text stringlengths 13.9k 38.7k | labels class label 4
classes | id stringlengths 10 10 |
|---|---|---|
# A PPROACHING R ATE -D ISTORTION L IMITS IN N EURAL C OMPRESSION WITH L ATTICE T RANSFORM C ODING **Eric Lei, Hamed Hassani & Shirin Saeedi Bidokhti** Department of Electrical and Systems Engineering, University of Pennsylvania {elei, hassani, saeedi}@seas.upenn.edu A BSTRACT Neural compression has brought tremendous ... | 0Conceptual Integration | Tv36j85SqR |
# - - S CALING I N THE -W ILD T RAINING FOR D IFFUSION BASED I LLUMINATION H ARMONIZATION AND E DITING BY I MPOSING C ONSISTENT L IGHT T RANSPORT **Lvmin Zhang** **[1]** **, Anyi Rao** **[2]** **, Maneesh Agrawala** **[1]** 1 Stanford University, 2 Hong Kong University of Science and Technology _{_ lvmin,maneesh _}_ @c... | 0Conceptual Integration | u1cQYxRI1H |
# - D IFFUSION O N S YNTAX T REES F OR P ROGRAM S YN ## THESIS **Shreyas Kapur, Erik Jenner & Stuart Russell** University of California, Berkeley _{_ srkp,jenner,russell _}_ @berkeley.edu A BSTRACT Large language models generate code one token at a time. Their autoregressive generation process lacks the feedback of obs... | 0Conceptual Integration | wN3KaUXA5X |
# - S EG LLM: M ULTI ROUND R EASONING S EGMENTATION WITH L ARGE L ANGUAGE M ODELS **XuDong Wang** _[∗]_ [1] **Shaolun Zhang** _[∗]_ [1] **Shufan Li** _[∗]_ [2] **Konstantinos Kallidromitis** [3] **Kehan Li** [1] _[,]_ [4] **Yusuke Kato** [3] **Kazuki Kozuka** [3] **Trevor Darrell** [1] 1 UC Berkeley 2 UCLA 3 Panasonic ... | 2Direct Enhancement | Pm1NXHgzyf |
# - D ATA ADAPTIVE D IFFERENTIALLY P RIVATE P ROMPT S YNTHESIS FOR I N -C ONTEXT L EARNING **Fengyu Gao** _[∗]_ University of Virginia wan6jj@virginia.edu **Tianhao Wang** University of Virginia tianhao@virginia.edu **Cong Shen** University of Virginia cong@virginia.edu A BSTRACT **Ruida Zhou** _[∗]_ University of Cali... | 2Direct Enhancement | sVNfWhtaJC |
# V ISION L ANGUAGE M ODELS ARE I N -C ONTEXT V ALUE L EARNERS **Yecheng Jason Ma** _[†][,]_ [1] _[,]_ [2] **, Joey Hejna** [1] _[,]_ [3] **, Ayzaan Wahid** [1] **, Chuyuan Fu** [1] **, Dhruv Shah** [1] **, Jacky Liang** [1] **,** **Zhuo Xu** [1] **, Sean Kirmani** [1] **, Peng Xu** [1] **, Danny Driess** [1] **, Ted X... | 0Conceptual Integration | friHAl5ofG |
# B ACKTRACKING I MPROVES G ENERATION S AFETY **Yiming Zhang** [1] _[,]_ [2] **Jianfeng Chi** [1] **Hailey Nguyen** [1] **Kartikeya Upasani** [1] **Daniel M. Bikel** [1] **Jason Weston** [1] _[∗]_ **Eric Michael Smith** [1] _[∗]_ 1 Meta 2 Carnegie Mellon University A BSTRACT Text generation has a fundamental limitation... | 2Direct Enhancement | Bo62NeU6VF |
# W ILD B ENCH : B ENCHMARKING LLM S WITH C HALLENGING T ASKS FROM R EAL U SERS IN THE W ILD **Bill Yuchen Lin** _[♡♢]_ **Yuntian Deng** _[♡]_ **Khyathi Chandu** _[♡]_ **Faeze Brahman** _[♡]_ **Abhilasha Ravichander** _[♡]_ **Valentina Pyatkin** _[♡]_ **Nouha Dziri** _[♡]_ **Ronan Le Bras** _[♡]_ **Yejin Choi** _[♡♢]_ ... | 3Other | MKEHCx25xp |
# H OLISTICALLY E VALUATING THE E NVIRONMENTAL I MPACT OF C REATING L ANGUAGE M ODELS **Jacob Morrison** **[1]** **Clara Na** **[2]** **Jared Fernandez** **[2]** **Tim Dettmers** **[1,2]** **Emma Strubell** **[1,2]** **Jesse Dodge** **[1]** 1 Allen Institute for AI 2 Carnegie Mellon University jacobm@allenai.org A BSTR... | 0Conceptual Integration | 04qx93Viwj |
# L ANGUAGE R EPRESENTATIONS C AN BE W HAT R ECOMMENDERS N EED : F INDINGS AND P OTENTIALS **Leheng Sheng** [1] **An Zhang** [1] _[∗]_ **Yi Zhang** [2] **Yuxin Chen** [1] **Xiang Wang** [2] **Tat-Seng Chua** [1] 1 National University of Singapore 2 University of Science and Technology of China leheng.sheng@u.nus.edu, a... | 0Conceptual Integration | eIJfOIMN9z |
# I NCLUDE : E VALUATING M ULTILINGUAL L ANGUAGE U NDERSTANDING WITH R EGIONAL K NOWLEDGE **Angelika Romanou** [1] _[,][∗]_ **, Negar Foroutan** [1] **, Anna Sotnikova** [1] **, Zeming Chen** [1] **,** **Sree Harsha Nelaturu** [3] _[,]_ [5] **, Shivalika Singh** [2] **, Rishabh Maheshwary** [7] **, Micol Altomare** [6]... | 3Other | k3gCieTXeY |
# D IFFERENTIABLE I NTEGER L INEAR P ROGRAMMING **Zijie Geng** **[1,2]** _[∗]_ **, Jie Wang** **[1,2]** _[†]_ **, Xijun Li** **[3]** **, Fangzhou Zhu** **[4]** **, Jianye Hao** **[4,5]** **, Bin Li** **[1]** **, Feng Wu** **[1,2]** 1 University of Science and Technology of China 2 MoE Key Laboratory of Brain-inspired I... | 0Conceptual Integration | FPfCUJTsCn |
# - P HYLO VAE: U NSUPERVISED L EARNING OF P HYLOGE NETIC T REES VIA V ARIATIONAL A UTOENCODERS **Tianyu Xie** [1] **, Harry Richman** [3] **, Jiansi Gao** [3] **, Frederick A. Matsen IV** [3] _[,]_ [4] **, Cheng Zhang** [1] _[,]_ [2] _[,][∗]_ 1 School of Mathematical Sciences, Peking University 2 Center for Statistica... | 2Direct Enhancement | Z8TglKXDWm |
# MAVIS: M ATHEMATICAL V ISUAL I NSTRUCTION T UNING WITH AN A UTOMATIC D ATA E NGINE **Renrui Zhang** [1] _[∗†]_ **, Xinyu Wei** [3] _[∗]_ **, Dongzhi Jiang** [1] **, Ziyu Guo** [2] **, Yichi Zhang** [3] **, Chengzhuo Tong** [4] **Jiaming Liu** [3], **Aojun Zhou** [1], **Shanghang Zhang** [3], **Peng Gao** [4], **Hongs... | 2Direct Enhancement | MnJzJ2gvuf |
# L O RA3D: L OW -R ANK S ELF -C ALIBRATION OF 3D G EOMETRIC F OUNDATION M ODELS **Ziqi Lu** **[1,2]** **, Heng Yang** **[1,3]** **, Danfei Xu** **[1,4]** **, Boyi Li** **[1,5]** **, Boris Ivanovic** **[1]** **, Marco Pavone** **[1,6]** **,** **Yue Wang** **[1,7]** 1 NVIDIA Research, 2 Massachusetts Institute of Techno... | 2Direct Enhancement | LSp4KBhAom |
# C ONTEXT C LUES : E VALUATING L ONG C ONTEXT M ODELS FOR C LINICAL P REDICTION T ASKS ON EHR S **Michael Wornow** _[∗]_ [1] **, Suhana Bedi** _[∗]_ [1] **, Miguel Angel Fuentes Hernandez** [1] **, Ethan Steinberg** [12] **,** **Jason Alan Fries** [1] **, Christopher R´e** [1] **, Sanmi Koyejo** [1] **, Nigam H. Shah*... | 1Cross-Domain Application | zg3ec1TdAP |
## - H UMAN ALIGNED C HESS WITH A B IT OF S EARCH **Yiming Zhang** [1] **Athul Paul Jacob** [2] **Vivian Lai** [3] **Daniel Fried** [1] **Daphne Ippolito** [1] 1 Carnegie Mellon University 2 MIT 3 Visa Research A BSTRACT Chess has long been a testbed for AI’s quest to match human intelligence, and in recent years, ches... | 0Conceptual Integration | bc2H72hGxB |
# P OINT OBB- V 2: T OWARDS S IMPLER, F ASTER, AND S TRONGER S INGLE P OINT S UPERVISED O RIENTED O BJECT D ETECTION **Botao Ren** [1] _[∗]_ **, Xue Yang** [2] _[∗]_ **, Yi Yu** [3] _[∗]_ **, Junwei Luo** [4] **, Zhidong Deng** [1] _[†]_ 1 BNRist, THUAI, Department of Computer Science, Tsinghua University 2 Department ... | 2Direct Enhancement | R22JPTQYWV |
# A CCESSING V ISION F OUNDATION M ODELS VIA I MAGE N ET -1K **Yitian Zhang** [1] **Xu Ma** [1] **Yue Bai** [1] **Huan Wang** [1] _[∗]_ **Yun Fu** [1] _[,]_ [2] 1 Department of Electrical and Computer Engineering, Northeastern University 2 Khoury College of Computer Science, Northeastern University {zhang.yitian, ma.xu... | 0Conceptual Integration | LC6ZtQV6u2 |
# - D YNAMIC G AUSSIANS M ESH : C ONSISTENT M ESH R E CONSTRUCTION F ROM D YNAMIC S CENES **Isabella Liu, Hao Su** **[†]** **, Xiaolong Wang** **[†]** UC San Diego, [†] Equal advising Figure 1: We propose DG-Mesh, a framework that reconstructs high-fidelity time-consistent mesh for dynamic scenes with complex non-rigid... | 0Conceptual Integration | LuGHbK8qTa |
# - R ETHINKING R EWARD M ODELING IN P REFERENCE BASED L ARGE L ANGUAGE M ODEL A LIGNMENT **Hao Sun** _[∗]_ **, Yunyi Shen** _[∗]_ **, Jean-Francois Ton** University of Cambridge, Massachusetts Institute of Technology, ByteDance Research A BSTRACT The Bradley-Terry (BT) model is a common and successful practice in rewa... | 0Conceptual Integration | rfdblE10qm |
# L EARNING TO D ISCRETIZE D ENOISING D IFFUSION ODE S **Vinh Tong** [1] _[,]_ [2] **, Trung-Dung Hoang** [4] **, Anji Liu** [1] _[,]_ [3] **, Guy Van den Broeck** [3] **, Mathias Niepert** [1] _[,]_ [2] 1 University of Stuttgart, 2 IMPRS-IS, 3 UCLA, 4 University of Bern vinh.tong@ki.uni-stuttgart.de A BSTRACT Diffusio... | 2Direct Enhancement | xDrFWUmCne |
# C IPHER P RUNE : E FFICIENT AND S CALABLE P RIVATE T RANSFORMER I NFERENCE Yancheng Zhang [1], Jiaqi Xue [1], Mengxin Zheng [1] Mimi Xie [2], Mingzhe Zhang [3], Lei Jiang [4], Qian Lou [1*] _1_ _University of Central Florida_ _2_ _University of Texas at San Antonio_ _3_ _Ant Research_ _4_ _Indiana University Blooming... | 0Conceptual Integration | mUMvr33FTu |
# - - D ISCRETIZATION INVARIANCE ? O N THE D ISCRETI ZATION M ISMATCH E RRORS IN N EURAL O PERATORS **Wenhan Gao** [1] **Ruichen Xu** [1] **Yuefan Deng** [1] **Yi Liu** [1] _[,]_ [2] _[∗]_ 1 Department of Applied Mathematics and Statistics, Stony Brook University 2 Department of Computer Science, Stony Brook University... | 0Conceptual Integration | J9FgrqOOni |
# - S IMPLE YET E FFECTIVE I NCOMPLETE M ULTI VIEW - C LUSTERING : S IMILARITY LEVEL I MPUTATION AND - - - I NTRA VIEW H YBRID GROUP P ROTOTYPE C ONSTRUC ## TION **Shengju Yu** [1] **, Zhibin Dong** [1] **, Siwei Wang** [2] _[∗]_ **, Pei Zhang** [1] **, Yi Zhang** [1] **, Xinwang Liu** [1] _[∗]_ **,** **Naiyang Guan** ... | 2Direct Enhancement | KijslFbfOL |
# S OFT CVI: C ONTRASTIVE VARIATIONAL INFERENCE ## WITH SELF GENERATED SOFT LABELS **Daniel Ward** [1] **, Mark Beaumont** [2] **, Matteo Fasiolo** [1] 1 School of Mathematics, Bristol University, UK 2 School of Biological Sciences, Bristol University, UK A BSTRACT Estimating a distribution given access to its unnormal... | 0Conceptual Integration | PiZtlzMWUj |
# C ROSS MPT: C ROSS -A TTENTION M ESSAGE -P ASSING T RANSFORMER FOR E RROR C ORRECTING C ODES **Seong-Joon Park** **[1,]** _[∗]_ **Hee-Youl Kwak** **[2,]** _[†]_ **Sang-Hyo Kim** **[3]** **Yongjune Kim** **[1,]** _[†]_ **Jong-Seon No** **[4]** 1 POSTECH, 2 University of Ulsan, 3 Sungkyunkwan University, 4 Seoul Nation... | 0Conceptual Integration | gFvRRCnQvX |
# - MMTEB: M ASSIVE M ULTILINGUAL T EXT E MBED DING B ENCHMARK **Kenneth Enevoldsen** **[1,†, ‡]** **, Isaac Chung** **[2,‡]** **, Imene Kerboua** **[3,4,‡]** **, Márton Kardos** **[1,‡]** **,** **Ashwin Mathur** **[2]**, **David Stap** **[5]**, **Jay Gala** **[6]**, **Wissam Siblini** **[2]**, **Dominik Krzemi´nski** ... | 3Other | zl3pfz4VCV |
# WEAK-TO-STRONG PREFERENCE O PTIMIZATION : S TEALING REWARD FROM WEAK ALIGNED MODEL **Wenhong Zhu** [1] _[,]_ [2] **, Zhiwei He** [1] **, Xiaofeng Wang** [1] **, Pengfei Liu** [1] _[,]_ [2] **, Rui Wang** [1] _[∗]_ 1 Shanghai Jiao Tong University, 2 Shanghai Innovation Institute _{_ zwhong714, wangrui12 _}_ @sjtu.edu.... | 0Conceptual Integration | f7KxfUrRSb |
### - E XPAND AND C OMPRESS : E XPLORING T UNING P RIN CIPLES FOR C ONTINUAL S PATIO -T EMPORAL G RAPH F ORECASTING **Wei Chen, Yuxuan Liang** [�] The Hong Kong University of Science and Technology (Guangzhou) onedeanxxx@gmail.com, yuxliang@outlook.com A BSTRACT The widespread deployment of sensing devices leads to a s... | 0Conceptual Integration | FRzCIlkM7I |
# M IN -K%++: I MPROVED B ASELINE FOR D ETECTING - P RE -T RAINING D ATA FROM L ARGE L ANGUAGE M OD ## ELS **Jingyang Zhang** [1] _[∗]_ **, Jingwei Sun** [1] **, Eric Yeats** _[∗]_ [1] **, Yang Ouyang** [1] **, Martin Kuo** [1] **,** **Jianyi Zhang** [1] **, Hao Frank Yang** [1] _[,]_ [2] **, Hai Li** [1] 1 Duke Univer... | 0Conceptual Integration | ZGkfoufDaU |
# - G AUSSIAN E NSEMBLE B ELIEF P ROPAGATION FOR E F - FICIENT I NFERENCE IN H IGH -D IMENSIONAL, B LACK BOX S YSTEMS **Dan MacKinlay** _[∗]_ CSIRO’s Data61 Dan.MacKinlay@data61.csiro.au **Russell Tsuchida** **Dan Pagendam** **Petra Kuhnert** Monash University _[†]_ CSIRO’s Data61 CSIRO’s Data61 A BSTRACT Efficient inf... | 0Conceptual Integration | PLskiLUBDW |
# C AT VTON: C ONCATENATION I S A LL Y OU N EED FOR V IRTUAL T RY -O N WITH D IFFUSION M ODELS **Zheng Chong** **[1,4,5]** **, Xiao Dong** **[2]** **, Haoxiang Li** **[3]** **, Shiyue Zhang** **[1]** **, Wenqing Zhang** **[1]** **, Xujie Zhang** **[1]** **,** **Hanqing Zhao** **[4,6]** **, Dongmei Jiang** **[4]** & **X... | 2Direct Enhancement | jt1h2dnmng |
# - T OWARDS U NDERSTANDING W HY F IX M ATCH G ENER ALIZES B ETTER T HAN S UPERVISED L EARNING **Jingyang Li** [1] **Jiachun Pan** [1] **Vincent Y. F. Tan** [1] **Kim-Chuan Toh** [1] **Pan Zhou** [2] _[∗]_ 1 National University of Singapore 2 Singapore Management University li_jingyang@u.nus.edu pan.jiachun@outlook.com... | 3Other | 25kAzqzTrz |
# REEF: R EPRESENTATION E NCODING F INGERPRINTS FOR L ARGE L ANGUAGE M ODELS **Jie Zhang** **[1,2]** _[⋆]_ **, Dongrui Liu** **[1]** _[⋆]_ **, Chen Qian** **[1,3]** **, Linfeng Zhang** **[4]** **, Yong Liu** **[3]** **, Yu Qiao** **[1]** **, Jing Shao** **[1]** _[†]_ 1 Shanghai Artificial Intelligence Laboratory 2 Univ... | 1Cross-Domain Application | SnDmPkOJ0T |
# - M ESH A NYTHING : A RTIST -C REATED M ESH G ENERA TION WITH A UTOREGRESSIVE T RANSFORMERS **Yiwen Chen** [1] _[,]_ [2] _[∗]_ **,** **Tong He** [2] _[†]_ **,** **Di Huang** [2] **,** **Weicai Ye** [2] **,** **Sijin Chen** [3] **,** **Jiaxiang Tang** [4] **Xin Chen** [5] **,** **Zhongang Cai** [6] **,** **Lei Yang** ... | 2Direct Enhancement | KGZAs8VcOM |
# O PTION Z ERO : P LANNING WITH L EARNED O PTIONS **Po-Wei Huang** **[1,2]** **, Pei-Chiun Peng** **[1,2]** **, Hung Guei** **[1]** **, Ti-Rong Wu** **[1]** _[†]_ 1 Institute of Information Science, Academia Sinica, Taiwan 2 Department of Computer Science, National Yang Ming Chiao Tung University, Taiwan A BSTRACT Pla... | 2Direct Enhancement | 3IFRygQKGL |
### - D IFFUSION -NPO: N EGATIVE P REFERENCE O PTI - MIZATION FOR B ETTER P REFERENCE A LIGNED G EN ERATION OF D IFFUSION M ODELS **Fu-Yun Wang** [1] **Yunhao Shui** [2] **Jingtan Piao** [1] **Keqiang Sun** [1] **Hongsheng Li** [1] _[,]_ [3] 1 MMLab, CUHK, Hong Kong 2 Shanghai Jiang Tong University, Shanghai 3 CPII und... | 2Direct Enhancement | iJi7nz5Cxc |
# - P ERTURBO LL A VA: R EDUCING M ULTIMODAL H ALLU CINATIONS WITH P ERTURBATIVE V ISUAL T RAINING **Cong Chen** [1] _[,]_ _[∗]_ **,** **Mingyu Liu** [1] _[,]_ _[∗]_ **,** **Chenchen Jing** [3] **,** **Yizhou Zhou** [2] **,** **Fengyun Rao** [2], **Hao Chen** [1] **,** **Bo Zhang** [1] **,** **Chunhua Shen** [3] _[,]_ ... | 0Conceptual Integration | j4LITBSUjs |
## H OMOMORPHISM C OUNTS AS S TRUCTURAL E NCODINGS FOR G RAPH L EARNING **Linus Bao** _[∗]_ **Emily Jin** _[∗]_ **Michael Bronstein** University of Oxford University of Oxford University of Oxford / AITHYRA **Ismail** ˙ ˙ **Ilkan Ceylan** **Matthias Lanzinger** University of Oxford TU Wien A BSTRACT Graph Transformers ... | 0Conceptual Integration | qFw2RFJS5g |
# - - F LOW D EC : A FLOW BASED FULL BAND GENERAL ## AUDIO CODEC WITH HIGH PERCEPTUAL QUALITY **Simon Welker** _[♯,][∗]_ **, Matthew Le** _[⋄]_ **, Ricky T.Q. Chen** _[⋄]_ **, Wei-Ning Hsu** _[⋄]_ **,** **Timo Gerkmann** _[♯]_ **, Alexander Richard** _[⋄]_ **, Yi-Chiao Wu** _[⋄]_ _♯_ Signal Processing, University of Ha... | 0Conceptual Integration | uxDFlPGRLX |
# - HELMET: H OW TO E VALUATE L ONG -C ONTEXT L AN GUAGE M ODELS E FFECTIVELY AND T HOROUGHLY **Howard Yen** _**[p]**_ **Tianyu Gao** _**[p]**_ **Minmin Hou** _**[i]**_ **Ke Ding** _**[i]**_ **Daniel Fleischer** _**[i]**_ **Peter Izsak** _**[i]**_ **Moshe Wasserblat** _**[i]**_ **Danqi Chen** _**[p]**_ _**p**_ Princeto... | 3Other | 293V3bJbmE |
# - W HAT IS W RONG WITH P ERPLEXITY FOR L ONG CONTEXT L ANGUAGE M ODELING ? **Lizhe Fang** [1] _[∗]_ **Yifei Wang** [2] _[∗]_ **Zhaoyang Liu** [3] **Chenheng Zhang** [1] **Stefanie Jegelka** [4] _[,]_ [5] **Jinyang Gao** [3] **Bolin Ding** [3] **Yisen Wang** [1] _[,]_ [6] _[†]_ 1 State Key Lab of General Artificial In... | 2Direct Enhancement | fL4qWkSmtM |
# - F INE TUNING WITH R ESERVED M AJORITY FOR N OISE R EDUCTION **Shuyang Jiang** _[♠][,][♣]_ **,Yusheng Liao** _[♢][,][♣]_ **,Ya Zhang** _[♢][,][♣][,][∗]_ **, Yanfeng Wang** _[♢][,][♣]_ **, Yu Wang** _[♢][,][♣][,][∗]_ _♠_ Fudan University _♢_ School of Artificial Intelligence, Shanghai Jiao Tong University _♣_ Shangha... | 2Direct Enhancement | ZV7CLf0RHK |
# C AN A MISL F LY ? A NALYSIS AND I NGREDIENTS FOR M UTUAL I NFORMATION S KILL L EARNING **Chongyi Zheng, Jens Tuyls** _[∗]_ _[∗]_ **, Joanne Peng & Benjamin Eysenbach** Department of Computer Science Princeton University _{_ chongyiz, jtuyls _}_ @cs.princeton.edu A BSTRACT Self-supervised learning has the potential o... | 0Conceptual Integration | xoIeVdFO7U |
# - - E FFICIENT R ESIDUAL L EARNING WITH M IXTURE OF E XPERTS FOR U NIVERSAL D EXTEROUS G RASPING **Ziye Huang** [2] **, Haoqi Yuan** [1] **, Yuhui Fu** [1] **, Zongqing Lu** [1] _[,]_ [3] _[†]_ 1 School of Computer Science, Peking University 2 School of EECS, Peking University 3 Beijing Academy of Artificial Intellig... | 0Conceptual Integration | BUj9VSCoET |
# E XPLORATORY P REFERENCE O PTIMIZATION : H ARNESSING I MPLICIT Q*-A PPROXIMATION FOR S AMPLE -E FFICIENT RLHF **Tengyang Xie** _[∗]_ UW-Madison [tx@cs.wisc.edu](mailto:tx@cs.wisc.edu) **Dylan J. Foster** _[∗]_ Microsoft Research [dylanfoster@microsoft.com](mailto:dylanfoster@microsoft.com) **Akshay Krishnamurthy** Mi... | 0Conceptual Integration | QYigQ6gXNw |
# D E LLM A : D ECISION M AKING U NDER U NCERTAINTY ## WITH L ARGE L ANGUAGE M ODELS **Ollie Liu** _[∗]_ **, Deqing Fu** _[∗]_ **, Dani Yogatama, Willie Neiswanger** Thomas Lord Department of Computer Science University of Southern California me@ollieliu.com, {deqingfu, yogatama, neiswang}@usc.edu A BSTRACT The potenti... | 0Conceptual Integration | Acvo2RGSCy |
# G RAPH B RIDGE : T OWARDS A RBITRARY T RANSFER L EARNING IN GNN S **Li Ju, Xingyi Yang, Qi Li, Xinchao Wang** _[∗]_ National University of Singapore _{_ jujulili,xyang,liqi _}_ @u.nus.edu, xinchao@nus.edu.sg A BSTRACT Graph neural networks (GNNs) are conventionally trained on a per-domain, pertask basis. It creates a... | 0Conceptual Integration | gjRhw5S3A4 |
# L EARNING LOCAL EQUIVARIANT REPRESENTATIONS ## FOR QUANTUM OPERATORS **Zhanghao Zhouyin** [2] _[,]_ [3] **Zixi Gan** [2] _[,]_ [4] **Shishir K. Pandey** [5] **Linfeng Zhang** [2] _[,]_ [6] **Qiangqiang Gu** [1] _[,]_ [2] _[,]_ [7] _[ ∗]_ 1 School of Artificial Intelligence and Data Science, University of Science and ... | 0Conceptual Integration | kpq3IIjUD3 |
# FlexPrefill : A C ONTEXT -A WARE S PARSE A TTENTION - M ECHANISM FOR E FFICIENT L ONG -S EQUENCE I N ## FERENCE **Xunhao Lai** [1] _[∗]_ **, Jianqiao Lu** [2] _[∗]_ **, Yao Luo** [3] **, Yiyuan Ma** [3] **, Xun Zhou** [3] **,** 1 Peking University 2 The University of Hong Kong 3 ByteDance Inc laixunhao@pku.edu.cn,jql... | 2Direct Enhancement | OfjIlbelrT |
# U NDERSTANDING AND M ITIGATING H ALLUCINATION - IN L ARGE V ISION -L ANGUAGE M ODELS VIA M ODU LAR A TTRIBUTION AND I NTERVENTION **Tianyun Yang** [1] _[,]_ [2] **Ziniu Li** [3] **Juan Cao** [1] _[,]_ [2] _[∗]_ **Chang Xu** [4] _[∗]_ 1 Institute of Computing Technology, Chinese Academy of Sciences, Beijing, China 2 U... | 0Conceptual Integration | Bjq4W7P2Us |
# - F AKE S HIELD : E XPLAINABLE I MAGE F ORGERY D E - TECTION AND L OCALIZATION VIA M ULTI MODAL L ARGE L ANGUAGE M ODELS **Zhipei Xu** [1] _[,]_ [2] **, Xuanyu Zhang** [1] **, Runyi Li** [1] **, Zecheng Tang** [1] **, Qing Huang** [4] **, Jian Zhang** [1] _[,]_ [2] _[,]_ [3] _[†]_ 1 School of Electronic and Computer ... | 0Conceptual Integration | pAQzEY7M03 |
# - VILA-U: A U NIFIED F OUNDATION M ODEL I NTEGRAT ING V ISUAL U NDERSTANDING AND G ENERATION **Yecheng Wu** [1] _[,]_ [2] _[∗]_ **Zhuoyang Zhang** [2] _[∗†]_ **Junyu Chen** [1] _[,]_ [2] **Haotian Tang** [2] _[†]_ **Dacheng Li** [4] _[†]_ **Yunhao Fang** [5] _[†]_ **Ligeng Zhu** [3] **Enze Xie** [3] **Hongxu Yin** [3... | 0Conceptual Integration | 02haSpO453 |
# P HYSICS OF L ANGUAGE M ODELS : P ART 2.2, H OW TO L EARN F ROM M ISTAKES ON G RADE -S CHOOL M ATH P ROBLEMS [ EXTENDED ABSTRACT ] _[∗]_ **Tian Ye** FAIR at Meta and Carnegie Mellon University sigma648@meta.com and tye2@andrew.cmu.edu **Zicheng Xu** FAIR at Meta (now Google Research) zichengxu@meta.com (now zichengxu... | 0Conceptual Integration | zpDGwcmMV4 |
# M ONITORING L ATENT W ORLD S TATES IN L ANGUAGE M ODELS WITH P ROPOSITIONAL P ROBES **Jiahai Feng, Stuart Russell & Jacob Steinhardt** _[∗]_ UC Berkeley A BSTRACT Language models (LMs) are susceptible to bias, sycophancy, backdoors, and other tendencies that lead to unfaithful responses to the input context. Interpre... | 0Conceptual Integration | 0yvZm2AjUr |
#### E LICITING H UMAN P REFERENCES WITH L ANGUAGE M ODELS **Belinda Z. Li** _[∗]_ MIT CSAIL bzl@mit.edu **Alex Tamkin** _[∗]_ Anthropic _[†]_ atamkin@cs.stanford.edu **Noah D. Goodman** Stanford ngoodman@stanford.edu **Jacob Andreas** MIT CSAIL jda@mit.edu A BSTRACT Language models (LMs) can be directed to perform use... | 0Conceptual Integration | LvDwwAgMEW |
# - V IDEO G RAIN : M ODULATING S PACE -T IME A TTEN - TION FOR M ULTI GRAINED V IDEO E DITING **Xiangpeng Yang** [1] **Linchao Zhu** [2] **Hehe Fan** [2] **Yi Yang** [2] _[∗]_ 1 ReLER Lab, AAII, University of Technology Sydney 2 ReLER Lab, CCAI, Zhejiang University [Project Page: https://knightyxp.github.io/VideoGrain... | 0Conceptual Integration | SSslAtcPB6 |
# A C OEFFICIENT M AKES SVRG E FFECTIVE **Yida Yin** [1] **Zhiqiu Xu** [2] **Zhiyuan Li** [3] **Trevor Darrell** [1] **Zhuang Liu** [4] 1 UC Berkeley 2 University of Pennsylvania 3 TTIC 4 Meta AI Research A BSTRACT Stochastic Variance Reduced Gradient (SVRG), introduced by Johnson & Zhang (2013), is a theoretically com... | 2Direct Enhancement | twtTLZnG0B |
# P ROBABILISTIC N EURAL P RUNING VIA S PARSITY E VOLUTIONARY F OKKER -P LANCK -K OLMOGOROV E QUATION **Zhanfeng Mo** _[†]_ **Haosen Shi** _[‡]_ **Sinno Jialin Pan** _[†‡]_ _†_ Nanyang Technological University, Singapore _‡_ The Chinese University of Hong Kong zhanfeng001@ntu.edu.sg; haosen.shi.ai@link.cuhk.edu.hk; sin... | 0Conceptual Integration | hJ1BaJ5ELp |
# - - I N VIVO CELL TYPE AND BRAIN REGION CLASSIFICA ## TION VIA MULTIMODAL CONTRASTIVE LEARNING **Han Yu** **[1]** **, Hanrui Lyu** **[2]** **, Ethan Yixun Xu** **[1]** **, Charlie Windolf** **[1]** **, Eric Kenji Lee** **[3]** **, Fan Yang** **[4]** **,** **Andrew M. Shelton** **[5]**, **Olivier Winter** **[6]**, **I... | 1Cross-Domain Application | 10JOlFIPjt |
# D ATA S CALING L AWS IN I MITATION L EARNING FOR R OBOTIC M ANIPULATION **Fanqi Lin** [1] _[,]_ [2] _[,]_ [3] _[∗]_ **Yingdong Hu** [1] _[,]_ [2] _[,]_ [3] _[∗]_ **Pingyue Sheng** [1] **Chuan Wen** [1] _[,]_ [2] _[,]_ [3] **Jiacheng You** [1] **Yang Gao** [1] _[,]_ [2] _[,]_ [3] 1 Tsinghua University, 2 Shanghai Qi Z... | 1Cross-Domain Application | pISLZG7ktL |
# G E S UB N ET : G ENE I NTERACTION I NFERENCE FOR D ISEASE S UBTYPE N ETWORK G ENERATION **Ziwei Yang** **[1,2]** **, Zheng Chen** **[2]** **, Xin Liu** _[∗]_ **[3*]** **, Rikuto Kotoge** **[2]** **, Peng Chen** **[4]** **, Yasuko Matsubara** **[2]** **,** **Yasushi Sakurai** **[2]** **, Jimeng Sun** **[5]** 1 Bioinf... | 0Conceptual Integration | ja4rpheN2n |
# S CALABLE AND C ERTIFIABLE G RAPH U NLEARNING : O VERCOMING THE A PPROXIMATION E RROR B ARRIER **Lu Yi & Zhewei Wei** _[∗]_ Renmin University of China _{_ yilu, zhewei _}_ @ruc.edu.cn A BSTRACT Graph unlearning has emerged as a pivotal research area for ensuring privacy protection, given the widespread adoption of Gr... | 0Conceptual Integration | pPyJyeLriR |
# R E D E EP: D ETECTING H ALLUCINATION IN R ETRIEVAL -A UGMENTED G ENERATION VIA M ECHANISTIC I NTERPRETABILITY **Zhongxiang Sun** **[1]** **, Xiaoxue Zang** **[2]** **, Kai Zheng** **[2]** **, Jun Xu** **[1]** _[∗]_ **, Xiao Zhang** **[1]** **, Weijie Yu** **[3]** **, Yang Song** **[2]** **, Han Li** **[2]** 1 Gaolin... | 0Conceptual Integration | ztzZDzgfrh |
# D ECOUPLING L AYOUT FROM G LYPH IN O NLINE C HINESE H ANDWRITING G ENERATION **Min-Si Ren** [1] _[,]_ [2] **, Yan-Ming Zhang** [1] _[,]_ [2] _[ ∗]_ **, Yi Chen** [1] _[,]_ [2] 1 School of Artificial Intelligence, University of Chinese Academy of Sciences, Beijing 100049, China 2 State Key Laboratory of Multimodal Art... | 0Conceptual Integration | DhHIw9Nbl1 |
# MMF AKE B ENCH : A M IXED -S OURCE M ULTIMODAL M ISINFORMATION D ETECTION B ENCHMARK FOR LVLM S **Xuannan Liu** [1] **Zekun Li** [2] **Peipei Li** [1] _[∗]_ **Huaibo Huang** [3] **Shuhan Xia** [1] **Xing Cui** [1] **Linzhi Huang** [1] **Weihong Deng** [1] **Zhaofeng He** [1] 1 Beijing University of Posts and Telecomm... | 3Other | D6zn6ozJs7 |
# M ECHANISTIC P ERMUTABILITY : M ATCH F EATURES A CROSS L AYERS **Nikita Balagansky** **[1,2]** _[∗]_ **, Ian Maksimov** **[3,1]** **, Daniil Gavrilov** **[1]** 1 T-Tech, 2 Moscow Institute of Physics and Technologies, 3 HSE University A BSTRACT Understanding how features evolve across layers in deep neural networks i... | 0Conceptual Integration | MDvecs7EvO |
# A CLIP-P OWERED F RAMEWORK FOR R OBUST AND G ENERALIZABLE D ATA S ELECTION **Suorong Yang** **[1,2*]** **, Peng Ye** **[2,3]** **, Wanli Ouyang** **[2]** **, Dongzhan Zhou** **[2 †]** **, Furao Shen** **[1 †]** 1 National Key Laboratory for Novel Software Technology, Nanjing University 2 Shanghai Artificial Intellige... | 0Conceptual Integration | 9bMZ29SPVx |
# - S HOW O : O NE S INGLE T RANSFORMER TO U NIFY M ULTIMODAL U NDERSTANDING AND G ENERATION **Jinheng Xie** [1] _[†]_ **Weijia Mao** [1] _[†]_ **Zechen Bai** [1] _[†]_ **David Junhao Zhang** [1] _[†]_ **Weihao Wang** [2] **Kevin Qinghong Lin** [1] **Yuchao Gu** [1] **Zhijie Chen** [2] **Zhenheng Yang** [2] **Mike Zhen... | 0Conceptual Integration | o6Ynz6OIQ6 |
# F AST AND A CCURATE B LIND F LEXIBLE D OCKING **Zizhuo Zhang** [1] **Lijun Wu** [2] _[†]_ **Kaiyuan Gao** [3] **Jiangchao Yao** [4] **Tao Qin** [5] **Bo Han** [1] _[†]_ 1 TMLR Group, Department of Computer Science, Hong Kong Baptist University 2 Shanghai AI Laboratory 3 Huazhong University of Science and Technology 4... | 0Conceptual Integration | iezDdA9oeB |
# C HEATING A UTOMATIC LLM B ENCHMARKS : N ULL M ODELS A CHIEVE H IGH W IN R ATES **Xiaosen Zheng** _[∗]_ [1] _[,]_ [2] **, Tianyu Pang** _[∗†]_ [1] **, Chao Du** [1] **, Qian Liu** [1] **, Jing Jiang** _[†]_ [2] _[,]_ [3] **, Min Lin** [1] 1 Sea AI Lab, Singapore 2 Singapore Management University 3 Australian National... | 3Other | syThiTmWWm |
# L EARNING G ENERAL -P URPOSE B IOMEDICAL V OLUME R EPRESENTATIONS USING R ANDOMIZED S YNTHESIS **Neel Dey** [1] **Benjamin Billot** [1] **Hallee E. Wong** [1] **Clinton J. Wang** [1] **Mengwei Ren** [2] **P. Ellen Grant** [3] **Adrian V. Dalca** [1, 3] **Polina Golland** [1] 1 MIT CSAIL 2 New York University 3 Harvar... | 1Cross-Domain Application | xOmC5LiVuN |
# F ROM S PARSE D EPENDENCE TO S PARSE A TTENTION : - U NVEILING H OW C HAIN OF -T HOUGHT E NHANCES T RANSFORMER S AMPLE E FFICIENCY **Kaiyue Wen** _[∗†]_ Stanford University kaiyuew@stanford.edu **Hongzhou Lin** _[‡]_ Amazon hongzhou.lin89@gmail.com **Huaqing Zhang** **[*]** IIIS, Tsinghua University zhanghq22@mails.t... | 0Conceptual Integration | AmEgWDhmTr |
# - M ETA -D YNAMICAL S TATE S PACE M ODELS FOR I NTE GRATIVE N EURAL D ATA A NALYSIS **Ayesha Vermani** [1] **, Josue Nassar** [2] **, Hyungju Jeon** [1] **, Matthew Dowling** [1] **, Il Memming Park** [1] 1 Champalimaud Centre for the Unknown, Champalimaud Foundation, Portugal 2 RyvivyR, USA {ayesha.vermani, memming.... | 0Conceptual Integration | SRpq5OBpED |
# - M ITIGATING I NFORMATION L OSS IN T REE -B ASED R E INFORCEMENT L EARNING VIA D IRECT O PTIMIZATION **Sascha Marton** [1] _[∗]_ **Tim Grams** [2] _[∗]_ **Florian Vogt** [1] **Stefan L¨udtke** [3] **Christian Bartelt** [2] **Heiner Stuckenschmidt** [1] 1 University of Mannheim 2 Technical University of Clausthal 3 U... | 0Conceptual Integration | qpXctF2aLZ |
# F AIR MT-B ENCH : B ENCHMARKING F AIRNESS FOR - M ULTI TURN D IALOGUE IN C ONVERSATIONAL LLM S **Zhiting Fan** _[†]_ **Ruizhe Chen** _[†]_ **Tianxiang Hu** _[†]_ **Zuozhu Liu** _[∗]_ Zhejiang University A BSTRACT The increasing deployment of large language model (LLM)-based chatbots has raised concerns regarding fair... | 3Other | RSGoXnS9GH |
# U NI CO: O N U NIFIED C OMBINATORIAL O PTIMIZATION - VIA P ROBLEM R EDUCTION TO M ATRIX -E NCODED G EN ERAL TSP **Wenzheng Pan** _[†]_ **, Hao Xiong** _[†]_ **, Jiale Ma, Wentao Zhao, Yang Li, Junchi Yan** _[∗]_ Sch. of Computer Science & Sch. of Artificial Intelligence, Shanghai Jiao Tong University _{_ pwz1121,taxu... | 3Other | yEwakMNIex |
# U NI M ATCH : U NIVERSAL M ATCHING FROM A TOM TO T ASK FOR F EW -S HOT D RUG D ISCOVERY **Ruifeng Li** [1,2] **, Mingqian Li** [2] **, Wei Liu** [3] **, Yuhua Zhou** [1,2] **, Xiangxin Zhou** [4] **,** **Yuan Yao** [1,5] **, Qiang Zhang** [1,5] _[†]_ **, Hongyang Chen** [2] _[†]_ 1 Zhejiang University 2 Zhejiang Lab ... | 0Conceptual Integration | v9EjwMM55Y |
# L O RA-P RO : A RE L OW -R ANK A DAPTERS P ROPERLY O PTIMIZED ? **Zhengbo Wang** [1] _[,]_ [2] **Jian Liang** [2] _[,]_ [3] _[∗]_ **Ran He** [2] _[,]_ [3] **Zilei Wang** [1] **Tieniu Tan** [2] _[,]_ [4] 1 University of Science and Technology of China 2 NLPR & MAIS, Institute of Automation, Chinese Academy of Sciences... | 2Direct Enhancement | gTwRMU3lJ5 |
# I NFINITE -R ESOLUTION I NTEGRAL N OISE W ARPING FOR D IFFUSION M ODELS **Yitong Deng** [1] _[,]_ [2] **, Winnie Lin** [1] **, Lingxiao Li** [1] **, Dmitriy Smirnov** [1] **, Ryan Burgert** [3] _[,]_ [4] **, Ning Yu** [3] **,** **Vincent Dedun** [1] **, Mohammad H. Taghavi** [1] 1 Netflix, 2 Stanford University, 3 Ne... | 2Direct Enhancement | Y6LPWBo2HP |
# R ETRI 3D: 3D N EURAL G RAPHICS R EPRESENTATION R ETRIEVAL **Yushi Guan** [1] **, Daniel Kwan** [1] **, Jean Sebastien Dandurand** [1] **, Xi Yan** [1] **, Ruofan Liang** [1] **,** **Yuxuan Zhang** [1] **, Nilesh Jain** [2] **, Nilesh Ahuja** [2] **, Selvakumar Panneer** [2] **, Nandita Vijaykumar** [1] {guanyushi, d... | 0Conceptual Integration | q3EbOXb4y1 |
# - P OSTERIOR -M EAN R ECTIFIED F LOW : T OWARDS M IN IMUM MSE P HOTO -R EALISTIC I MAGE R ESTORATION **Guy Ohayon, Tomer Michaeli, Michael Elad** Technion – Israel Institute of Technology _{_ ohayonguy@cs,tomer.m@ee,elad@cs _}_ .technion.ac.il A BSTRACT Photo-realistic image restoration algorithms are typically evalu... | 0Conceptual Integration | hPOt3yUXii |
# - L ONG VILA: S CALING L ONG -C ONTEXT V ISUAL L AN GUAGE M ODELS FOR L ONG V IDEOS **Yukang Chen** [1] _[∗]_ **Fuzhao Xue** [1] _[∗]_ **Dacheng Li** [3] _[†]_ **Qinghao Hu** [2] _[†]_ **Ligeng Zhu** [1] **Xiuyu Li** [3] **Yunhao Fang** [1] **Haotian Tang** [1] _[,]_ [2] **Shang Yang** [2] **Zhijian Liu** [1] **Ethan... | 0Conceptual Integration | wCXAlfvCy6 |
# A LPHA E DIT : N ULL -S PACE C ONSTRAINED K NOWLEDGE E DITING FOR L ANGUAGE M ODELS **Junfeng Fang** [1] _[∗]_ **, Houcheng Jiang** [1] _[ ∗]_ **, Kun Wang** [1] **, Yunshan Ma** [2] **,** **Jie Shi** [2] **, Xiang Wang** [1] _[†]_ **, Xiangnan He** [1] _[†]_ **, Tat-Seng Chua** [2] 1 University of Science and Techno... | 0Conceptual Integration | HvSytvg3Jh |
# V ISUAL P REDICATOR : L EARNING A BSTRACT W ORLD M ODELS WITH N EURO -S YMBOLIC P REDICATES FOR R OBOT P LANNING **Yichao Liang** [1] **, Nishanth Kumar** [3] **, Hao Tang** [2] **, Adrian Weller** [1] _[,]_ [6] **, Joshua B. Tenenbaum** [3] **,** **Tom Silver** [4] **, Jo˜ao F. Henriques** [5] **, Kevin Ellis** [2] ... | 0Conceptual Integration | QOfswj7hij |
### - E NHANCING L EARNING WITH L ABEL L OCAL D IFFER ENTIAL P RIVACY BY V ECTOR A PPROXIMATION **Puning Zhao, Jiafei Wu, Zhe Liu** _[∗]_ Zhejiang Lab Hangzhou, Zhejiang, China _{_ pnzhao, wujiafei, zhe.liu _}_ @zhejianglab.com **Li Shen** Shenzhen Campus of Sun Yat-sen University Shenzhen, Guangdong, China mathshenli@... | 1Cross-Domain Application | IwPXYk6BV9 |
# T RANSFORMERS P ROVABLY S OLVE P ARITY E FFICIENTLY WITH C HAIN OF T HOUGHT **Juno Kim** **[1,2]** _[∗]_ **Taiji Suzuki** **[1,2]** 1 Department of Mathematical Informatics, University of Tokyo 2 Center for Advanced Intelligence Project, RIKEN _∗_ `junokim@berkeley.edu` A BSTRACT This work provides the first theoreti... | 3Other | n2NidsYDop |
# C HART M IMIC : E VALUATING LMM’ S C ROSS -M ODAL - - R EASONING C APABILITY VIA C HART TO -C ODE G EN ## ERATION **Cheng Yang** [1] _[∗]_ **Chufan Shi** [1] _[∗]_ **Yaxin Liu** [1] _[∗]_ **Bo Shui** [1] _[∗]_ **Junjie Wang** [1] _[∗]_ **Mohan Jing** [1] **Linran Xu** [1] **Xinyu Zhu** [1] **Siheng Li** [1] **Yuxiang... | 3Other | sGpCzsfd1K |
# S ELF -S UPERVISED D IFFUSION M ODELS FOR E LECTRON -A WARE M OLECULAR R EPRESENTATION L EARNING **Gyoung S. Na** KRICT, Republic of Korea ngs0@krict.re.kr **Chanyoung Park** KAIST, Republic of Korea cy.park@kaist.ac.kr A BSTRACT Physical properties derived from electronic distributions are essential information that... | 0Conceptual Integration | UQ0RqfhgCk |
# FLIP: F LOW -C ENTRIC G ENERATIVE P LANNING AS G ENERAL -P URPOSE M ANIPULATION W ORLD M ODEL **Chongkai Gao** National University of Singapore gaochongkai@u.nus.edu **Zhixuan Xu, Zhehao Cai** National University of Singapore _{_ zhixuanxu,e1373791 _}_ @u.nus.edu **Haozhuo Zhang** Peking University 2100013132@stu.pku... | 0Conceptual Integration | B2N0nCVC91 |
# N OT A LL LLM-G ENERATED D ATA A RE E QUAL : - R ETHINKING D ATA W EIGHTING IN T EXT C LASSIFICA ## TION **Hsun-Yu Kuo** _[∗]_ [1] _[,]_ [2] _[,]_ [3] _[,][†]_ **, Yin-Hsiang Liao** _[∗]_ [1] _[,]_ [2] _[,][†]_ **, Yu-Chieh Chao** [1] **, Wei-Yun Ma** [1] _[,][†][,][‡]_ **, Pu-Jen Cheng** [2] _[,][†]_ 1 Academia Sini... | 0Conceptual Integration | oI5tZaWkF9 |
# - L OOKING B ACKWARD : S TREAMING V IDEO TO -V IDEO T RANSLATION WITH F EATURE B ANKS **Feng Liang** [1] **Akio Kodaira** [2] **Chenfeng Xu** [2] **Masayoshi Tomizuka** [2] **Kurt Keutzer** [2] **Diana Marculescu** [1] 1 UT Austin 2 UC Berkeley _{_ jeffliang, dianam _}_ @utexas.edu, _{_ akio.kodaira, xuchenfeng _}_ @... | 0Conceptual Integration | AMkf7h7HER |
# K NOWLEDGE E NTROPY D ECAY DURING L ANGUAGE M ODEL P RETRAINING H INDERS N EW K NOWLEDGE A CQUISITION **Jiyeon Kim** [1] _[∗]_ **Hyunji Lee** [1] _[∗]_ **Hyowon Cho** [1] **Joel Jang** [2] **Hyeonbin Hwang** [1] **Seungpil Won** [3] **Youbin Ahn** [3] **Dohaeng Lee** [3] **Minjoon Seo** [1] 1 KAIST AI 2 University of... | 0Conceptual Integration | eHehzSDUFp |
# - C ONTEXT AWARE D YNAMIC P RUNING FOR S PEECH F OUNDATION M ODELS **Masao Someki, Yifan Peng, Siddhant Arora, Shinji Watanabe** Language Technologies Institute Carnegie Mellon University Pittsburgh, PA 15213, USA _{_ msomeki,yifanpen,siddhana _}_ @andrew.cmu.edu, shinjiw@ieee.org, **Markus M¨uller, Thanasis Mouchtar... | 2Direct Enhancement | u2QdCiOgwA |
# - E FFICIENT AND C ONTEXT -A WARE L ABEL P ROPAGA - TION FOR Z ERO -/F EW -S HOT T RAINING -F REE A DAP TATION OF V ISION -L ANGUAGE M ODEL **Yushu Li** [1] _[,]_ [2] _[,]_ [4] _[∗]_ **, Yongyi Su** [1] _[,]_ [2] _[∗]_ **, Adam Goodge** [2] **, Kui Jia** [3] **, Xun Xu** [2] _[†]_ 1 South China University of Technolo... | 0Conceptual Integration | D10yarGQNk |
# O VERCOMING F ALSE I LLUSIONS IN R EAL -W ORLD F ACE R ESTORATION WITH M ULTI -M ODAL G UIDED D IFFUSION M ODEL **Keda Tao** **[1]** **,** **Jinjin Gu** **[2]** **,** _[∗]_ **Yulun Zhang** **[3]** **,** **Xiucheng Wang** **[1]** **,** **Nan Cheng** **[1, 4]** _[∗]_ 1 Xidian University, 2 The University of Sydney, 3 S... | 0Conceptual Integration | m9RNBZewW2 |
# - L EARNING STOCHASTIC DYNAMICS FROM SNAP ## SHOTS THROUGH REGULARIZED UNBALANCED OPTI MAL TRANSPORT **Zhenyi Zhang** [1] **, Tiejun Li** [1] _[,]_ [2] _[,][∗]_ **, and Peijie Zhou** [2] _[,]_ [3] _[,]_ [4] _[,]_ [5] _[,][∗]_ 1 LMAM and School of Mathematical Sciences, Peking University. 2 Center for Machine Learning... | 0Conceptual Integration | gQlxd3Mtru |
# - F LICKER F USION : I NTRA TRAJECTORY D OMAIN G ENERALIZING M ULTI -A GENT RL **Woosung Koh** **[1]** **, Wonbeen Oh** _[∗]_ **[1]** **, Siyeol Kim** **[1]** **, Suhin Shin** **[1]** **, Hyeongjin Kim** **[1]** **, Jaein Jang** **[1]** **,** **Junghyun Lee** **[2]** **, Se-Young Yun** **[2]** _[†]_ 1 Yonsei Universi... | 1Cross-Domain Application | MRYyOaNxh3 |
## - I NTERPRETING E MERGENT P LANNING IN M ODEL F REE R EINFORCEMENT L EARNING **Thomas Bush** [1] **, Stephen Chung** [1] **[†]** **, Usman Anwar** [1] **[†]** **, Adri`a Garriga-Alonso** [2] **, David Krueger** [3] 1 University of Cambridge, 2 FAR AI, 3 Mila, University of Montreal 28tbush@gmail.com, _{_ mhc48,ua237... | 0Conceptual Integration | DzGe40glxs |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.