∮Historical archive tag
#onn
13 documents preserved from the archived ONN programme. These records are not the current Main research.
ℓNotes
1χONN documents
6- ONN documents
ONN Canonical Audit & Status
What the Ontology Neural Network programme set out to build, and what its own rigorous audit established: an identifiability boundary, a modest surviving positive, and narrowly scoped model-level control results. The programme has since moved to ULR.
- ONN documents
ONN Research Status — 2026
Current, audited status of the ONN / ORTSF programme: the central higher-order question resolved to a scoped No-Go boundary, one modest positive survived, control results were narrowed to model-level certificates, and the programme moved to ULR.
- ONN documents
Audit — the 'Cohomological Stability Certificate'
The claimed 'cohomology = Lyapunov certificate' does not exist in the research source. What remains is a model-level phase-margin gate and a separate single-mode scalar delay certificate, neither derived from cohomology.
- ONN documents
ORTSF Delay-Margin Certificate
An honestly scoped ORTSF control note: a model-level phase-margin gate implemented in a guarded pipeline, plus a single-mode scalar delay-independent certificate. Multimode and deployed-controller stability remain open.
- ONN documents
ONN — Feb-2026 Snapshot (superseded)
A dated 2026-02-18 development snapshot of the ONN repository, kept for history. Its framing (a 'PLAUSIBLE' stability theorem heading to deployment) has been superseded by the programme's audited resolution — a scoped No-Go boundary. See the current research status.
- ONN documents
ONN vs Transformer — Experiment Plan
A Feb-2026 pre-registered protocol for a constraint-satisfaction (CSR/ASR/SVR) benchmark of an ONN LOGOS solver vs. Transformer baselines. Kept as a methodology record; its sibling results docs are not reflected in the current authoritative research source.
¶Papers
4- Papers · 2026
Ontology Neural Network and ORTSF: A Framework for Topological Reasoning and Delay-Robust Control
Paper of record for the original ONN + ORTSF framing. A later audit found that the higher-order/cohomological thesis adds no information beyond pairwise data un
- Papers · 2025
Advanced Topology-Preserving Neural Networks: An Extension of ONN/ORTSF Framework with Dynamic Structural Optimization
The headline empirical figures and transfer results are not reproducible from the current authoritative research source. This preprint is retained as a historic
- Papers · 2025
Constructive Lyapunov Functions via Topology-Preserving Neural Networks
The claimed constructive cohomological Lyapunov certificate, together with the 177 microsecond / 3M-node and 99.75% figures, is not supported by the current aut
- Papers · 2025
Ontology Neural Networks for Topologically Conditioned Constraint Satisfaction
The reported topology-loss and 95% constraint-satisfaction figures are not reproducible from the current authoritative research source. This manuscript is retai
∂Journal
2- Journal · 2026
ONN project status report -- codebase, tests, and paper readiness
Snapshot of the ONN-ES codebase at ~45,800 lines across 154 files: module completion, 98.7% test pass rate, 11/12 paper figures done, and a prioritized punch list.
- Journal · 2026
ONN + ORTSF — a field note on testable claims and proxy controllers
A compact field note distilled from a multi-round research-planning session: testable claims for the ONN + ORTSF pipeline, an evaluation matrix, acceptance thresholds, and the risks that most threaten faithful reproduction.