlo EM, our method simultaneously infers anomalous points while improving model performance on nominal data. We demonstrate the effectiveness of the approach by combining it with a simple feed-forward forecasting model. We investigate how anomalies in the train set affect the training of forecasting models, which are commonly used for time series anomaly detection, and show that our method improves the training of the model.", "authors": ["François-Xavier Aubet", "Daniel Zügner", "Jan Gasthaus"], "categories": ["cs.LG", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2021-12-29", "url": "https://arxiv.org/abs/2112.14436", "pdf_url": "https://arxiv.org/pdf/2112.14436v1", "arxiv_id": "2112.14436", "doi": null, "citation_count": 7, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.2258}
{"id": "46aababb953d0a5422dd8cf84ecab7beaccf1e25c1dbb4587a8a74a8768a1a49", "sources": ["arxiv", "semantic_scholar"], "title": "Parameter Efficient Deep Probabilistic Forecasting", "abstract": "Probabilistic time series forecasting is crucial in many application domains such as retail, ecommerce, finance, or biology. With the increasing availability of large volumes of data, a number of neural architectures have been proposed for this problem. In particular, Transformer-based methods achieve state-of-the-art performance on real-world benchmarks. However, these methods require a large number of parameters to be learned, which imposes high memory requirements on the computational resources for training such models. To address this problem, we introduce a novel Bidirectional Temporal Convolutional Network (BiTCN), which requires an order of magnitude less parameters than a common Transformer-based approach. Our model combines two Temporal Convolutional Networks (TCNs): the first network encodes future covariates of the time series, whereas the second network encodes past observations and covariates. We jointly estimate the parameters of an output distribution via these two networks. Experiments on four real-world datasets show that our method performs on par with four state-of-the-art probabilistic forecasting methods, including a Transformer-based approach and WaveNet, on two point metrics (sMAPE, NRMSE) as well as on a set of range metrics (quantile loss percentiles) in the majority of cases. Secondly, we demonstrate that our method requires significantly less parameters than Transformer-based methods, which means the model can be trained faster with significantly lower memory requirements, which as a consequence reduces the infrastructure cost for deploying these models.", "authors": ["Olivier Sprangers", "Sebastian Schelter", "Maarten de Rijke"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2021-12-06", "url": "https://arxiv.org/abs/2112.02905", "pdf_url": "https://arxiv.org/pdf/2112.02905v2", "arxiv_id": "2112.02905", "doi": "10.1016/j.ijforecast.2021.11.011", "citation_count": 44, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "International Journal of Forecasting", "quality_score": 0.4133}
{"id": "e5d29cf8bed93bcc34992b22f95771359effa8a45296e8ca8cc92c034483ed08", "sources": ["arxiv", "semantic_scholar"], "title": "Learning dynamical systems from data: A simple cross-validation perspective, part III: Irregularly-Sampled Time Series", "abstract": "A simple and interpretable way to learn a dynamical system from data is to interpolate its vector-field with a kernel. In particular, this strategy is highly efficient (both in terms of accuracy and complexity) when the kernel is data-adapted using Kernel Flows (KF)\\cite{Owhadi19} (which uses gradient-based optimization to learn a kernel based on the premise that a kernel is good if there is no significant loss in accuracy if half of the data is used for interpolation). Despite its previous successes, this strategy (based on interpolating the vector field driving the dynamical system) breaks down when the observed time series is not regularly sampled in time. In this work, we propose to address this problem by directly approximating the vector field of the dynamical system by incorporating time differences between observations in the (KF) data-adapted kernels. We compare our approach with the classical one over different benchmark dynamical systems and show that it significantly improves the forecasting accuracy while remaining simple, fast, and robust.", "authors": ["Jonghyeon Lee", "Edward De Brouwer", "Boumediene Hamzi", "Houman Owhadi"], "categories": ["stat.ML", "cs.LG", "math.DS", "stat.CO"], "fields_of_study": ["Mathematics", "Computer Science"], "published_date": "2021-11-25", "url": "https://arxiv.org/abs/2111.13037", "pdf_url": "https://arxiv.org/pdf/2111.13037v2", "arxiv_id": "2111.13037", "doi": "10.1016/j.physd.2023.133853", "citation_count": 25, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "Physica D: Nonlinear Phenomena Volume 454 , 15 November 2023, 133853", "quality_score": 0.3537}
{"id": "4c372c13a41adb302467ce40b973e6d4fd66e3b16f57a47ae6e325cbb20fcf28", "sources": ["arxiv", "semantic_scholar"], "title": "LoMEF: A Framework to Produce Local Explanations for Global Model Time Series Forecasts", "abstract": "Global Forecasting Models (GFM) that are trained across a set of multiple time series have shown superior results in many forecasting competitions and real-world applications compared with univariate forecasting approaches. One aspect of the popularity of statistical forecasting models such as ETS and ARIMA is their relative simplicity and interpretability (in terms of relevant lags, trend, seasonality, and others), while GFMs typically lack interpretability, especially towards particular time series. This reduces the trust and confidence of the stakeholders when making decisions based on the forecasts without being able to understand the predictions. To mitigate this problem, in this work, we propose a novel local model-agnostic interpretability approach to explain the forecasts from GFMs. We train simpler univariate surrogate models that are considered interpretable (e.g., ETS) on the predictions of the GFM on samples within a neighbourhood that we obtain through bootstrapping or straightforwardly as the one-step-ahead global black-box model forecasts of the time series which needs to be explained. After, we evaluate the explanations for the forecasts of the global models in both qualitative and quantitative aspects such as accuracy, fidelity, stability and comprehensibility, and are able to show the benefits of our approach.", "authors": ["Dilini Rajapaksha", "Christoph Bergmeir", "Rob J Hyndman"], "categories": ["cs.LG", "cs.AI", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2021-11-13", "url": "https://arxiv.org/abs/2111.07001", "pdf_url": "https://arxiv.org/pdf/2111.07001v1", "arxiv_id": "2111.07001", "doi": "10.1016/j.ijforecast.2022.06.006", "citation_count": 15, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "International Journal of Forecasting", "quality_score": 0.301}
{"id": "a1d0c1d51ac901bd6d05869bcfc1a7d52098a4e7273650e22b6aa8a84f2f2122", "sources": ["arxiv", "semantic_scholar"], "title": "Probabilistic Hierarchical Forecasting with Deep Poisson Mixtures", "abstract": "Hierarchical forecasting problems arise when time series have a natural group structure, and predictions at multiple levels of aggregation and disaggregation across the groups are needed. In such problems, it is often desired to satisfy the aggregation constraints in a given hierarchy, referred to as hierarchical coherence in the literature. Maintaining coherence while producing accurate forecasts can be a challenging problem, especially in the case of probabilistic forecasting. We present a novel method capable of accurate and coherent probabilistic forecasts for time series when reliable hierarchical information is present. We call it Deep Poisson Mixture Network (DPMN). It relies on the combination of neural networks and a statistical model for the joint distribution of the hierarchical multivariate time series structure. By construction, the model guarantees hierarchical coherence and provides simple rules for aggregation and disaggregation of the predictive distributions. We perform an extensive empirical evaluation comparing the DPMN to other state-of-the-art methods which produce hierarchically coherent probabilistic forecasts on multiple public datasets. Comparing to existing coherent probabilistic models, we obtain a relative improvement in the overall Continuous Ranked Probability Score (CRPS) of 11.8% on Australian domestic tourism data, and 8.1% on the Favorita grocery sales dataset, where time series are grouped with geographical hierarchies or travel intent hierarchies. For San Francisco Bay Area highway traffic, where the series' hierarchical structure is randomly assigned, and their correlations are less informative, our method does not show significant performance differences over statistical baselines.", "authors": ["Kin G. Olivares", "O. Nganba Meetei", "Ruijun Ma", "Rohan Reddy", "Mengfei Cao", "Lee Dicker"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2021-10-25", "url": "https://arxiv.org/abs/2110.13179", "pdf_url": "https://arxiv.org/pdf/2110.13179v8", "arxiv_id": "2110.13179", "doi": "10.1016/j.ijforecast.2023.04.007", "citation_count": 36, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "International Journal of Forecasting", "quality_score": 0.3921}
{"id": "36b7711563033f931657c7c877083d1a6f4b60d8aafe922f2c7e48c414ec3614", "sources": ["arxiv", "semantic_scholar"], "title": "DMS, AE, DAA: methods and applications of adaptive time series model selection, ensemble, and financial evaluation", "abstract": "We introduce three adaptive time series learning methods, called Dynamic Model Selection (DMS), Adaptive Ensemble (AE), and Dynamic Asset Allocation (DAA). The methods respectively handle model selection, ensembling, and contextual evaluation in financial time series. Empirically, we use the methods to forecast the returns of four key indices in the US market, incorporating information from the VIX and Yield curves. We present financial applications of the learning results, including fully-automated portfolios and dynamic hedging strategies. The strategies strongly outperform long-only benchmarks over our testing period, spanning from Q4 2015 to the end of 2021. The key outputs of the learning methods are interpreted during the 2020 market crash.", "authors": ["Parley Ruogu Yang", "Ryan Lucas"], "categories": ["stat.AP", "econ.EM", "q-fin.ST", "stat.ML", "stat.OT"], "fields_of_study": ["Mathematics", "Economics"], "published_date": "2021-10-21", "url": "https://arxiv.org/abs/2110.11156", "pdf_url": "https://arxiv.org/pdf/2110.11156v3", "arxiv_id": "2110.11156", "doi": null, "citation_count": 0, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.0}
{"id": "cac2bc2b8d3bbab98f444715aa1247d1bf517f0d625476ba4eea574f5eaceafe", "sources": ["arxiv", "semantic_scholar"], "title": "Probabilistic Time Series Forecasts with Autoregressive Transformation Models", "abstract": "Probabilistic forecasting of time series is an important matter in many applications and research fields. In order to draw conclusions from a probabilistic forecast, we must ensure that the model class used to approximate the true forecasting distribution is expressive enough. Yet, characteristics of the model itself, such as its uncertainty or its feature-outcome relationship are not of lesser importance. This paper proposes Autoregressive Transformation Models (ATMs), a model class inspired by various research directions to unite expressive distributional forecasts using a semi-parametric distribution assumption with an interpretable model specification. We demonstrate the properties of ATMs both theoretically and through empirical evaluation on several simulated and real-world forecasting datasets.", "authors": ["David Rügamer", "Philipp F. M. Baumann", "Thomas Kneib", "Torsten Hothorn"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2021-10-15", "url": "https://arxiv.org/abs/2110.08248", "pdf_url": "https://arxiv.org/pdf/2110.08248v3", "arxiv_id": "2110.08248", "doi": "10.1007/s11222-023-10212-8", "citation_count": 14, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "Statistics and computing", "quality_score": 0.294}
{"id": "d2fd751119db495fec11f1661b1351f351ac98cce5283dc3d623ca7b1a51cce0", "sources": ["arxiv", "semantic_scholar"], "title": "Well Googled is Half Done: Multimodal Forecasting of New Fashion Product Sales with Image-based Google Trends", "abstract": "New fashion product sales forecasting is a challenging problem that involves many business dynamics and cannot be solved by classical forecasting approaches. In this paper, we investigate the effectiveness of systematically probing exogenous knowledge in the form of Google Trends time series and combining it with multi-modal information related to a brand-new fashion item, in order to effectively forecast its sales despite the lack of past data. In particular, we propose a neural network-based approach, where an encoder learns a representation of the exogenous time series, while the decoder forecasts the sales based on the Google Trends encoding and the available visual and metadata information. Our model works in a non-autoregressive manner, avoiding the compounding effect of large first-step errors. As a second contribution, we present VISUELLE, a publicly available dataset for the task of new fashion product sales forecasting, containing multimodal information for 5577 real, new products sold between 2016-2019 from Nunalie, an Italian fast-fashion company. The dataset is equipped with images of products, metadata, related sales, and associated Google Trends. We use VISUELLE to compare our approach against state-of-the-art alternatives and several baselines, showing that our neural network-based approach is the most accurate in terms of both percentage and absolute error. It is worth noting that the addition of exogenous knowledge boosts the forecasting accuracy by 1.5% in terms of Weighted Absolute Percentage Error (WAPE), revealing the importance of exploiting informative external information. The code and dataset are both available at https://github.com/HumaticsLAB/GTM-Transformer.", "authors": ["Geri Skenderi", "Christian Joppi", "Matteo Denitto", "Marco Cristani"], "categories": ["cs.CV", "cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2021-09-20", "url": "https://arxiv.org/abs/2109.09824", "pdf_url": "https://arxiv.org/pdf/2109.09824v6", "arxiv_id": "2109.09824", "doi": "10.1002/for.3104", "citation_count": 40, "influential_citation_count": 8, "has_code": true, "code_url": "https://github.com/HumaticsLAB/GTM-Transformer", "venue": "Journal of Forecasting", "quality_score": 0.4771}
{"id": "7a8595eb54601b5090e22a1852812319848d752b537ec0efab20140aa67dbb77", "sources": ["arxiv", "semantic_scholar"], "title": "A Study of Joint Graph Inference and Forecasting", "abstract": "We study a recent class of models which uses graph neural networks (GNNs) to improve forecasting in multivariate time series. The core assumption behind these models is that there is a latent graph between the time series (nodes) that governs the evolution of the multivariate time series. By parameterizing a graph in a differentiable way, the models aim to improve forecasting quality. We compare four recent models of this class on the forecasting task. Further, we perform ablations to study their behavior under changing conditions, e.g., when disabling the graph-learning modules and providing the ground-truth relations instead. Based on our findings, we propose novel ways of combining the existing architectures.", "authors": ["Daniel Zügner", "François-Xavier Aubet", "Victor Garcia Satorras", "Tim Januschowski", "Stephan Günnemann", "Jan Gasthaus"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2021-09-10", "url": "https://arxiv.org/abs/2109.04979", "pdf_url": "https://arxiv.org/pdf/2109.04979v1", "arxiv_id": "2109.04979", "doi": null, "citation_count": 12, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.2785}
{"id": "d23043f4a8768bdba685f09702a8a1db88b4d20e98bbdf6cfc16225b34a20517", "sources": ["arxiv", "semantic_scholar"], "title": "TCCT: Tightly-Coupled Convolutional Transformer on Time Series Forecasting", "abstract": "Time series forecasting is essential for a wide range of real-world applications. Recent studies have shown the superiority of Transformer in dealing with such problems, especially long sequence time series input(LSTI) and long sequence time series forecasting(LSTF) problems. To improve the efficiency and enhance the locality of Transformer, these studies combine Transformer with CNN in varying degrees. However, their combinations are loosely-coupled and do not make full use of CNN. To address this issue, we propose the concept of tightly-coupled convolutional Transformer(TCCT) and three TCCT architectures which apply transformed CNN architectures into Transformer: (1) CSPAttention: through fusing CSPNet with self-attention mechanism, the computation cost of self-attention mechanism is reduced by 30% and the memory usage is reduced by 50% while achieving equivalent or beyond prediction accuracy. (2) Dilated causal convolution: this method is to modify the distilling operation proposed by Informer through replacing canonical convolutional layers with dilated causal convolutional layers to gain exponentially receptive field growth. (3) Passthrough mechanism: the application of passthrough mechanism to stack of self-attention blocks helps Transformer-like models get more fine-grained information with negligible extra computation costs. Our experiments on real-world datasets show that our TCCT architectures could greatly improve the performance of existing state-of-art Transformer models on time series forecasting with much lower computation and memory costs, including canonical Transformer, LogTrans and Informer.", "authors": ["Li Shen", "Yangzhu Wang"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2021-08-29", "url": "https://arxiv.org/abs/2108.12784", "pdf_url": "https://arxiv.org/pdf/2108.12784v2", "arxiv_id": "2108.12784", "doi": "10.1016/j.neucom.2022.01.039", "citation_count": 123, "influential_citation_count": 8, "has_code": false, "code_url": null, "venue": "Neurocomputing", "quality_score": 0.5234}
{"id": "4b93102d1db908cd47c4ae9a678ef6a44cae8bcc6ca2b1d50eee4e0f1d211b4a", "sources": ["arxiv", "semantic_scholar"], "title": "Transformers predicting the future. Applying attention in next-frame and time series forecasting", "abstract": "Recurrent Neural Networks were, until recently, one of the best ways to capture the timely dependencies in sequences. However, with the introduction of the Transformer, it has been proven that an architecture with only attention-mechanisms without any RNN can improve on the results in various sequence processing tasks (e.g. NLP). Multiple studies since then have shown that similar approaches can be applied for images, point clouds, video, audio or time series forecasting. Furthermore, solutions such as the Perceiver or the Informer have been introduced to expand on the applicability of the Transformer. Our main objective is testing and evaluating the effectiveness of applying Transformer-like models on time series data, tackling susceptibility to anomalies, context awareness and space complexity by fine-tuning the hyperparameters, preprocessing the data, applying dimensionality reduction or convolutional encodings, etc. We are also looking at the problem of next-frame prediction and exploring ways to modify existing solutions in order to achieve higher performance and learn generalized knowledge.", "authors": ["Radostin Cholakov", "Todor Kolev"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2021-08-18", "url": "https://arxiv.org/abs/2108.08224", "pdf_url": "https://arxiv.org/pdf/2108.08224v1", "arxiv_id": "2108.08224", "doi": null, "citation_count": 24, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.3495}
{"id": "2294105a11853b818bf889384b2defe26b08a64792b36f8f14ada9d873f53fdf", "sources": ["arxiv", "semantic_scholar"], "title": "Tsformer: Time series Transformer for tourism demand forecasting", "abstract": "AI-based methods have been widely applied to tourism demand forecasting. However, current AI-based methods are short of the ability to process long-term dependency, and most of them lack interpretability. The Transformer used initially for machine translation shows an incredible ability to long-term dependency processing. Based on the Transformer, we proposed a time series Transformer (Tsformer) with Encoder-Decoder architecture for tourism demand forecasting. The proposed Tsformer encodes long-term dependency with encoder, captures short-term dependency with decoder, and simplifies the attention interactions under the premise of highlighting dominant attention through a series of attention masking mechanisms. These improvements make the multi-head attention mechanism process the input sequence according to the time relationship, contributing to better interpretability. What's more, the context processing ability of the Encoder-Decoder architecture allows adopting the calendar of days to be forecasted to enhance the forecasting performance. Experiments conducted on the Jiuzhaigou valley and Siguniang mountain tourism demand datasets with other nine baseline methods indicate that the proposed Tsformer outperformed all baseline models in the short-term and long-term tourism demand forecasting tasks. Moreover, ablation studies demonstrate that the adoption of the calendar of days to be forecasted contributes to the forecasting performance of the proposed Tsformer. For better interpretability, the attention weight matrix visualization is performed. It indicates that the Tsformer concentrates on seasonal features and days close to days to be forecast in short-term forecasting.", "authors": ["Siyuan Yi", "Xing Chen", "Chuanming Tang"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2021-07-22", "url": "https://arxiv.org/abs/2107.10977", "pdf_url": "https://arxiv.org/pdf/2107.10977v1", "arxiv_id": "2107.10977", "doi": null, "citation_count": 4, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.1747}
{"id": "1c5ea74fae5b88a3641e612ba334206924e38d005a0ca79f89ca02ae0ac34916", "sources": ["arxiv", "semantic_scholar"], "title": "Dynamic functional time-series forecasts of foreign exchange implied volatility surfaces", "abstract": "This paper presents static and dynamic versions of univariate, multivariate, and multilevel functional time-series methods to forecast implied volatility surfaces in foreign exchange markets. We find that dynamic functional principal component analysis generally improves out-of-sample forecast accuracy. More specifically, the dynamic univariate functional time-series method shows the greatest improvement. Our models lead to multiple instances of statistically significant improvements in forecast accuracy for daily EUR-USD, EUR-GBP, and EUR-JPY implied volatility surfaces across various maturities, when benchmarked against established methods. A stylised trading strategy is also employed to demonstrate the potential economic benefits of our proposed approach.", "authors": ["Han Lin Shang", "Fearghal Kearney"], "categories": ["q-fin.ST", "stat.AP", "stat.CO"], "fields_of_study": ["Economics", "Mathematics"], "published_date": "2021-07-22", "url": "https://arxiv.org/abs/2107.14026", "pdf_url": "https://arxiv.org/pdf/2107.14026v1", "arxiv_id": "2107.14026", "doi": "10.1016/j.ijforecast.2021.07.011", "citation_count": 23, "influential_citation_count": 3, "has_code": false, "code_url": null, "venue": "International Journal of Forecasting", "quality_score": 0.3451}
{"id": "601422d9a129048af6914f4c4a0f629b2db33398abb7247348bbeba746286d57", "sources": ["arxiv", "semantic_scholar"], "title": "Topological Attention for Time Series Forecasting", "abstract": "The problem of (point) forecasting $ \\textit{univariate} $ time series is considered. Most approaches, ranging from traditional statistical methods to recent learning-based techniques with neural networks, directly operate on raw time series observations. As an extension, we study whether $\\textit{local topological properties}$, as captured via persistent homology, can serve as a reliable signal that provides complementary information for learning to forecast. To this end, we propose $\\textit{topological attention}$, which allows attending to local topological features within a time horizon of historical data. Our approach easily integrates into existing end-to-end trainable forecasting models, such as $\\texttt{N-BEATS}$, and in combination with the latter exhibits state-of-the-art performance on the large-scale M4 benchmark dataset of 100,000 diverse time series from different domains. Ablation experiments, as well as a comparison to a broad range of forecasting methods in a setting where only a single time series is available for training, corroborate the beneficial nature of including local topological information through an attention mechanism.", "authors": ["Sebastian Zeng", "Florian Graf", "Christoph Hofer", "Roland Kwitt"], "categories": ["cs.LG", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2021-07-19", "url": "https://arxiv.org/abs/2107.09031", "pdf_url": "https://arxiv.org/pdf/2107.09031v1", "arxiv_id": "2107.09031", "doi": null, "citation_count": 40, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "Neural Information Processing Systems", "quality_score": 0.4032}
{"id": "a2648627852380ba511149030091f8c8f16e5b871be02e276c3a0cded133ec08", "sources": ["arxiv", "semantic_scholar"], "title": "Probabilistic Time Series Forecasting with Implicit Quantile Networks", "abstract": "Here, we propose a general method for probabilistic time series forecasting. We combine an autoregressive recurrent neural network to model temporal dynamics with Implicit Quantile Networks to learn a large class of distributions over a time-series target. When compared to other probabilistic neural forecasting models on real- and simulated data, our approach is favorable in terms of point-wise prediction accuracy as well as on estimating the underlying temporal distribution.", "authors": ["Adèle Gouttes", "Kashif Rasul", "Mateusz Koren", "Johannes Stephan", "Tofigh Naghibi"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2021-07-08", "url": "https://arxiv.org/abs/2107.03743", "pdf_url": "https://arxiv.org/pdf/2107.03743v1", "arxiv_id": "2107.03743", "doi": null, "citation_count": 31, "influential_citation_count": 6, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.4225}
{"id": "68f5d4603dc5fc84f663c353325d037678e8e78041b33e8c68b989fc7e90bf3d", "sources": ["arxiv", "semantic_scholar"], "title": "SCINet: Time Series Modeling and Forecasting with Sample Convolution and Interaction", "abstract": "One unique property of time series is that the temporal relations are largely preserved after downsampling into two sub-sequences. By taking advantage of this property, we propose a novel neural network architecture that conducts sample convolution and interaction for temporal modeling and forecasting, named SCINet. Specifically, SCINet is a recursive downsample-convolve-interact architecture. In each layer, we use multiple convolutional filters to extract distinct yet valuable temporal features from the downsampled sub-sequences or features. By combining these rich features aggregated from multiple resolutions, SCINet effectively models time series with complex temporal dynamics. Experimental results show that SCINet achieves significant forecasting accuracy improvements over both existing convolutional models and Transformer-based solutions across various real-world time series forecasting datasets. Our codes and data are available at https://github.com/cure-lab/SCINet.", "authors": ["Minhao Liu", "Ailing Zeng", "Muxi Chen", "Zhijian Xu", "Qiuxia Lai", "Lingna Ma", "Qiang Xu"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2021-06-17", "url": "https://arxiv.org/abs/2106.09305", "pdf_url": "https://arxiv.org/pdf/2106.09305v3", "arxiv_id": "2106.09305", "doi": "10.52202/068431-0421", "citation_count": 819, "influential_citation_count": 47, "has_code": true, "code_url": "https://github.com/cure-lab/SCINet", "venue": "Neural Information Processing Systems", "quality_score": 0.8406}
{"id": "d4bb68e4ba77ce7a1ed6342ebbbdd29420a799d6e22f88c60e79cd4644b39615", "sources": ["arxiv", "semantic_scholar"], "title": "Monash Time Series Forecasting Archive", "abstract": "Many businesses and industries nowadays rely on large quantities of time series data making time series forecasting an important research area. Global forecasting models that are trained across sets of time series have shown a huge potential in providing accurate forecasts compared with the traditional univariate forecasting models that work on isolated series. However, there are currently no comprehensive time series archives for forecasting that contain datasets of time series from similar sources available for the research community to evaluate the performance of new global forecasting algorithms over a wide variety of datasets. In this paper, we present such a comprehensive time series forecasting archive containing 20 publicly available time series datasets from varied domains, with different characteristics in terms of frequency, series lengths, and inclusion of missing values. We also characterise the datasets, and identify similarities and differences among them, by conducting a feature analysis. Furthermore, we present the performance of a set of standard baseline forecasting methods over all datasets across eight error metrics, for the benefit of researchers using the archive to benchmark their forecasting algorithms.", "authors": ["Rakshitha Godahewa", "Christoph Bergmeir", "Geoffrey I. Webb", "Rob J. Hyndman", "Pablo Montero-Manso"], "categories": ["cs.LG", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2021-05-14", "url": "https://arxiv.org/abs/2105.06643", "pdf_url": "https://arxiv.org/pdf/2105.06643v1", "arxiv_id": "2105.06643", "doi": null, "citation_count": 267, "influential_citation_count": 36, "has_code": false, "code_url": null, "venue": "Neural Information Processing Systems Track on Datasets and Benchmarks (2021) - forthcoming", "quality_score": 0.7841}
{"id": "90c80c524ce35900f82e99763f0634d86154645839ad12e7910cef0c82256af6", "sources": ["arxiv", "semantic_scholar"], "title": "Time Series Forecasting via Learning Convolutionally Low-Rank Models", "abstract": "Recently, Liu and Zhang studied the rather challenging problem of time series forecasting from the perspective of compressed sensing. They proposed a no-learning method, named Convolution Nuclear Norm Minimization (CNNM), and proved that CNNM can exactly recover the future part of a series from its observed part, provided that the series is convolutionally low-rank. While impressive, the convolutional low-rankness condition may not be satisfied whenever the series is far from being seasonal, and is in fact brittle to the presence of trends and dynamics. This paper tries to approach the issues by integrating a learnable, orthonormal transformation into CNNM, with the purpose for converting the series of involute structures into regular signals of convolutionally low-rank. We prove that the resultant model, termed Learning-Based CNNM (LbCNNM), strictly succeeds in identifying the future part of a series, as long as the transform of the series is convolutionally low-rank. To learn proper transformations that may meet the required success conditions, we devise an interpretable method based on Principal Component Pursuit (PCP). Equipped with this learning method and some elaborate data argumentation skills, LbCNNM not only can handle well the major components of time series (including trends, seasonality and dynamics), but also can make use of the forecasts provided by some other forecasting methods; this means LbCNNM can be used as a general tool for model combination. Extensive experiments on 100,452 real-world time series from Time Series Data Library (TSDL) and M4 Competition (M4) demonstrate the superior performance of LbCNNM.", "authors": ["Guangcan Liu"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2021-04-23", "url": "https://arxiv.org/abs/2104.11510", "pdf_url": "https://arxiv.org/pdf/2104.11510v5", "arxiv_id": "2104.11510", "doi": "10.1109/TIT.2022.3144605", "citation_count": 19, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "IEEE Transactions on Information Theory", "quality_score": 0.3253}
{"id": "1b054a417f50ba51f9fac9e82a39a6f9d774504bd3502d9f3995f72012519355", "sources": ["arxiv", "semantic_scholar"], "title": "Boosted Embeddings for Time Series Forecasting", "abstract": "Time series forecasting is a fundamental task emerging from diverse data-driven applications. Many advanced autoregressive methods such as ARIMA were used to develop forecasting models. Recently, deep learning based methods such as DeepAr, NeuralProphet, Seq2Seq have been explored for time series forecasting problem. In this paper, we propose a novel time series forecast model, DeepGB. We formulate and implement a variant of Gradient boosting wherein the weak learners are DNNs whose weights are incrementally found in a greedy manner over iterations. In particular, we develop a new embedding architecture that improves the performance of many deep learning models on time series using Gradient boosting variant. We demonstrate that our model outperforms existing comparable state-of-the-art models using real-world sensor data and public dataset.", "authors": ["Sankeerth Rao Karingula", "Nandini Ramanan", "Rasool Tahmasbi", "Mehrnaz Amjadi", "Deokwoo Jung", "Ricky Si", "Charanraj Thimmisetty", "Luisa Polania Cabrera", "Marjorie Sayer", "Claudionor Nunes Coelho"], "categories": ["cs.LG", "cs.AI", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2021-04-10", "url": "https://arxiv.org/abs/2104.04781", "pdf_url": "https://arxiv.org/pdf/2104.04781v2", "arxiv_id": "2104.04781", "doi": "10.1007/978-3-030-95470-3_1", "citation_count": 17, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "International Conference on Machine Learning, Optimization, and Data Science", "quality_score": 0.3138}
{"id": "dbade99861019a3756df13218363dd4e55780884cafb055b356f3470ff6d7f27", "sources": ["arxiv", "semantic_scholar"], "title": "Time Series Analysis and Modeling to Forecast: a Survey", "abstract": "Time series modeling for predictive purpose has been an active research area of machine learning for many years. However, no sufficiently comprehensive and meanwhile substantive survey was offered so far. This survey strives to meet this need. A unified presentation has been adopted for entire parts of this compilation. A red thread guides the reader from time series preprocessing to forecasting. Time series decomposition is a major preprocessing task, to separate nonstationary effects (the deterministic components) from the remaining stochastic constituent, assumed to be stationary. The deterministic components are predictable and contribute to the prediction through estimations or extrapolation. Fitting the most appropriate model to the remaining stochastic component aims at capturing the relationship between past and future values, to allow prediction. We cover a sufficiently broad spectrum of models while nonetheless offering substantial methodological developments. We describe three major linear parametric models, together with two nonlinear extensions, and present five categories of nonlinear parametric models. Beyond conventional statistical models, we highlight six categories of deep neural networks appropriate for time series forecasting in nonlinear framework. Finally, we enlighten new avenues of research for time series modeling and forecasting. We also report software made publicly available for the models presented.", "authors": ["Fatoumata Dama", "Christine Sinoquet"], "categories": ["cs.LG", "cs.AI"], "fields_of_study": ["Computer Science"], "published_date": "2021-03-31", "url": "https://arxiv.org/abs/2104.00164", "pdf_url": "https://arxiv.org/pdf/2104.00164v2", "arxiv_id": "2104.00164", "doi": null, "citation_count": 39, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.4005}
{"id": "e03c67e740db87a9c61a7be6df0fcb75875bfd09b3e64ad766ab2406155c5c1e", "sources": ["arxiv", "semantic_scholar"], "title": "Hierarchical forecasting with a top-down alignment of independent level forecasts", "abstract": "Hierarchical forecasting with intermittent time series is a challenge in both research and empirical studies. Extensive research focuses on improving the accuracy of each hierarchy, especially the intermittent time series at bottom levels. Then hierarchical reconciliation could be used to improve the overall performance further. In this paper, we present a \\emph{hierarchical-forecasting-with-alignment} approach that treats the bottom level forecasts as mutable to ensure higher forecasting accuracy on the upper levels of the hierarchy. We employ a pure deep learning forecasting approach N-BEATS for continuous time series at the top levels and a widely used tree-based algorithm LightGBM for the intermittent time series at the bottom level. The \\emph{hierarchical-forecasting-with-alignment} approach is a simple yet effective variant of the bottom-up method, accounting for biases that are difficult to observe at the bottom level. It allows suboptimal forecasts at the lower level to retain a higher overall performance. The approach in this empirical study was developed by the first author during the M5 Forecasting Accuracy competition, ranking second place. The method is also business orientated and could benefit for business strategic planning.", "authors": ["Matthias Anderer", "Feng Li"], "categories": ["stat.ML", "cs.LG", "stat.AP"], "fields_of_study": ["Mathematics", "Computer Science"], "published_date": "2021-03-15", "url": "https://arxiv.org/abs/2103.08250", "pdf_url": "https://arxiv.org/pdf/2103.08250v4", "arxiv_id": "2103.08250", "doi": "10.1016/j.ijforecast.2021.12.015", "citation_count": 21, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": "International Journal of Forecasting", "quality_score": 0.3356}
{"id": "1fd9ab3c6a3c586ed1afc58566269c4c8d13c16006534c1046d5ad813e8196ea", "sources": ["arxiv", "semantic_scholar"], "title": "Forecasting high-frequency financial time series: an adaptive learning approach with the order book data", "abstract": "This paper proposes a forecast-centric adaptive learning model that engages with the past studies on the order book and high-frequency data, with applications to hypothesis testing. In line with the past literature, we produce brackets of summaries of statistics from the high-frequency bid and ask data in the CSI 300 Index Futures market and aim to forecast the one-step-ahead prices. Traditional time series issues, e.g. ARIMA order selection, stationarity, together with potential financial applications are covered in the exploratory data analysis, which pave paths to the adaptive learning model. By designing and running the learning model, we found it to perform well compared to the top fixed models, and some could improve the forecasting accuracy by being more stable and resilient to non-stationarity. Applications to hypothesis testing are shown with a rolling window, and further potential applications to finance and statistics are outlined.", "authors": ["Parley Ruogu Yang"], "categories": ["q-fin.ST", "econ.EM", "q-fin.TR", "stat.AP"], "fields_of_study": ["Computer Science", "Economics", "Mathematics"], "published_date": "2021-02-27", "url": "https://arxiv.org/abs/2103.00264", "pdf_url": "https://arxiv.org/pdf/2103.00264v1", "arxiv_id": "2103.00264", "doi": "10.20944/PREPRINTS202103.0269.V1", "citation_count": 3, "influential_citation_count": 0, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.1505}
{"id": "4616817c6a5b27b06738b89c15e51521a78f31fc8c437b33f09c822391cc7a63", "sources": ["arxiv", "semantic_scholar"], "title": "NAST: Non-Autoregressive Spatial-Temporal Transformer for Time Series Forecasting", "abstract": "Although Transformer has made breakthrough success in widespread domains especially in Natural Language Processing (NLP), applying it to time series forecasting is still a great challenge. In time series forecasting, the autoregressive decoding of canonical Transformer models could introduce huge accumulative errors inevitably. Besides, utilizing Transformer to deal with spatial-temporal dependencies in the problem still faces tough difficulties.~To tackle these limitations, this work is the first attempt to propose a Non-Autoregressive Transformer architecture for time series forecasting, aiming at overcoming the time delay and accumulative error issues in the canonical Transformer. Moreover, we present a novel spatial-temporal attention mechanism, building a bridge by a learned temporal influence map to fill the gaps between the spatial and temporal attention, so that spatial and temporal dependencies can be processed integrally. Empirically, we evaluate our model on diversified ego-centric future localization datasets and demonstrate state-of-the-art performance on both real-time and accuracy.", "authors": ["Kai Chen", "Guang Chen", "Dan Xu", "Lijun Zhang", "Yuyao Huang", "Alois Knoll"], "categories": ["cs.LG", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2021-02-10", "url": "https://arxiv.org/abs/2102.05624", "pdf_url": "https://arxiv.org/pdf/2102.05624v2", "arxiv_id": "2102.05624", "doi": null, "citation_count": 28, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.3656}
{"id": "8dd00a2249f8dfe7305cbe6258791ed9a4076f80a13d211222ccc59dcf2bf45d", "sources": ["arxiv", "semantic_scholar"], "title": "Learning Interpretable Deep State Space Model for Probabilistic Time Series Forecasting", "abstract": "Probabilistic time series forecasting involves estimating the distribution of future based on its history, which is essential for risk management in downstream decision-making. We propose a deep state space model for probabilistic time series forecasting whereby the non-linear emission model and transition model are parameterized by networks and the dependency is modeled by recurrent neural nets. We take the automatic relevance determination (ARD) view and devise a network to exploit the exogenous variables in addition to time series. In particular, our ARD network can incorporate the uncertainty of the exogenous variables and eventually helps identify useful exogenous variables and suppress those irrelevant for forecasting. The distribution of multi-step ahead forecasts are approximated by Monte Carlo simulation. We show in experiments that our model produces accurate and sharp probabilistic forecasts. The estimated uncertainty of our forecasting also realistically increases over time, in a spontaneous manner.", "authors": ["Longyuan Li", "Junchi Yan", "Xiaokang Yang", "Yaohui Jin"], "categories": ["cs.LG", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2021-01-31", "url": "https://arxiv.org/abs/2102.00397", "pdf_url": "https://arxiv.org/pdf/2102.00397v1", "arxiv_id": "2102.00397", "doi": "10.24963/ijcai.2019/402", "citation_count": 71, "influential_citation_count": 6, "has_code": false, "code_url": null, "venue": "International Joint Conference on Artificial Intelligence", "quality_score": 0.4643}
{"id": "7c7e107e5e6d0769aad861715c417df215e81b370e3e275305a14822aa8a9b25", "sources": ["arxiv", "semantic_scholar"], "title": "Multi-Task Time Series Forecasting With Shared Attention", "abstract": "Time series forecasting is a key component in many industrial and business decision processes and recurrent neural network (RNN) based models have achieved impressive progress on various time series forecasting tasks. However, most of the existing methods focus on single-task forecasting problems by learning separately based on limited supervised objectives, which often suffer from insufficient training instances. As the Transformer architecture and other attention-based models have demonstrated its great capability of capturing long term dependency, we propose two self-attention based sharing schemes for multi-task time series forecasting which can train jointly across multiple tasks. We augment a sequence of paralleled Transformer encoders with an external public multi-head attention function, which is updated by all data of all tasks. Experiments on a number of real-world multi-task time series forecasting tasks show that our proposed architectures can not only outperform the state-of-the-art single-task forecasting baselines but also outperform the RNN-based multi-task forecasting method.", "authors": ["Zekai Chen", "Jiaze E", "Xiao Zhang", "Hao Sheng", "Xiuzheng Cheng"], "categories": ["cs.LG", "stat.ML"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2021-01-24", "url": "https://arxiv.org/abs/2101.09645", "pdf_url": "https://arxiv.org/pdf/2101.09645v1", "arxiv_id": "2101.09645", "doi": "10.1109/ICDMW51313.2020.00132", "citation_count": 27, "influential_citation_count": 3, "has_code": false, "code_url": null, "venue": null, "quality_score": 0.3618}
{"id": "0aca506570b77208161b7bbc0f5e5d6edd94d0729a70612592ed480d40640878", "sources": ["arxiv", "semantic_scholar"], "title": "A Trainable Reconciliation Method for Hierarchical Time-Series", "abstract": "In numerous applications, it is required to produce forecasts for multiple time-series at different hierarchy levels. An obvious example is given by the supply chain in which demand forecasting may be needed at a store, city, or country level. The independent forecasts typically do not add up properly because of the hierarchical constraints, so a reconciliation step is needed. In this paper, we propose a new general, flexible, and easy-to-implement reconciliation strategy based on an encoder-decoder neural network. By testing our method on four real-world datasets, we show that it can consistently reach or surpass the performance of existing methods in the reconciliation setting.", "authors": ["Davide Burba", "Trista Chen"], "categories": ["cs.LG"], "fields_of_study": ["Computer Science"], "published_date": "2021-01-05", "url": "https://arxiv.org/abs/2101.01329", "pdf_url": "https://arxiv.org/pdf/2101.01329v1", "arxiv_id": "2101.01329", "doi": null, "citation_count": 10, "influential_citation_count": 2, "has_code": false, "code_url": null, "venue": "arXiv.org", "quality_score": 0.2603}
{"id": "82ebb6afc23df529847b66f35268ad1557a8cafdb774eb2810ebe50f6e5795d8", "sources": ["arxiv", "semantic_scholar"], "title": "Improving forecasting by subsampling seasonal time series", "abstract": "Time series forecasting plays an increasingly important role in modern business decisions. In today's data-rich environment, people often aim to choose the optimal forecasting model for their data. However, identifying the optimal model requires professional knowledge and experience, making accurate forecasting a challenging task. To mitigate the importance of model selection, we propose a simple and reliable algorithm to improve the forecasting performance. Specifically, we construct multiple time series with different sub-seasons from the original time series. These derived series highlight different sub-seasonal patterns of the original series, making it possible for the forecasting methods to capture diverse patterns and components of the data. Subsequently, we produce forecasts for these multiple series separately with classical statistical models (ETS or ARIMA). Finally, the forecasts are combined. We evaluate our approach on widely-used forecasting competition data sets (M1, M3, and M4) in terms of both point forecasts and prediction intervals. We observe performance improvements compared with the benchmarks. Our approach is particularly suitable and robust for the data with higher frequency. To demonstrate the practical value of our proposition, we showcase the performance improvements from our approach on hourly load data that exhibit multiple seasonal patterns.", "authors": ["Xixi Li", "Fotios Petropoulos", "Yanfei Kang"], "categories": ["stat.AP"], "fields_of_study": ["Computer Science", "Mathematics"], "published_date": "2021-01-04", "url": "https://arxiv.org/abs/2101.00827", "pdf_url": "https://arxiv.org/pdf/2101.00827v4", "arxiv_id": "2101.00827", "doi": "10.1080/00207543.2021.2022800", "citation_count": 14, "influential_citation_count": 1, "has_code": false, "code_url": null, "venue": "International Journal of Production Research", "quality_score": 0.294}