| <!doctype html> |
| <html lang="en"> |
| <head> |
| <meta charset="utf-8" /> |
| <meta name="viewport" content="width=device-width, initial-scale=1" /> |
| <meta |
| name="description" |
| content="Traceable historic 100-ID Gaussian-mixture evidence for Diffusion differentiable resampling." |
| /> |
| <link rel="icon" href="data:," /> |
| <title>DiffRes: traceable 100-ID Gaussian-mixture study</title> |
| <style> |
| :root { |
| color-scheme: light dark; |
| --paper: #f8fafc; |
| --ink: #172033; |
| --muted: #526076; |
| --panel: #ffffff; |
| --line: #d8e0ec; |
| --accent: #215bd8; |
| --warning: #7a4d00; |
| --warning-bg: #fff6e3; |
| } |
| |
| @media (prefers-color-scheme: dark) { |
| :root { |
| --paper: #10151f; |
| --ink: #edf2ff; |
| --muted: #afbbce; |
| --panel: #182131; |
| --line: #2b384d; |
| --accent: #8bb2ff; |
| --warning: #ffce72; |
| --warning-bg: #2b2417; |
| } |
| } |
| |
| * { box-sizing: border-box; } |
| body { |
| margin: 0; |
| background: var(--paper); |
| color: var(--ink); |
| font: 16px/1.55 system-ui, -apple-system, BlinkMacSystemFont, "Segoe UI", sans-serif; |
| } |
| main { max-width: 900px; margin: 0 auto; padding: 56px 24px 72px; } |
| h1, h2 { line-height: 1.2; } |
| h1 { font-size: clamp(2rem, 5vw, 3.25rem); margin: 0.45rem 0 1rem; } |
| h2 { margin: 2.4rem 0 0.7rem; font-size: 1.3rem; } |
| p, li { max-width: 76ch; } |
| a { color: var(--accent); } |
| .eyebrow { color: var(--muted); font-weight: 700; letter-spacing: 0.08em; text-transform: uppercase; } |
| .status { |
| display: inline-block; |
| border: 1px solid var(--line); |
| border-radius: 999px; |
| padding: 0.2rem 0.65rem; |
| color: var(--muted); |
| font-size: 0.85rem; |
| font-weight: 650; |
| } |
| .card, .warning { |
| border: 1px solid var(--line); |
| border-radius: 12px; |
| padding: 1.1rem 1.2rem; |
| background: var(--panel); |
| } |
| .warning { background: var(--warning-bg); color: var(--warning); } |
| table { border-collapse: collapse; width: 100%; margin: 1rem 0; } |
| th, td { border-bottom: 1px solid var(--line); padding: 0.7rem 0.55rem; text-align: left; vertical-align: top; } |
| th:first-child, td:first-child { padding-left: 0; } |
| th:last-child, td:last-child { padding-right: 0; } |
| .number { font-variant-numeric: tabular-nums; text-align: right; white-space: nowrap; } |
| footer { border-top: 1px solid var(--line); color: var(--muted); font-size: 0.9rem; margin-top: 3rem; padding-top: 1rem; } |
| </style> |
| </head> |
| <body> |
| <main> |
| <div class="eyebrow">ICML 2026 evidence package</div> |
| <h1>DiffRes: traceable 100-ID Gaussian-mixture study</h1> |
| <p class="status">Historic stored output · static package · no numerical execution</p> |
| <p> |
| A local study wrapper using released implementation, stored numerical random keys and pinned source |
| config recorded a traceable 100-ID Gaussian-mixture result; a separate validator recomputed |
| raw-record aggregates, paired intervals, and source hashes. |
| </p> |
|
|
| <section class="warning" aria-labelledby="boundary-title"> |
| <h2 id="boundary-title">Interpretation boundary</h2> |
| <p> |
| This is not an exact reproduction, independent replication, complete Table 1 or Table 2 |
| reproduction, publication score, or broad outperformance claim. It is a stored, |
| traceable comparison against the multinomial baseline only. This static package |
| does not verify a source checkout or numerical execution, and it does not establish |
| source-hash identity, execution provenance, paper-level results, or significance. |
| </p> |
| </section> |
|
|
| <section aria-labelledby="record-title"> |
| <h2 id="record-title">Recorded configuration</h2> |
| <div class="card"> |
| <ul> |
| <li>Released <code>zgbkdlm/diffres</code> source at <code>767effe3e755067eb8a04422597fbf37eb8ab754</code>.</li> |
| <li>Stored Monte Carlo IDs 0–99; 10,000 particles; 1,000 sliced-Wasserstein projections.</li> |
| <li>Probability-flow ODE, <code>T=3</code>, 128 steps, Jentzen--Kloeden integrator.</li> |
| <li>Diffusion resampling and multinomial baseline; no other Table 1 baselines.</li> |
| </ul> |
| </div> |
| </section> |
|
|
| <section aria-labelledby="results-title"> |
| <h2 id="results-title">Historical stored aggregates</h2> |
| <table> |
| <thead> |
| <tr> |
| <th scope="col">Metric (lower is better)</th> |
| <th scope="col" class="number">Diffusion resampling</th> |
| <th scope="col" class="number">Multinomial baseline</th> |
| </tr> |
| </thead> |
| <tbody> |
| <tr> |
| <td>Sliced Wasserstein L1, mean ± population SD</td> |
| <td class="number">0.0807686 ± 0.0212899</td> |
| <td class="number">0.0824334 ± 0.0250323</td> |
| </tr> |
| <tr> |
| <td>Squared posterior-mean residual L2, mean ± population SD</td> |
| <td class="number">0.0373908 ± 0.0299358</td> |
| <td class="number">0.0378330 ± 0.0442877</td> |
| </tr> |
| </tbody> |
| </table> |
| <p> |
| The recorded paired 95% t intervals cross zero for both metrics. They are not evidence of |
| a statistically significant improvement. |
| </p> |
| </section> |
|
|
| <section aria-labelledby="evidence-title"> |
| <h2 id="evidence-title">Inspect the evidence</h2> |
| <ul> |
| <li><a href="evidence/gms-selected-ids-0-99.json">Raw 100-ID result</a></li> |
| <li><a href="evidence/gms-selected-ids-0-99-validation.json">Historic validation record</a></li> |
| <li><a href="evidence/validate_gms_selected_result.py">Validator source</a></li> |
| <li><a href="evidence/verify_stored_gms_evidence.py">Self-contained stored-evidence verifier</a></li> |
| <li><a href="evidence/README.md">Evidence guide and validator limits</a></li> |
| <li><a href="MANIFEST.md">Source and artefact manifest</a></li> |
| <li><a href="README.md">Full README, safety note, and licensing</a></li> |
| </ul> |
| </section> |
|
|
| <section aria-labelledby="safety-title"> |
| <h2 id="safety-title">Selection and safety</h2> |
| <p> |
| The package contains only static presentation files, two historical JSON records, a |
| non-networked audit script, and licensing/provenance notes. It deliberately excludes the |
| upstream source tree, virtual environments, caches, runners, model artefacts, partial |
| studies, and LGSSM material. |
| </p> |
| </section> |
|
|
| <footer> |
| <a href="https://arxiv.org/abs/2512.10401">Paper</a> · |
| <a href="https://openreview.net/forum?id=M0e5XORjAW">OpenReview</a> · |
| <a href="https://github.com/zgbkdlm/diffres">released source</a> · |
| <a href="NOTICE">attribution and licence notice</a> |
| </footer> |
| </main> |
| </body> |
| </html> |
|
|
