diff options
| -rw-r--r-- | RESULTS.md | 12 | ||||
| -rw-r--r-- | REVIEW_SCORECARD.md | 1 | ||||
| -rw-r--r-- | ROADMAP.md | 7 | ||||
| -rw-r--r-- | STABLE_INNOVATION.md | 16 |
4 files changed, 36 insertions, 0 deletions
@@ -949,6 +949,18 @@ the forward local feedback mode. The linearized update in `THEORY.md` exposes the multiplicative term `D W C` and proves that any positive effective scalar mode has momentum transition spectral radius above one. +The separately frozen training-only S0 screen then tests whether a one-sided +neutral-regression upper margin is sufficient. None of +`{0.001,0.003,0.01,0.03}` is eligible. The two smaller margins produce +nonfinite BatchNorm running variances and `1e17`-scale weights. The two larger +margins keep tensors finite for 352 steps but produce maximum losses of +`1.40e5`/`4.97e6` and weights of `8.83e6`/`1.19e8`, violating the predeclared +loss, signal-ratio, weight, and momentum envelopes. Every candidate preserves +the four-to-one intervention, passes its one-sided residual-soma certificate, +freezes the predictor, restores loader state, and evaluates no validation or +test example. Sign-only dissipativity is therefore closed before endpoint +access; no margin is selected and the score remains 5/10. + ## How to run `experiments/run.py --mode {bp,fa,dfa,sdil} --dataset {mnist,fmnist,cifar10} --depth D --residual {0,1} --act {tanh,gelu,silu,relu}` Batteries: `experiments/run_v2.sh <ds> "<depths>" <res> <act> "<seeds>" <ep> <pfx>`. diff --git a/REVIEW_SCORECARD.md b/REVIEW_SCORECARD.md index 035448f..5443510 100644 --- a/REVIEW_SCORECARD.md +++ b/REVIEW_SCORECARD.md @@ -103,6 +103,7 @@ retroactively reopened by success on standard vision benchmarks. | 2026-07-22 / mixed-traffic MT-0 | Exact raw/matched/innovation mechanics pass; synthetic ResNet-20 matched batch peaks at 0.857 GB allocated on GTX 1080 without task data | 5 → 5 | Removes implementation, graph, and memory objections before endpoint access; supplies no evidence yet that innovation is useful | | 2026-07-22 / `f310ad5` mixed-traffic MT-1 | All three frozen conditions become nonfinite in epoch 1 and end at 10%; ratio calibration, predictor warmup, zero-query, and cost checks pass | 5 → 5 | Fails to make innovation load-bearing on a standard ResNet and closes MT-2/MT-3; the prior mechanism-only evidence survives but empirical support does not improve | | 2026-07-22 / `6531aed`, `d818fc2` MT-1 diagnosis | Training-only localization finds the first active failure at stem steps 23/69/74 for raw/matched/innovation; residual coupling yields the verified multiplicative operator `D W C` | 5 → 5 | Sharpens the failure into an operator-stability requirement and improves soundness of the negative analysis, but adds no successful held-out SDIL endpoint | +| 2026-07-22 / `8d28fd9` stability S0 | A frozen training-only four-margin grid has no eligible candidate; sign-certified margins either overflow BN state or produce enormous transient loss and parameter growth | 5 → 5 | Closes fixed one-sided predictor bias and motivates a two-sided dynamic gain condition, but no validation endpoint or empirical support is gained | Future rows are appended only after an audited frozen stage. A score staying flat is informative: engineering, theory exposition, or visualization may make the paper more defensible without @@ -508,6 +508,13 @@ respectively, with no validation or test evaluation. This narrows any future independent method to operator-level gain control or dissipativity; more endpoint tuning of the same residual-RMS gate is not justified. +The follow-up S0 one-sided stability branch also closes before held-out access. +Its frozen four-margin training-prefix grid has no eligible candidate: small +margins still overflow BN state and grow weights to `1e17`, while larger +margins retain finite tensors but violate loss, signal-ratio, weight, and +momentum envelopes by orders of magnitude. No validation protocol opens. This +rules out fixed sign-only predictor bias as the missing gain-control mechanism. + Prepare convolutional local-update primitives and ResNet-20/32/56 protocols early. Queue frozen runs opportunistically on authorized idle GPUs. Because BurstCCN already reports CIFAR-10 and ImageNet scaling, dataset scale alone is not novel. The oral-level target is a memorable joint diff --git a/STABLE_INNOVATION.md b/STABLE_INNOVATION.md index d499dc5..ecaf224 100644 --- a/STABLE_INNOVATION.md +++ b/STABLE_INNOVATION.md @@ -64,3 +64,19 @@ one-sided predictor branch without extending the grid. A pass only authorizes writing a separately frozen validation protocol with a new evaluation boundary; it supplies no accuracy claim and leaves the paper score at 5/10. +## Audited S0 outcome + +S0 fails with no eligible margin. `delta=0.001` and `0.003` first make +BatchNorm running variances nonfinite at steps 252 and 301; their maximum +forward/feedback weights reach `8.16e16` and `1.48e17`. `delta=0.01` and +`0.03` keep every tensor finite for 352 steps but have maximum minibatch losses +of `1.40e5` and `4.97e6`, maximum used/instruction RMS ratios of 538.4 and +6.59, and maximum weights of `8.83e6` and `1.19e8`. All four pass the +one-sided slope certificate, frozen-predictor, traffic-calibration, loader, and +zero-held-out-evaluation checks. + +The result shows that sign control alone is insufficient: a discrete momentum +mode also has a lower stability boundary, and time-varying deep covariances can +drive an initially negative effective mode beyond it. The grid is closed +without another margin. No validation protocol is opened and the reviewer +score remains 5/10. |