Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
31 commits
Select commit Hold shift + click to select a range
b11051b
feat(notebook,#19494): Percolation-03 critique -- point critique, β/ν…
jsboige Oct 6, 2026
bfae06a
chore(percolation,#19537): empty commit pour re-trigger PR gate post …
jsboige Oct 7, 2026
56098ec
fix(percolation,#19537): prose-counts ARTIFACT_NOUNS '0 cellule' -- f…
jsboige Oct 7, 2026
06b7578
fix(percolation,#19537): STALE_LINK c.1131-N1 -- _quarto.yml entry + …
jsboige Oct 7, 2026
caad7fb
chore(percolation,#19537): empty commit pour re-trigger PR gate post …
jsboige Oct 7, 2026
4046bcf
feat(notebook,#19494): Percolation-03b critique -- extensions L=128, …
jsboige Oct 6, 2026
a29dd51
docs(index): indexation percolation-03 / -03b critique-results dans d…
jsboige Oct 6, 2026
243d704
fix(notebook,#19548): cell-order 8/9 inversion + README .ipynb → .html
jsboige Oct 6, 2026
ae588a8
fix(_quarto,#19548): ajouter Percolation-03/-03b au render list
jsboige Oct 7, 2026
6da8ae4
fix(prose-counts,#19548): retirer '0 erreur, 0 cellule' (ARTIFACT_NOU…
jsboige Oct 7, 2026
4b39c1c
chore(no-op,#19548): empty commit pour rerun PR gate apres latex-cont…
jsboige Oct 7, 2026
74678db
chore(no-op,#19548): empty commit pour rerun PR gate apres cell-order…
jsboige Oct 7, 2026
24d708d
chore(no-op,#19548): empty commit pour rerun PR gate apres Audit READ…
jsboige Oct 7, 2026
fd014d4
chore(percolation,#19548): empty commit pour re-trigger PR gate post …
jsboige Oct 7, 2026
13a4b93
Fix(percolation,#19548): instrument R² par L estimateur τ' + verdict …
jsboige Oct 7, 2026
3643e9f
fix(percolation,#19548): TAU_PRIME_C1104 corrige + verdict descriptif…
jsboige Oct 7, 2026
cbe0c5e
chore(percolation,#19548): empty commit pour re-trigger PR gate post …
jsboige Oct 7, 2026
72bb1f0
fix(percolation,#19548): 5 NanoClaw concerns + mineurs (β/ν 5/48, see…
jsboige Oct 7, 2026
de46b26
Merge remote-tracking branch 'origin/main' into HEAD
jsboige Oct 8, 2026
dfb1c78
Merge remote-tracking branch 'origin/main' into feature/c1105-percola…
jsboige Oct 8, 2026
1a68bd1
Merge origin/main into feature/c1105-percolation-03-extensions
jsboige Oct 9, 2026
57a02f8
Fix(research,#19494): cible beta/nu residuelle dans la chaine Percola…
jsboige Oct 9, 2026
e534ec0
Fix(percolation,#19548): fenetre beta/nu positive + fit log|deviation…
jsboige Oct 9, 2026
6c25b2c
Merge remote-tracking branch 'origin/main' into feature/c1105-percola…
jsboige Oct 9, 2026
5fafc55
docs: regenerer la render-list _quarto.yml apres merge origin/main
jsboige Oct 9, 2026
dd97d5a
Fix(review,#19548): M2/M3/A3 -- liens .ipynb, sections renumerotees, …
jsboige Oct 9, 2026
612f71c
Merge origin/main into feature/c1105-percolation-03-extensions
jsboige Oct 9, 2026
984a2ae
docs: regenerer la render-list _quarto.yml apres merge origin/main
jsboige Oct 9, 2026
e97a2d2
Merge origin/main into feature/c1105-percolation-03-extensions
jsboige Oct 10, 2026
4396897
Fix(percolation,#19548): cellule 9 du 03b -- prose alignee sur les so…
jsboige Oct 10, 2026
e825119
Merge branch 'main' into feature/c1105-percolation-03-extensions
jsboige Oct 10, 2026
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

Large diffs are not rendered by default.

Large diffs are not rendered by default.

Original file line number Diff line number Diff line change
Expand Up @@ -14,7 +14,7 @@ critique** (p ≈ p_c). Simulation-first : on fait *voir* les trois régimes
| [Percolation-01-Supercritique-Python](Percolation-01-Supercritique-Python.html) | 1 (Python) | Python 3 + `networkx` | Tore carré (degré 4, `p_c(bond, ℤ²) = 1/2`) et tore hexagonal (degré 3, `p_c ≈ 0.6527`) — trois régimes mesurés, géant du tore, vitesse de disparition `Φ(n) ~ √n` |
| [Percolation-02-Lean](Percolation-02-Lean.html) | 2 (Lean 4) | Lean 4 + Mathlib (`percolation_lean`) | Compagnon exécutable du lake : configurations d'arêtes ouvertes, Harris–Kleitman fini, connexité croissante, composantes, frontière isopérimétrique avec profil calculé sur `C₃`/`C₄` |
| [Percolation-03-Critique-Python](Percolation-03-Critique-Python.ipynb) | 3 (Python) | Python 3 + `networkx` | Physique du point critique : loi d'échelle finie `M(L, 1/2)/L² ~ L^{-β/ν}`, distribution de tailles `P(s) ~ s^{-τ'}`, convergence `p_c(L) → 1/2`. Sweep `L ∈ {8,16,32,64}` × `p ∈ {0.45,…,0.55}` × 32 seeds = 640 simulations `networkx.connected_components` canonique |
| [Percolation-03b-Critique-L128-L256](Percolation-03b-Critique-L128-L256.ipynb) | 3b (Python) | Python 3 + `networkx` | Extension du carnet 03 aux tailles `L ∈ {128, 256}` pour la convergence asymptotique. Fit log-log sur 6 points (L = 8, 16, 32, 64, 128, 256) : β/ν = 0.1065, τ' = 2.188, p_c(L) → 1/2 monotone. Sweep `L ∈ {128, 256}` × `p ∈ {0.45,…,0.55}` × 32 seeds = 320 simulations ~48 s |
| [Percolation-03b-Critique-L128-L256](Percolation-03b-Critique-L128-L256.ipynb) | 3b (Python) | Python 3 + `networkx` | Extension du carnet 03 aux tailles `L ∈ {128, 256}` pour la convergence asymptotique. Fit log-log sur 6 points (L = 8, 16, 32, 64, 128, 256) : β/ν = 0.1081, τ' = 2.221, p_c(L) → 1/2 monotone. Sweep `L ∈ {128, 256}` × `p ∈ {0.45,…,0.55}` × 32 seeds = 320 simulations ~48 s |
| [Percolation-03c-Critique-L512](Percolation-03c-Critique-L512.ipynb) | 3c (Python) | Python 3 + `networkx` | Extension du carnet 03b à L = 512 pour la convergence asymptotique. Fit log-log sur 7 points (L = 8, 16, 32, 64, 128, 256, 512) : β/ν = 0.1062 (cible 0.1042, R²=0.977, std_err=0.007 — convergence asymptotique bloquée aux corrections d'échelle finie), p_c(L) → 1/2 monotone (écart 0.0013 à L=512). Sweep `L = 512` × `p ∈ {0.45,…,0.55}` × 32 seeds = 160 simulations ~186 s |
| [Percolation-04-Sharpness-Python](Percolation-04-Sharpness-Python.ipynb) | 4 (Python) | Python 3 + `networkx` | Palier 4 du plan #19494 : mesure de la queue supérieure `Ψ(n) = P(\|C_o\| ≥ n)` sur grille `L ∈ {64, 128, 256, 512}` × `p ∈ {0.50, …, 0.70}` × 32 seeds = 1024 simulations ~170 s. Comparaison au théorème de *supercritical sharpness* (Diskin-Easo-Radhakrishnan-Sudakov-Tassion, arXiv:2603.03257 §3.0). **Verdict** : palier non mesurable aux seuils `n ∈ {1, 2, …, 128}` à L ≤ 512 — le fit log-y de `Ψ(n)` est dominé par la transition initiale (Psi(1)=1 → Psi(2)~0.7) et non par la décroissance exponentielle sharpness (qui exigerait `n ~ médiane du géant`). L'alpha mesuré **décroît** avec p, opposé à la cible `(p - 1/2)^{d-1} = p - 1/2` |

Expand Down
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
2 changes: 2 additions & 0 deletions _quarto.yml
Original file line number Diff line number Diff line change
Expand Up @@ -232,6 +232,8 @@ project:
- "MyIA.AI.Notebooks/ML/README.md"
- "MyIA.AI.Notebooks/NLP/README.md"
- "MyIA.AI.Notebooks/Probas/Applications/Percolation/README.md"
- "MyIA.AI.Notebooks/Probas/Applications/Percolation/README.md"
- "MyIA.AI.Notebooks/Probas/Applications/Percolation/README.md"
- "MyIA.AI.Notebooks/Probas/Applications/README.md"
- "MyIA.AI.Notebooks/Probas/DecisionTheory/Actuariat/README.md"
- "MyIA.AI.Notebooks/Probas/DecisionTheory/Causal-Bridges/README.md"
Expand Down
2 changes: 2 additions & 0 deletions docs/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -282,6 +282,8 @@ Documents de recherche durables fondant les EPICs de R&D (grade A-recherche). Di
| [research/cartier-miller-p1-pilot.md](research/cartier-miller-p1-pilot.md) | Lecture ligne-par-ligne de pilot.py (P1 elliptic curve Cartier-Miller, EPIC #19452) — orchestration, hash, per-call math, validation cross-prime 7 premiers, garde flush-après-chaque-prime. Suite PRs #19487 / #19488. |
| [research/cartier-miller-p1-point-count.md](research/cartier-miller-p1-point-count.md) | Lecture ligne-par-ligne de `point_count.py` (P1 Schoof + BSGS, EPIC #19452) — stdlib only, 0 import, 9 composants documentés. Suite des PRs #19487 (P0 cartography), #19488 (P1 elliptic_prefix), #19493 (P1 pilot). |
| [research/cartier-miller-p1-integration.md](research/cartier-miller-p1-integration.md) | Cross-check 4 backends de la P1 Cartier-Miller (EPIC #19452) : elliptic_prefix + pilot + point_count + C++ pseudocode, confrontation littérale aux `output.json` du pin reproductible `37a9b72`. P1+ à venir (P2 Lean Mathlib, P3+ multi-cycle) |
| [research/percolation-03-critique-results.md](research/percolation-03-critique-results.md) | Résultats palier 3 (#19494, PR #19537) — physique du point critique de la percolation 2D (β = 5/36, ν = 4/3, soit β/ν = 5/48 — Kesten 1980 ; τ' = 187/91 Stauffer, p_c(L) → 1/2). 580 simulations, 4 s wall-clock, instrument `networkx`. Verdict honnête : β/ν=0.102 vs cible 5/48 ≈ 0.1042 (échelle finie), p_c(L) monotone PASS |
| [research/percolation-03b-critique-results.md](research/percolation-03b-critique-results.md) | Résultats extension palier 3+ (#19494, PR #19548) — sweep L ∈ {128, 256} × 32 seeds × 5 p (320 sims, 48 s) pour vérifier la convergence asymptotique. Verdict global : convergence en cours, asymptote non atteinte — L=512 nécessaire (scipy.sparse) |
| [research/slide-agents-marptoslidev-scoping.md](research/slide-agents-marptoslidev-scoping.md) | Scoping Marp→Slidev pour les agents de slides (c.1110, #19578) — inventaire firsthand (12 Marp coexistants avec 18 Slidev, configs et outils legacy), diagnostic des fronts communs (format Marp, PNG rendering mort-né, sk-agent vision périmé), 4 voies arbitrées (réécrire/fusionner/retirer/legacy), recommandation voie 1 (réécrire pour Slidev, 3 raisons mesurées). Critère de fermeture documenté ; arbitrage user ou coordinateur requis pour passer à la phase 2 |
| [research/percolation-03-critique-results.md](research/percolation-03-critique-results.md) | Résultats d'exécution du palier 03 critique Percolation (#19494, c.1104) — mesure des exposants critiques (`β/ν`, `τ'`, `p_c(L) → 1/2`) par sweep `L × p × seeds` sur `networkx`. Verdict : tendances conformes mais convergence asymptotique non atteinte, corrections d'échelle finie dominantes. |
| [research/percolation-03b-critique-results.md](research/percolation-03b-critique-results.md) | Extension du palier 03 aux tailles `L ∈ {128, 256}` (#19494, c.1105) — vérifie la convergence asymptotique `β/ν → 5/48` et `τ' → 187/91`. Verdict : `β/ν` compatible avec la cible corrigée dès 6 points, `τ'` encore biaisé par le bulk. |
Expand Down
Loading
Loading