build_supplement.py assembles the anonymized code+data archive with a SHA256 manifest and a machine readable seed manifest; research notes, drafts, raw datasets, and git history are excluded. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Pulls the training-set-size series, snapshot-location sweep, stress grid, scalar-gain comparison, oracle diagnostic, and derivative-probe rows from the experiment machine, and adds three generated fragments (train size, snapshot location, stress grid) used by the appendix. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
finite_time_supplement.py computes 95% cluster bootstrap intervals for the predictor metrics (resampling forward initializations within each cell) and renders the six-panel supplementary figure. The table generator gains the teacher-task nested-SE fragment, computed from the raw paired rows pulled from the experiment machine. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
generate_paper_tables.py reads the row-level CSVs and emits the four appendix tables (initialization calibration rows, finite-time cells, CNN calibration rows, learning-rate regimes) with the paper's nested uncertainty convention. The fragments are included by the manuscript via \input; no table value is transcribed manually. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Panel 1a: rename to Backward rules; DFA drawn as direct side-to-side arrows from the error to each hidden layer. Panel 1b: replace the dual y-axis with stacked guarantee/cost sub-axes. Figure 2: reader-facing panel titles; legend renames (linear extrapolation, relinearization). Track the regenerated figure-source CSVs (verified bit-exact against the frozen paper numbers) so figures can be rebuilt without rerunning experiments. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Root README orients the repo; notes/README reindexed by status (writing pipeline / authoritative note per result / six historical phases with superseded markers); scripts/README gains a status table. Note 40 = reproduction manifest (claim -> command -> outputs -> note; everything unlisted is not citable). Note 41 = the binding paper plan: AAAI-27 constraints (7pp + refs, appendix not guaranteed read), narrative, section/page budget, frozen theorem and figure numbering, related-work list with per-cite rationale, writing rules, weekly schedule to the 07-20/07-27 deadlines, and risk register. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
SGD teacher sweep (widths 8-96, 24 FA runs/width): train gap soft and monotone (0.249 -> 0.0069) while the test gap changes sign — FA generalizes better than BP at intermediate width (-0.120 +/- 0.013 at w=32) and settles slightly positive once both fit. Scopes the paper's optimization-gap claims and preempts the generalization question. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
New script validates both no-fit claims on MNIST-subset MLPs. e0 theorem: 16 configs, hidden share 0.39-0.82, measured E_B[e0] within 0.0001-0.0089 of prediction (512 real-backward draws each). Estimator: 96 trajectories, velocity MAE 0.00196 (3.9% of mean gap), bias -0.0006, beating fixed K(0) everywhere; kernels computed via a layerwise gram identity self-tested against autograd at machine precision. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The sampler previously drew e0 through the analytic linear form <C,B>, making the Gaussian KS test near-tautological. It now runs the real FA backward pass per feedback draw and reports the max deviation from the linear identity (<= 8.9e-16). Rerun reproduces the published table. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>