From 3471c1f12c33ea3e05c58b40d790872051dcc0b4 Mon Sep 17 00:00:00 2001 From: Jammy2211 Date: Fri, 23 Oct 2026 16:27:28 +0100 Subject: [PATCH] analytic_gaussian: log the #1713 NaN-evidence verdict Co-Authored-By: Claude Fable 5.1 --- dashboard.html | 6 +++--- dashboard.md | 4 ++-- projects/analytic_gaussian.md | 1 + state.json | 2 +- 4 files changed, 7 insertions(+), 6 deletions(-) diff --git a/dashboard.html b/dashboard.html index 628eb41..45b5fe4 100644 --- a/dashboard.html +++ b/dashboard.html @@ -334,7 +334,7 @@ - Open autofit_inference: https://github.com/PyAutoLabs/autofit_inference - Open euclid_detailed: https://github.com/PyAutoLabs/euclid_detailed - Open cowls: https://github.com/PyAutoLabs/cowls -- Open cowls_ansofi: https://github.com/PyAutoLabs/cowls_ansofi
Last updated 2026-10-23 14:36:18 UTC
↻ Update

+- Open cowls_ansofi: https://github.com/PyAutoLabs/cowls_ansofi
Last updated 2026-10-23 15:27:28 UTC
↻ Update

@@ -517,13 +517,13 @@

Current position

Runs (0)

nothing on the cluster

-
Recent log — 5 of 8 entries +
Recent log — 5 of 9 entries
    +
  • 2026-10-23 — note —
    second wave-1 finding closed (PyAutoFit#1713, PR #1714 merged, autofit_workspace_test#116): the `invalid value encountered in add` warnings on seeds 25 and 172 are a readout-only defect — under the mode projection an accepted truncated-nor…show less

    second wave-1 finding closed (PyAutoFit#1713, PR #1714 merged, autofit_workspace_test#116): the `invalid value encountered in add` warnings on seeds 25 and 172 are a readout-only defect — under the mode projection an accepted truncated-normal scatter message sits ~19 sigma into its tail, its log_partition underflows to −inf (true log Z ≈ −187.9), the sigma evidence goes +inf and the EP log_evidence NaN from sweep 4; posteriors bit-identical to the ens_n5 sidecars. The library now names the term once per fit and returns an explicit NaN. The cure (tail-stable log_partition via log_ndtr) is message maths, filed in the Mind as draft/bug/autofit/truncated_normal_log_partition_tail_underflow.md for the human.

  • 2026-10-23 — note —
    criterion 6 amended (PyAutoFit#1711, PR #1712): the 19/400 zero-SUCCESS seed-legs of wave 1 split into 13 NO_CHANGE-only prior factors (zero-iteration Laplace convergence at the start point, projection accepted, closed-form match a 5e-5 /…show less

    criterion 6 amended (PyAutoFit#1711, PR #1712): the 19/400 zero-SUCCESS seed-legs of wave 1 split into 13 NO_CHANGE-only prior factors (zero-iteration Laplace convergence at the start point, projection accepted, closed-form match a 5e-5 / b 2e-5 on seed 140 leg A — a converged update, not stale; library reporting aligned, no EP maths changed) and 6 BAD_PROJECTION-only hierarchical factors (the #1405 scatter state, cured by projection="moments" #1656; seeds 29 and 71 re-run under moments give HierarchicalFactor SUCCESS 28 and 23). Criterion 6 now counts SUCCESS or NO_CHANGE as a completed update; the matching ensemble_seed.py any_factor_never_succeeded change is the human's to apply in the science checkout. Side finding: seed 29 leg B under moments still misses the sigma cell (a 0.225 / b 0.416) while sigma stays inside [q05, q95].

  • 2026-10-14 — note —
    PyAutoFit#1656 (merged 2026-09-30, released 2026.10.7.1) ports the minimal-EP referee's moment matching as LaplaceOptimiser(projection="moments"); on analytic_gaussian.py leg B autofit EP sigma 6.3401 ± 2.4609 vs closed form 6.5667 ± 2.883…show less

    PyAutoFit#1656 (merged 2026-09-30, released 2026.10.7.1) ports the minimal-EP referee's moment matching as LaplaceOptimiser(projection="moments"); on analytic_gaussian.py leg B autofit EP sigma 6.3401 ± 2.4609 vs closed form 6.5667 ± 2.8832 (mode path 9.3722, BAD_PROJECTION=19). A wave-2 ensemble on the moments path needs a private PyAutoFit checkout on RAL: the shared mirror is 404b3e5f7 and lacks it. The criterion-2 mu-threshold diagnostic is still unrun.

  • 2026-10-14 — note — output/ vaulted to z_vault/ on laptop and RAL (mv, nothing deleted), anticipating reruns; output/ left empty, results/ untouched; z_vault/ gitignored locally
  • 2026-09-30 — note —
    astra (Codex gpt-6-astra) on criterion 2, verbatim: 'Neither is established as "wrong" by this result alone. I favour inadequate calibration of the threshold, with moderate confidence, but would not yet certify the control.' Decisive diagn…show less

    astra (Codex gpt-6-astra) on criterion 2, verbatim: 'Neither is established as "wrong" by this result alone. I favour inadequate calibration of the threshold, with moderate confidence, but would not yet certify the control.' Decisive diagnostic: independently reconstruct the same EP factorisation at the fixed point on a passing, an edge and the worst mu seed and compare tilted vs q moments; audit the per-site sigma>0 clip (analytic_ep_minimal.py:333). Opus reviewer concurs independently (a_mu = 0.041*(E_ref[mu]-m0)/std_ref over 200 seeds, corr 0.98; calibration seeds 0-4 had |a_mu| 0.002-0.020 vs 95th pct 0.077). Full answer: PyAutoCortex#50 comment https://github.com/PyAutoLabs/PyAutoCortex/issues/50#issuecomment-5908448194

    • -
  • 2026-09-10 — result — accepted ensemble_parity (R-20260910-04): ok accept and intake the things to address. we can do another run down the line so also achieve the results in the analytic_gaussian project for future comparison

full log

active · ral · PyAutoCortex#34

diff --git a/dashboard.md b/dashboard.md index 68c7f34..4151eac 100644 --- a/dashboard.md +++ b/dashboard.md @@ -321,13 +321,13 @@ Use the cortex skill. — resume analytic_gaussian: read PyAutoCortex projects/a _nothing on the cluster_ -
Recent log — 5 of 8 entries +
Recent log — 5 of 9 entries +- 2026-10-23 — *note* — second wave-1 finding closed (PyAutoFit#1713, PR #1714 merged, autofit_workspace_test#116): the `invalid value encountered in add` warnings on seeds 25 and 172 are a readout-only defect — under the mode projection an accepted truncated-normal scatter message sits ~19 sigma into its tail, its log_partition underflows to −inf (true log Z ≈ −187.9), the sigma evidence goes +inf and the EP log_evidence NaN from sweep 4; posteriors bit-identical to the ens_n5 sidecars. The library now names the term once per fit and returns an explicit NaN. The cure (tail-stable log_partition via log_ndtr) is message maths, filed in the Mind as draft/bug/autofit/truncated_normal_log_partition_tail_underflow.md for the human. - 2026-10-23 — *note* — criterion 6 amended (PyAutoFit#1711, PR #1712): the 19/400 zero-SUCCESS seed-legs of wave 1 split into 13 NO_CHANGE-only prior factors (zero-iteration Laplace convergence at the start point, projection accepted, closed-form match a 5e-5 / b 2e-5 on seed 140 leg A — a converged update, not stale; library reporting aligned, no EP maths changed) and 6 BAD_PROJECTION-only hierarchical factors (the #1405 scatter state, cured by projection="moments" #1656; seeds 29 and 71 re-run under moments give HierarchicalFactor SUCCESS 28 and 23). Criterion 6 now counts SUCCESS or NO_CHANGE as a completed update; the matching ensemble_seed.py any_factor_never_succeeded change is the human's to apply in the science checkout. Side finding: seed 29 leg B under moments still misses the sigma cell (a 0.225 / b 0.416) while sigma stays inside [q05, q95]. - 2026-10-14 — *note* — PyAutoFit#1656 (merged 2026-09-30, released 2026.10.7.1) ports the minimal-EP referee's moment matching as LaplaceOptimiser(projection="moments"); on analytic_gaussian.py leg B autofit EP sigma 6.3401 ± 2.4609 vs closed form 6.5667 ± 2.8832 (mode path 9.3722, BAD_PROJECTION=19). A wave-2 ensemble on the moments path needs a private PyAutoFit checkout on RAL: the shared mirror is 404b3e5f7 and lacks it. The criterion-2 mu-threshold diagnostic is still unrun. - 2026-10-14 — *note* — output/ vaulted to z_vault/ on laptop and RAL (mv, nothing deleted), anticipating reruns; output/ left empty, results/ untouched; z_vault/ gitignored locally - 2026-09-30 — *note* — astra (Codex gpt-6-astra) on criterion 2, verbatim: 'Neither is established as "wrong" by this result alone. I favour inadequate calibration of the threshold, with moderate confidence, but would not yet certify the control.' Decisive diagnostic: independently reconstruct the same EP factorisation at the fixed point on a passing, an edge and the worst mu seed and compare tilted vs q moments; audit the per-site sigma>0 clip (analytic_ep_minimal.py:333). Opus reviewer concurs independently (a_mu = 0.041*(E_ref[mu]-m0)/std_ref over 200 seeds, corr 0.98; calibration seeds 0-4 had |a_mu| 0.002-0.020 vs 95th pct 0.077). Full answer: PyAutoCortex#50 comment https://github.com/PyAutoLabs/PyAutoCortex/issues/50#issuecomment-5908448194 -- 2026-09-10 — *result* — accepted ensemble_parity (R-20260910-04): ok accept and intake the things to address. we can do another run down the line so also achieve the results in the analytic_gaussian project for future comparison [full log](https://github.com/PyAutoLabs/PyAutoCortex/blob/main/projects/analytic_gaussian.md) diff --git a/projects/analytic_gaussian.md b/projects/analytic_gaussian.md index 7890f4f..c394571 100644 --- a/projects/analytic_gaussian.md +++ b/projects/analytic_gaussian.md @@ -11,6 +11,7 @@ Wave 1 (342413, 200 seeds at N=5) is accepted as the baseline: autofit EP is exa ## Log +- 2026-10-23 — note — second wave-1 finding closed (PyAutoFit#1713, PR #1714 merged, autofit_workspace_test#116): the `invalid value encountered in add` warnings on seeds 25 and 172 are a readout-only defect — under the mode projection an accepted truncated-normal scatter message sits ~19 sigma into its tail, its log_partition underflows to −inf (true log Z ≈ −187.9), the sigma evidence goes +inf and the EP log_evidence NaN from sweep 4; posteriors bit-identical to the ens_n5 sidecars. The library now names the term once per fit and returns an explicit NaN. The cure (tail-stable log_partition via log_ndtr) is message maths, filed in the Mind as draft/bug/autofit/truncated_normal_log_partition_tail_underflow.md for the human. - 2026-10-23 — note — criterion 6 amended (PyAutoFit#1711, PR #1712): the 19/400 zero-SUCCESS seed-legs of wave 1 split into 13 NO_CHANGE-only prior factors (zero-iteration Laplace convergence at the start point, projection accepted, closed-form match a 5e-5 / b 2e-5 on seed 140 leg A — a converged update, not stale; library reporting aligned, no EP maths changed) and 6 BAD_PROJECTION-only hierarchical factors (the #1405 scatter state, cured by projection="moments" #1656; seeds 29 and 71 re-run under moments give HierarchicalFactor SUCCESS 28 and 23). Criterion 6 now counts SUCCESS or NO_CHANGE as a completed update; the matching ensemble_seed.py any_factor_never_succeeded change is the human's to apply in the science checkout. Side finding: seed 29 leg B under moments still misses the sigma cell (a 0.225 / b 0.416) while sigma stays inside [q05, q95]. - 2026-10-14 — note — PyAutoFit#1656 (merged 2026-09-30, released 2026.10.7.1) ports the minimal-EP referee's moment matching as LaplaceOptimiser(projection="moments"); on analytic_gaussian.py leg B autofit EP sigma 6.3401 ± 2.4609 vs closed form 6.5667 ± 2.8832 (mode path 9.3722, BAD_PROJECTION=19). A wave-2 ensemble on the moments path needs a private PyAutoFit checkout on RAL: the shared mirror is 404b3e5f7 and lacks it. The criterion-2 mu-threshold diagnostic is still unrun. - 2026-10-14 — note — output/ vaulted to z_vault/ on laptop and RAL (mv, nothing deleted), anticipating reruns; output/ left empty, results/ untouched; z_vault/ gitignored locally diff --git a/state.json b/state.json index 14ebf6a..38df967 100644 --- a/state.json +++ b/state.json @@ -4,7 +4,7 @@ "repo": "PyAutoCortex", "status": "yellow", "headline": "YELLOW \u2014 12 active \u00b7 16 running", - "updated": "2026-10-23T14:36:18Z", + "updated": "2026-10-23T15:27:28Z", "pages_url": "https://pyautolabs.github.io/PyAutoCortex/", "items": [ {