Skip to content

chore(twin-parity,#8057): re-attest ML-5 TimeSeries — seul drift restant du registre (cf #8264) #14494

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
myia-ai-01 merged 1 commit into from
Sep 4, 2026
Merged

chore(twin-parity,#8057): re-attest ML-5 TimeSeries — seul drift restant du registre (cf #8264) #14494

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 8 additions & 0 deletions scripts/notebook_tools/twin_pairs.d/ml-5-timeseries.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -24,7 +24,15 @@
csharp_sha: b0c0a5297ace0d8fe490901b3dc8e522af15d649
content_python_sha: 083c392f8d9070ae5b22b7bfcbb6b57c5a6a5f8365929cf45134c5a93f2ca3eb
content_csharp_sha: e24fbc32219e7ebaeea9223982e906434465dd83c3394fc9a3be59b3d0384db0
- date: "2026-09-03"
by: myia-po-2026:CoursIA
python_sha: 5e863ede175ce3bf689752df4e3750ac08f2a71f
csharp_sha: 1b66717cf364decb17a23edbd87c1df5df9099cb
content_python_sha: 083c392f8d9070ae5b22b7bfcbb6b57c5a6a5f8365929cf45134c5a93f2ca3eb
content_csharp_sha: 5f5671cb8d5021a4237c9f61a1a47f29baf213a6baf32c90bbe5383868081f85
reason: "Re-attestation drift pre-existant (PR dediee cf #8264) : #13001 (fix eval SSA #12988, commit 9e03b82e3) a modifie le twin C# SANS re-attester -- csharp b0c0a529 -> 1b66717c, python 5e863ede inchange (cote OK). Audit firsthand du delta : moteurs intacts (ForecastBySsa + Microsoft.ML.TimeSeries 5.0.0), 16/16 cellules code execution_count non-null, 0 output erreur, SelectMany.Zip buggé absent (0 occurrence), protocole origine-horizon explicite + baseline naive executee (verifie grep), +1 Exercice 6 temperature (stub TODO C.1-compliant). Parite native-both intacte."
known_differences:
- "2026-09-03 : protocole d evaluation C# refondu par #13001 (#12988) sans changement de moteur : table alignee (origine x horizon, reel vs predit vs naive) remplace le SelectMany.Zip non aligne (MAE 33,8 / RMSE 41,8 bugges), baseline naive executee, + Exercice 6 temperature. Les metriques C# ne sont PAS comparables au twin Python (STL/SARIMAX, cf non-comparabilite documentee ci-dessus) -- divergence de protocole d evaluation, pas de parite de moteur."
- "2026-08-19 : reclassification surface -> native-both (adjudication dispatch #10382 '3 paires surface', lane myia-po-2026:CoursIA). Les DEUX cotes font le travail du notebook au moyen d'un moteur de production nomme de leur ecosysteme : C# = Microsoft.ML.TimeSeries 5.0.0 ForecastBySsa (15/15 cellules code executees), Python = statsmodels STL + SARIMAX (12/12 executees). La difference de famille algorithmique reste documentee ci-dessous -- difference de METHODE, pas absence de moteur."
- "2026-08-15 : correction d'auteurs falsifies (issue #11044, tranche W3C SPARQL / ML.NET KDD) — citations remises a leurs auteurs reels (Amizadeh S., DOI KDD, Protocol et al., etc.), markdown-only, aucun changement d'engine/outputs. Parite semantique conservee ; rebaseline content SHA (re-audit 2026-08-15, myia-po-2026:CoursIA-2)."
- "FAMILLES ALGORITHMIQUES DIFFERENTES (pas un simple idiome swap) : C# = `Microsoft.ML.Transforms.TimeSeries` -> `ForecastBySsa` (Singular Spectrum Analysis, decomposition spectrale ML.NET-native). Python = `statsmodels` -> `STL` (seasonal-trend LOESS decomposition) + `SARIMAX` (seasonal ARIMA classique)."
Expand Down
Loading