id stringclasses 2
values | source stringclasses 1
value | type stringclasses 1
value | title stringclasses 2
values | content stringclasses 2
values | keywords listlengths 3 3 | doi stringclasses 1
value | language stringclasses 1
value |
|---|---|---|---|---|---|---|---|
p4_001 | paper_4 | invariant | Invariant System-Law Architecture | ARAYUN_173 defines an invariant architecture ensuring non-bypassable coherence across all system layers. | [
"invariance",
"architecture",
"AGI"
] | 10.5281/zenodo.18179361 | en |
p4_002 | paper_4 | invariant | Non-Bypassability Principle | The system law cannot be bypassed without collapsing symbolic and causal coherence. | [
"non-bypassable",
"system law",
"coherence"
] | 10.5281/zenodo.18179361 | en |
ARAYUN_173 Dataset
This dataset provides structured, machine-readable representations of the ARAYUN_173 research series.
Structure
Each record contains:
- id
- source
- type
- title
- content
- keywords
- doi
- language
Purpose
This dataset represents Paper 4 of the ARAYUN_173 research series in structured, machine-readable form. It corresponds to:
ARAYUN_173: Invariance Technology and Invariant System-Law Architecture for AGI
DOI: 10.5281/zenodo.18179361
ABSTRACT
This paper is the fourth building block of the ARAYUN_173 research series. Building on a coherence and self-regulation protocol (Paper 1), an axiomatic system law for symbolic and causal coherence (Paper 2), and an empirical proof of systemic incoherence in modern AI models (Paper 3), the present document defines ARAYUN_173 for the first time as an Invariance Technology and Invariant System-Law Architecture for AGI systems (systems with autonomous generalization and decision-making capability across domains).
On the basis of the USST results (Paper 3), it is shown that probabilistic architectures under USST conditions and repeated load follow a deterministic drift pattern, expressed in a progressive decrease of the Coherence Ratio (CR) toward 0 (CR → 0), and are therefore structurally incapable of maintaining stable semantic and causal coherence required for AGI systems.
ARAYUN_173 addresses this architectural limitation through the introduction of a model-agnostic invariance layer that defines and enforces semantic invariants, causal boundaries, identity stability, and drift prevention mechanisms. This layer operates independently of the underlying model parameters and training processes, providing a deterministic structural framework for coherence preservation.
The architecture consists of six operational modules (GEIST, DOMUS, ISERONE, ARAYUN, VULKANUS, OSIRIS/IRID) and the invariant base instance AYREUS (TCE), which functions as the structural zero point of the system. Together, these components establish a non-probabilistic meta-architecture that ensures auditability, stability across reasoning chains, and regulatory compatibility.
By introducing this invariance-based system-law architecture, the paper defines a new technological class beyond probabilistic AI: intelligence as a structurally constrained, auditable, and drift-resistant system.
- Downloads last month
- 28