Back to PublicationsResearch Archive

The permanent record.

Every NeuroGenesys publication, in chronological order — preprints, defensive publications, research notes, and negative results. Each entry links to the full paper, reproducible from fixed seeds.

2026 — 9 publications

Jul 27, 2026
Which sequence architecture keeps analog in-memory compute's advantage? A cost-model answer with a design rule
Analog in-memory compute accelerates static-weight matmuls but cannot touch attention's growing KV cache, so its decode advantage dilutes with context length. The attention window policy — sliding vs. full — not the attention fraction, decides whether a hybrid keeps the advantage. Distilled from ~70 pre-registered falsification experiments.
Defensive publicationSilicon & Analog Compute
Jul 26, 2026
The displacement theorem, and the aspect no-go
Two small results, stated precisely and machine-checked (22/22 pre-registered predictions across the project): a pigeonhole-style displacement theorem — the metalanguage, not the primitive, is what cannot be eliminated — and a two-line model-theoretic no-go for aspect. Both are plainly labeled where they are trivial.
Machine-checked resultsMathematics & Logic
Jul 19, 2026
A closed-form spectral response for balanced perturbations of positive recurrences
A tidy, fully explicit special case of the weak-noise Lyapunov-exponent expansion: perturb a positive two-term recurrence so the coefficient sum is conserved, and the growth response has a closed form — with a pleasant golden-ratio corollary. Honestly labeled a worked example, not a new theorem.
Expository noteMathematics
Jul 19, 2026
Spectral response of positive recurrences under memory-filtered balanced coefficient perturbations
When an AR(1) memory process perturbs a positive two-term recurrence while conserving total coefficient weight, memory systematically accelerates growth. The main theorem gives the exact spectral response and is numerically confirmed; the proof architecture is presented as a skeleton.
Research noteMathematics
Jun 27, 2026
Catastrophic forgetting in class-incremental learning is a verifier-exploitability problem
Freeze one near-perfect expert per task and all of class-incremental difficulty becomes selection — which expert to trust. Accuracy factorizes exactly as selection accuracy × within-task accuracy, and the selection term is pinned by the selector's exploitability ε, the same quantity that bounds test-time compute.
Preprint draftContinual Learning
Jun 27, 2026
When more sampling hurts: verifier exploitability, not accuracy, bounds test-time compute
Best-of-N sampling lifts coverage log-linearly, but realized accuracy is ceilinged by the verifier — and the property that matters is not discrimination but exploitability ε: the rate at which wrong candidates score high. Any ε > 0 makes accuracy non-monotone in N, peaking at a computable budget before declining.
Preprint draftTest-Time Compute
Jun 27, 2026
Where optics-native inference actually pays: a cost-model study
Speed-of-light inference is real but narrow and conditional: an optics-native state-space model beats a GPU on batch-1 decode latency only inside a tight box — WDM microrings, small-to-mid width, ~5-bit precision, error-tolerant workloads. The model reproduces a published 2025 chip's throughput exactly; the defensible contribution is the cost model itself.
Cost-model studyPhotonics
Jun 27, 2026
Reverse-engineering a trained network for all-optical inference via random-feature kernels
Can inference on an already-trained network run entirely in light? An envelope exists, bounded by a quantified depth→width feature wall — and an honest requote against a digital ASIC collapses the headline speedup to near parity. Two genuine advantages survive: precision robustness and zero-forgetting frozen features.
Falsification studyPhotonics
Jun 27, 2026
Complex amplitudes are decorative: a pre-registered falsification of the wave/quantum thesis for learning and cognition
Across continual learning, cognition, and compositional binding, the complex/wave degree of freedom is load-bearing nowhere at equal real-parameter budget. The geometric intuition survives in real-valued form; the transferable contribution is the equal-budget falsification methodology itself.
Negative resultsFoundations

The archive is append-only: published results stay published, including the ones that did not survive. Corrections are issued in place and noted in the paper, never silently.

The permanent record.

2026 — 9 publications

Which sequence architecture keeps analog in-memory compute's advantage? A cost-model answer with a design rule

The displacement theorem, and the aspect no-go

A closed-form spectral response for balanced perturbations of positive recurrences

Spectral response of positive recurrences under memory-filtered balanced coefficient perturbations

Catastrophic forgetting in class-incremental learning is a verifier-exploitability problem

When more sampling hurts: verifier exploitability, not accuracy, bounds test-time compute

Where optics-native inference actually pays: a cost-model study

Reverse-engineering a trained network for all-optical inference via random-feature kernels

Complex amplitudes are decorative: a pre-registered falsification of the wave/quantum thesis for learning and cognition

Leadership

"The brain is the universe's only proof that general intelligence is possible. We treat it as the blueprint."

Will Schulz