Buckets:
| <!-- | |
| Vendored + adapted into posterly from ARIS (Auto-claude-code-research-in-sleep), | |
| skill paper-poster-html. Origin: https://github.com/wanshuiyin/Auto-claude-code-research-in-sleep | |
| Upstream tokenization: MIT (c) 2026 wanshuiyin -- ../LICENSES/aris-MIT.txt | |
| This is the tokenized form of posterly's own neutral template; the class names and | |
| content are posterly's own (c) 2026 Ruishuo Chen. As an adapted derivative this file | |
| ships as part of posterly under AGPL-3.0. See ../NOTICE.md. | |
| --> | |
| <!-- | |
| ============================================================ | |
| TEMPLATE: landscape_4col (ARIS fork) | |
| CANVAS: 60in × 36in landscape (ICML / NeurIPS / generic landscape) | |
| LAYOUT: header → optional banner → 4 body columns → optional takeaways → footer | |
| USE WHEN: standard ML conference poster with 3-5 content cards per column, | |
| mix of figures + equations + small tables. | |
| HOW TO USE: | |
| 1. Copy this file to your working directory as `poster.html`. | |
| 2. Edit the DESIGN TOKENS in `:root` to your lab/venue colors. | |
| 3. Replace TODO content placeholders (search for "TODO"). | |
| 4. Run `python tools/run_gates.py poster.html` (the default driver: preflight -> style -> measure -> polish) to align columns and check style/structure. | |
| 5. Run `python tools/render_preview.py poster.html` to produce the PDF. | |
| NOTE: as shipped, this scaffold passes `preflight` (structure) but is | |
| EXPECTED to FAIL `measure`/`polish` -- figures are commented out and copy | |
| is TODO stubs, so columns only fill the top of the canvas. Those two gates | |
| judge a FILLED poster; balance them once you've added real content. See | |
| templates/README.md ("Scaffolds, not finished posters"). | |
| MEASURE ROLES: every layout-critical element carries `data-measure-role`. | |
| Generic measurement scripts depend on these — do not remove. | |
| CANVAS RETARGETING: to print at another size, change the canvas in exactly | |
| TWO places that must stay in sync -- the `@page { size: ... }` rule and the | |
| `.poster { width/height }` (also the `@media print .poster { width/height }`). | |
| Example: ICLR 2026 main conference uses 185cm 90cm landscape (official print | |
| service spec) -- set @page to `185cm 90cm`, width to `1850 * var(--u)` and | |
| height to `900 * var(--u)` (and the print-override .poster to `185cm`/`90cm`). | |
| Adapted from posterly (MIT, © 2026 Ruishuo Chen) — see LICENSES/ & NOTICE.md; | |
| ARIS modifications: flat de-gradient, --fs token scale, zero-inline-style | |
| utilities, data-source/data-color-exempt contracts. | |
| ============================================================ | |
| --> | |
| <html lang="en"> | |
| <head> | |
| <meta charset="UTF-8"> | |
| <meta name="viewport" content="width=device-width, initial-scale=1.0"> | |
| <title>Repro: Stable Deep RL via Isotropic Gaussian Representations — ICML 2026</title> | |
| <!-- MathJax v3 for inline equations. CDN by default; to go offline, | |
| download a MathJax v3 release into ./mathjax/ and change the <script> | |
| `src` below to "mathjax/es5/tex-svg.js". --> | |
| <script> | |
| window.MathJax = { | |
| tex: { | |
| inlineMath: [['$', '$'], ['\\(', '\\)']], | |
| displayMath: [['$$', '$$'], ['\\[', '\\]']], | |
| packages: {'[+]': ['ams']} | |
| }, | |
| svg: { fontCache: 'global' } | |
| }; | |
| </script> | |
| <script id="MathJax-script" async src="https://cdn.jsdelivr.net/npm/mathjax@3/es5/tex-svg.js"></script> | |
| <style> | |
| /* ========================================================= | |
| CANVAS — 60" × 36" landscape | |
| ========================================================= */ | |
| @page { size: 60in 36in; margin: 0; } | |
| :root { | |
| /* ===== DESIGN TOKENS ===== */ | |
| /* Primary accent (header underline, card-highlight bar, section nums, .keyword) */ | |
| --accent: #CB4743; | |
| --accent-deep: #8E322F; | |
| --accent-light: #FAEDEC; | |
| --accent-soft: #F6DEDD; | |
| /* Secondary accent for "ours / best" emphasis (table .ours row, "★" callouts) */ | |
| --gold: #3D4A5C; | |
| --gold-soft: #ECEDEF; | |
| /* Text */ | |
| --text-primary: #1A1A1A; | |
| --text-secondary: #555555; | |
| --text-muted: #888888; | |
| /* Backgrounds */ | |
| --bg-page: #F6F2F0; | |
| --bg-card: #FFFFFF; | |
| --bg-card-tint: #FAFAFB; | |
| --bg-emphasis: var(--accent-light); | |
| /* Borders */ | |
| --border-soft: #D8D8D8; | |
| --border-strong: var(--accent); | |
| /* Screen-only dark mat behind the poster (the off-canvas viewport | |
| background). Print resets html/body to white, so this never reaches | |
| paper; kept as a token so no color literal lives outside this block. */ | |
| --bg-viewport: #2B2B2B; | |
| /* Base unit. Print: 1mm; screen preview: 1.6px (~3.78px = 1mm at 96dpi). */ | |
| --u: 1.6px; | |
| /* Font-size scale (9 archetypes). Every `font-size` in this template | |
| references one of these. `calc(var(--fs-N) * k)` is permitted ONLY for | |
| a COMPONENTS.md-defined variant (e.g. `.eqn--large`). */ | |
| --fs-1: calc(9 * var(--u)); /* micro label */ | |
| --fs-2: calc(10 * var(--u)); /* small caption */ | |
| --fs-3: calc(11 * var(--u)); /* caption / table */ | |
| --fs-4: calc(12 * var(--u)); /* body text */ | |
| --fs-5: calc(13 * var(--u)); /* equation / emphasis */ | |
| --fs-6: calc(15 * var(--u)); /* subtitle */ | |
| --fs-7: calc(16 * var(--u)); /* section title */ | |
| --fs-8: calc(22 * var(--u)); /* banner number */ | |
| --fs-9: calc(32 * var(--u)); /* main title */ | |
| /* Fonts. Override at :root if your venue mandates a specific family. */ | |
| --font-serif: "Charter", "Source Serif Pro", "Georgia", serif; | |
| --font-sans: "Inter", "Helvetica Neue", sans-serif; | |
| /* Shadows & watermark ink — tokenized so rules 1/3 stay literal-free | |
| (the faint radial page tint is the ONE allowed literal outside this | |
| block; rule 5 validates it WHEN ENABLED -- the style gate ships with rules 4-5 off). */ | |
| --shadow-screen: 0 0 60px rgba(0, 0, 0, 0.5); | |
| --shadow-card: 0 calc(2 * var(--u)) calc(6 * var(--u)) rgba(142, 50, 47, 0.05); | |
| --ornament-ink: rgba(142, 50, 47, 0.06); | |
| /* ===== END DESIGN TOKENS ===== */ | |
| } | |
| /* ========================================================= | |
| RESET + BASE | |
| ========================================================= */ | |
| * { box-sizing: border-box; margin: 0; padding: 0; } | |
| html, body { | |
| background: var(--bg-viewport); | |
| font-family: var(--font-serif); | |
| color: var(--text-primary); | |
| -webkit-font-smoothing: antialiased; | |
| } | |
| /* ========================================================= | |
| UTILITY CLASSES — replace inline `style=` in the body markup. | |
| The template (and any finished poster) must carry ZERO `style=` | |
| attributes. Use these classes instead. Two documented exceptions | |
| exist ONLY as examples and may appear in comments, never live: | |
| - a logo/seal SVG flagged data-color-exempt="logo" | |
| - a paper figure `<img ... data-source="paper" style="width: NN%">` | |
| (AR width tweak; prefer the .w-NN classes below where possible) | |
| ========================================================= */ | |
| /* Poster container — exact print dimensions; data-measure-role="poster" so poster_check.py | |
| can verify the canvas size. */ | |
| .poster { | |
| width: calc(1524 * var(--u)); | |
| height: calc(914 * var(--u)); | |
| background: var(--bg-page); | |
| /* Page tint kept (radial only, all color stops alpha <= 0.06 per the | |
| de-gradient policy: linear-gradient is banned, this low-alpha radial | |
| wash is the single allowed exception). */ | |
| background-image: | |
| radial-gradient(ellipse at top left, rgba(142, 50, 47, 0.06), transparent 40%), | |
| radial-gradient(ellipse at bottom right, rgba(61, 74, 92, 0.05), transparent 50%); | |
| margin: 20px auto; | |
| padding: calc(10 * var(--u)) calc(14 * var(--u)); | |
| display: grid; | |
| grid-template-rows: auto auto 1fr auto; /* header | banner | body | footer */ | |
| gap: calc(6 * var(--u)); | |
| box-shadow: var(--shadow-screen); | |
| position: relative; | |
| overflow: hidden; | |
| } | |
| /* Decorative top bar — flat solid accent (de-gradient). */ | |
| .poster::before { | |
| content: ""; | |
| position: absolute; top: 0; left: 0; right: 0; | |
| height: calc(8 * var(--u)); | |
| background: var(--accent); | |
| } | |
| /* ========================================================= | |
| HEADER | |
| ========================================================= */ | |
| .header { | |
| display: grid; | |
| grid-template-columns: 1fr minmax(50%, auto) 1fr; /* equal side tracks (1fr) -> the title track is centred on the poster, not just between the side blocks; the centre track is floored at 50% so a one-line title still fills it (else a short title measures narrow and trips a false HEADER/TITLE-SQUEEZED). Best-effort: a side block wide enough to clamp its 1fr track can still pull the title off-centre. */ | |
| align-items: center; | |
| gap: calc(16 * var(--u)); | |
| padding: calc(2 * var(--u)) calc(4 * var(--u)) calc(5 * var(--u)); | |
| border-bottom: calc(2 * var(--u)) solid var(--accent); | |
| } | |
| .venue-badge { | |
| justify-self: start; /* anchor to the far-left edge so the centre track stays centred */ | |
| display: flex; flex-direction: column; | |
| align-items: center; justify-content: center; | |
| min-width: calc(95 * var(--u)); | |
| text-align: center; | |
| border-right: calc(1 * var(--u)) solid var(--border-soft); | |
| padding-right: calc(12 * var(--u)); | |
| } | |
| .venue-badge .vb-venue { | |
| font-family: var(--font-sans); | |
| font-weight: 800; | |
| font-size: var(--fs-9); | |
| color: var(--accent-deep); | |
| line-height: 1; | |
| letter-spacing: -0.5px; | |
| } | |
| .venue-badge .vb-year { | |
| font-family: var(--font-sans); | |
| font-size: var(--fs-5); | |
| color: var(--text-secondary); | |
| margin-top: calc(3 * var(--u)); | |
| letter-spacing: 1.2px; | |
| } | |
| .venue-badge .vb-tag { | |
| font-family: var(--font-sans); | |
| font-size: var(--fs-2); | |
| color: var(--accent); | |
| font-weight: 700; | |
| margin-top: calc(2 * var(--u)); | |
| letter-spacing: 1.2px; | |
| } | |
| .title-block { text-align: center; min-width: 0; } | |
| .title { | |
| font-family: var(--font-sans); | |
| font-weight: 800; | |
| font-size: var(--fs-9); | |
| line-height: 1.05; | |
| color: var(--accent-deep); | |
| letter-spacing: -0.5px; | |
| } | |
| .title .accent { color: var(--gold); } | |
| .subtitle { | |
| font-family: var(--font-sans); | |
| font-weight: 500; | |
| font-size: var(--fs-6); | |
| color: var(--text-secondary); | |
| margin-top: calc(2 * var(--u)); | |
| font-style: italic; | |
| } | |
| .authors-line { | |
| font-family: var(--font-sans); | |
| font-size: var(--fs-4); | |
| color: var(--accent); | |
| font-weight: 600; | |
| margin-top: calc(3 * var(--u)); | |
| } | |
| .authors-line .author { margin: 0 calc(4 * var(--u)); } | |
| .authors-line sup { font-size: 0.7em; color: var(--accent); } | |
| .authors-line .aff { | |
| color: var(--text-secondary); | |
| font-weight: 400; | |
| display: block; | |
| margin-top: calc(2 * var(--u)); | |
| font-size: var(--fs-4); | |
| } | |
| .right-block { | |
| justify-self: end; /* keep logo+QR in the far-right corner */ | |
| display: flex; align-items: center; | |
| gap: calc(10 * var(--u)); | |
| } | |
| .qr-block { display: flex; flex-direction: column; align-items: center; gap: calc(2 * var(--u)); } | |
| .qr-block img { | |
| width: calc(85 * var(--u)); | |
| height: calc(85 * var(--u)); | |
| border: calc(2 * var(--u)) solid var(--accent); | |
| border-radius: calc(4 * var(--u)); | |
| background: white; | |
| padding: calc(2 * var(--u)); | |
| } | |
| .qr-label { | |
| font-family: var(--font-sans); | |
| font-size: var(--fs-3); | |
| color: var(--accent); | |
| font-weight: 600; | |
| } | |
| /* Optional logo slot — drop your lab logo here. Keep its rendered height | |
| close to the QR (~85u) so the header doesn't grow disproportionately. | |
| A raster logo needs no exemption. An inline-SVG logo with brand colors | |
| must carry data-color-exempt="logo" so style_check skips its fills, e.g.: | |
| <svg data-color-exempt="logo" ...> ... </svg> */ | |
| .logo-slot img { height: calc(85 * var(--u)); width: auto; max-width: calc(360 * var(--u)); object-fit: contain; display: block; } | |
| /* Logo size classes on .logo-slot -- pick from the file's aspect ratio (SKILL.md Gate E). | |
| tall/square sit at the QR-matched slot height; wide is intentionally shorter (~68% of the | |
| QR) and width-capped so a long wordmark doesn't out-mass the title. Do NOT combine these | |
| with logo-stack (that row is width-normalized, below). */ | |
| .logo-slot.logo-tall img, | |
| .logo-slot.logo-square img { height: calc(85 * var(--u)); } | |
| .logo-slot.logo-wide img { height: calc(58 * var(--u)); max-width: calc(300 * var(--u)); } | |
| /* Logo chip: a solid backing that keeps a transparent / edge-white logo legible on a | |
| colored/dark header (white chip) or a light header (.logo-chip-dark); the padding + radius | |
| fold a stray white box into a deliberate rounded tile (SKILL.md Gate E, Background). */ | |
| .logo-chip { | |
| display: inline-flex; align-items: center; justify-content: center; | |
| background: var(--bg-card); border-radius: calc(3 * var(--u)); | |
| padding: calc(3 * var(--u)) calc(5 * var(--u)); | |
| } | |
| .logo-chip.logo-chip-dark { background: var(--text-primary); } | |
| /* logo-row: institution logos in the header right block (REAL logos the | |
| user provided — never fabricate a seal). Each img MUST carry | |
| data-color-exempt="logo"; height pairs with the QR. */ | |
| .logo-row { display: flex; align-items: center; gap: calc(5 * var(--u)); } | |
| .logo-row img { height: calc(68 * var(--u)); width: auto; display: block; } | |
| /* --boxed variant: each logo in a labeled tile (logo + institution name) — | |
| more presence at poster distance; white tile + soft border, no new hues. */ | |
| .logo-row .lr-item { | |
| display: flex; flex-direction: column; align-items: center; | |
| gap: calc(2 * var(--u)); | |
| background: var(--bg-card); | |
| border: 1px solid var(--border-soft); | |
| border-radius: calc(3 * var(--u)); | |
| padding: calc(4 * var(--u)) calc(6 * var(--u)); | |
| } | |
| .logo-row .lr-item img { height: calc(58 * var(--u)); } | |
| .logo-row .lr-label { | |
| font-family: var(--font-sans); font-weight: 600; font-size: var(--fs-1); | |
| color: var(--text-secondary); text-align: center; line-height: 1.15; | |
| } | |
| /* logo-stack variant: WIDE WORDMARKS (AR >= ~2) normalized to EQUAL WIDTH | |
| and stacked vertically, left-aligned. Use when two wide wordmarks of | |
| different aspect ratio read unbalanced height-matched in a row (equal | |
| width lets the less-wide mark grow taller and aligns a clean block). | |
| NOT for a square seal or tall mark (equal width blows it up — keep those | |
| height-matched). Do NOT also apply the logo-wide/tall/square classes. */ | |
| .logo-row.logo-stack { flex-direction: column; align-items: flex-start; gap: calc(8 * var(--u)); } | |
| .logo-row.logo-stack img { width: calc(170 * var(--u)); height: auto; } | |
| /* venue badge may carry the official venue logo above its text line */ | |
| .venue-badge img { height: calc(62 * var(--u)); width: auto; display: block; margin: 0 auto calc(2 * var(--u)); } | |
| .venue-badge .vb-title { font-family: var(--font-sans); font-weight: 800; font-size: var(--fs-5); color: var(--accent-deep); letter-spacing: 0.5px; } | |
| /* ========================================================= | |
| OPTIONAL FRAMEWORK BANNER (delete this whole section if you don't need it) | |
| ========================================================= */ | |
| .framework-banner { | |
| display: flex; | |
| align-items: center; | |
| gap: calc(16 * var(--u)); | |
| /* Flat solid emphasis fill (de-gradient). */ | |
| background: var(--bg-emphasis); | |
| border: calc(1 * var(--u)) solid var(--border-soft); | |
| border-left: calc(6 * var(--u)) solid var(--accent); | |
| border-radius: calc(6 * var(--u)); | |
| padding: calc(6 * var(--u)) calc(14 * var(--u)); | |
| } | |
| .framework-banner img { height: calc(120 * var(--u)); width: auto; display: block; } | |
| .framework-banner .banner-stats { | |
| flex: 1; | |
| display: grid; | |
| grid-template-columns: 1fr 1fr; | |
| gap: calc(6 * var(--u)); | |
| } | |
| .framework-banner .bs-item { | |
| background: white; | |
| border: calc(1 * var(--u)) solid var(--border-soft); | |
| border-left: calc(3 * var(--u)) solid var(--accent); | |
| border-radius: calc(3 * var(--u)); | |
| padding: calc(2 * var(--u)) calc(8 * var(--u)); | |
| text-align: center; | |
| } | |
| .framework-banner .bs-num { | |
| font-family: var(--font-sans); | |
| font-weight: 800; | |
| font-size: var(--fs-8); | |
| color: var(--accent); | |
| line-height: 1; | |
| } | |
| .framework-banner .bs-label { | |
| font-family: var(--font-sans); | |
| font-size: var(--fs-3); | |
| color: var(--text-secondary); | |
| margin-top: calc(2 * var(--u)); | |
| line-height: 1.2; | |
| } | |
| .framework-banner .fb-text { | |
| flex: 1.6; | |
| font-family: var(--font-serif); | |
| font-size: var(--fs-6); | |
| line-height: 1.5; | |
| text-wrap: pretty; /* prose: fill each line, protect last-line orphan. NOT balance — on multi-sentence prose balance shortens+hyphenates line 1 (crammed-left, gap-right). */ | |
| text-align: center; | |
| } | |
| .framework-banner .fb-text strong { color: var(--accent-deep); } | |
| .framework-banner .fb-label { | |
| display: inline-block; | |
| background: var(--accent); | |
| color: white; | |
| font-family: var(--font-sans); | |
| font-size: var(--fs-5); | |
| font-weight: 700; | |
| padding: calc(2 * var(--u)) calc(8 * var(--u)); | |
| border-radius: calc(4 * var(--u)); | |
| text-transform: uppercase; | |
| letter-spacing: 1px; | |
| vertical-align: middle; | |
| line-height: 1; | |
| position: relative; | |
| top: calc(-1 * var(--u)); | |
| } | |
| /* Method-overview figure for the banner (banner-figure component, | |
| catalogued in COMPONENTS.md). Usually CAPTIONLESS -- the banner text | |
| block beside it is the figure's explanation. width:min-content collapses | |
| the slot to the IMAGE (a captionless figure's slot IS the image); if a | |
| short caption is used it wraps at the image box and can NEVER set the | |
| flex-item width; margin-inline centres a block image (text-align does | |
| not); overflow-wrap guards a long unbreakable token. Do NOT hand-roll a | |
| .fb-fig wrapper or a bare <img class="w-100"> here. */ | |
| .framework-banner .banner-figure { | |
| flex: 0 0 auto; | |
| width: min-content; | |
| margin: 0; | |
| text-align: center; | |
| } | |
| .framework-banner .banner-figure img { | |
| height: calc(120 * var(--u)); | |
| width: auto; | |
| display: block; | |
| margin-inline: auto; | |
| } | |
| .framework-banner .banner-figure figcaption { | |
| width: 100%; | |
| margin-top: calc(2 * var(--u)); | |
| font-family: var(--font-sans); | |
| font-size: var(--fs-2); | |
| line-height: 1.2; | |
| color: var(--text-secondary); | |
| text-align: center; | |
| text-wrap: pretty; | |
| overflow-wrap: anywhere; | |
| } | |
| .framework-banner .banner-figure figcaption strong { color: var(--accent-deep); } | |
| /* ========================================================= | |
| BODY: 4 columns, variable-width hint via fr values | |
| ========================================================= */ | |
| .body-grid { | |
| display: grid; | |
| grid-template-columns: 1fr 1.05fr 1.05fr 1fr; | |
| gap: calc(10 * var(--u)); | |
| /* No overflow:hidden here — it would clip card shadows (pitfall #6). | |
| .poster already clips at the page boundary; min-height:0 is the | |
| grid-blowout guard. */ | |
| min-height: 0; | |
| } | |
| .column { | |
| display: flex; flex-direction: column; | |
| gap: calc(6 * var(--u)); | |
| min-height: 0; | |
| height: 100%; | |
| padding-bottom: calc(4 * var(--u)); /* shadow breathing room above takeaways */ | |
| } | |
| /* ========================================================= | |
| CARDS | |
| ========================================================= */ | |
| .card { | |
| background: var(--bg-card); | |
| border-radius: calc(5 * var(--u)); | |
| padding: calc(4 * var(--u)) calc(9 * var(--u)); | |
| border: calc(1 * var(--u)) solid var(--border-soft); | |
| box-shadow: var(--shadow-card); | |
| position: relative; | |
| } | |
| .card.tinted { background: var(--bg-card-tint); } | |
| .card.card--compact { padding: calc(3 * var(--u)) calc(6 * var(--u)); } /* predefined variant: tighter padding (fix (f)) */ | |
| .card.highlight { | |
| border-left: calc(6 * var(--u)) solid var(--accent); | |
| /* Flat emphasis tint (de-gradient). */ | |
| background: var(--bg-emphasis); | |
| } | |
| .section-title { | |
| font-family: var(--font-sans); | |
| font-weight: 700; | |
| font-size: var(--fs-7); | |
| color: var(--accent-deep); | |
| margin-bottom: calc(3 * var(--u)); | |
| display: flex; align-items: center; | |
| gap: calc(5 * var(--u)); | |
| } | |
| /* Title text + any ★ marker share ONE .st-text span, so the heading wraps as | |
| natural text (hanging indent) instead of flex-wrapping atomic items: the | |
| number badge is never stranded on its own line and the ★ never widows. */ | |
| .section-title .st-text { flex: 1; min-width: 0; line-height: 1.18; } | |
| /* Graceful fallback if a title is NOT wrapped in .st-text: float the badge so | |
| bare inline text still wraps beside it rather than dropping below it. */ | |
| .section-title:not(:has(.st-text)) { display: block; line-height: 1.18; } | |
| .section-title:not(:has(.st-text)) .num { float: left; margin-right: calc(5 * var(--u)); } | |
| .section-title .num { | |
| display: inline-flex; align-items: center; justify-content: center; | |
| width: calc(22 * var(--u)); height: calc(22 * var(--u)); | |
| background: var(--accent); color: white; | |
| border-radius: 50%; | |
| font-size: var(--fs-5); font-weight: 700; | |
| flex-shrink: 0; | |
| } | |
| /* Inline "★ KEY" / "★ Headline" marker beside a section title. */ | |
| .section-title .key-mark { color: var(--gold); font-size: var(--fs-3); } | |
| .body-text, .card p, .card li { | |
| font-family: var(--font-serif); | |
| font-size: var(--fs-4); | |
| line-height: 1.3; | |
| color: var(--text-primary); | |
| } | |
| .card ul, .card ol { padding-left: calc(18 * var(--u)); } | |
| .card li { margin-bottom: calc(2 * var(--u)); } | |
| .keyword { color: var(--accent); font-weight: 700; } | |
| .keyword-gold { color: var(--gold); font-weight: 700; } | |
| .highlight-text { | |
| background: var(--bg-emphasis); | |
| padding: 0 calc(3 * var(--u)); | |
| border-radius: calc(2 * var(--u)); | |
| } | |
| /* Equation block */ | |
| .eqn { | |
| background: var(--bg-emphasis); | |
| border-left: calc(3 * var(--u)) solid var(--accent); | |
| padding: calc(4 * var(--u)) calc(10 * var(--u)); | |
| margin: calc(4 * var(--u)) 0; | |
| font-size: var(--fs-5); | |
| overflow-x: hidden; | |
| } | |
| /* Predefined variant for a larger / more legible display equation. | |
| `calc(var(--fs-N) * k)` is allowed here because this is a | |
| COMPONENTS.md-registered variant (DESIGN_FINAL §A / §10 f). */ | |
| .eqn--large { font-size: calc(var(--fs-5) * 1.25); } | |
| .eqn .label { | |
| display: block; | |
| font-family: var(--font-sans); | |
| font-size: var(--fs-2); | |
| color: var(--accent); | |
| font-weight: 600; | |
| margin-bottom: calc(2 * var(--u)); | |
| text-transform: uppercase; | |
| letter-spacing: 1px; | |
| } | |
| /* Callout: solid accent for primary, solid gold for theorems / "★ key" strips */ | |
| .callout { | |
| background: var(--accent); | |
| color: white; | |
| padding: calc(5 * var(--u)) calc(10 * var(--u)); | |
| border-radius: calc(4 * var(--u)); | |
| font-size: var(--fs-4); | |
| margin: calc(4 * var(--u)) 0; | |
| } | |
| .callout strong { color: var(--gold-soft); } | |
| /* Flat solid gold (de-gradient). */ | |
| .callout.gold { | |
| background: var(--gold-soft); | |
| color: var(--gold); | |
| } | |
| .callout.gold strong { color: var(--gold); } | |
| /* Figure container */ | |
| .figure { margin: calc(4 * var(--u)) 0; text-align: center; } | |
| .figure img:not([class*="w-"]) { width: 100%; } | |
| .figure--wide img { width: 100%; } /* predefined variant: force full card width over any .w-NN (fix (f)) */ | |
| .figure img { | |
| border-radius: calc(4 * var(--u)); | |
| border: calc(1 * var(--u)) solid var(--border-soft); | |
| background: white; | |
| } | |
| .figure .caption { | |
| font-family: var(--font-sans); | |
| font-size: var(--fs-3); | |
| color: var(--text-secondary); | |
| margin-top: calc(3 * var(--u)); | |
| line-height: 1.3; | |
| text-align: left; | |
| } | |
| .figure .caption strong { color: var(--accent-deep); } | |
| /* Float a figure BESIDE text in a TEXT-RICH card: text wraps to its side | |
| and then below it. The clearfix grows the card to contain the float; | |
| mark the <img> data-fig-layout="beside-text" so the AR gates honour the | |
| intentionally small width. Use ONLY when there is enough text to fill the | |
| figure's height -- a text-sparse card should center .figure instead. A | |
| short-text float leaves an L-shaped void below the text, which polish's | |
| FIG/BESIDE-TEXT-VOID flags. */ | |
| .fig-wrap::after { content: ""; display: table; clear: both; } | |
| .ff-fig { | |
| float: right; | |
| width: 48%; max-width: 58%; min-width: 38%; | |
| margin: calc(1 * var(--u)) 0 calc(3 * var(--u)) calc(11 * var(--u)); | |
| text-align: center; | |
| } | |
| .ff-fig.left { | |
| float: left; | |
| margin: calc(1 * var(--u)) calc(11 * var(--u)) calc(3 * var(--u)) 0; | |
| } | |
| .ff-fig img { | |
| display: block; | |
| width: 100%; | |
| border-radius: calc(4 * var(--u)); | |
| border: calc(1 * var(--u)) solid var(--border-soft); | |
| background: white; | |
| } | |
| .ff-fig .caption { | |
| font-family: var(--font-sans); | |
| font-size: var(--fs-3); | |
| color: var(--text-secondary); | |
| margin-top: calc(3 * var(--u)); | |
| line-height: 1.3; | |
| text-align: center; | |
| } | |
| /* Result table with .ours row highlighted gold */ | |
| .result-table { | |
| width: 100%; | |
| border-collapse: collapse; | |
| font-family: var(--font-sans); | |
| font-size: var(--fs-3); | |
| margin-top: calc(3 * var(--u)); | |
| } | |
| .result-table th, .result-table td { | |
| padding: calc(2 * var(--u)) calc(4 * var(--u)); | |
| text-align: center; | |
| border-bottom: calc(1 * var(--u)) solid var(--border-soft); | |
| } | |
| .result-table thead th { | |
| background: var(--accent); color: white; | |
| font-weight: 600; font-size: var(--fs-2); | |
| } | |
| .result-table tbody tr.group-row td { | |
| background: var(--bg-emphasis); font-weight: 700; | |
| text-align: left; | |
| color: var(--accent-deep); | |
| padding-left: calc(8 * var(--u)); | |
| border-bottom: calc(2 * var(--u)) solid var(--accent); | |
| } | |
| .result-table tbody tr.ours td { background: var(--gold-soft); font-weight: 700; } | |
| .result-table tbody tr.ours td:first-child { color: var(--accent-deep); } | |
| /* Greyed reference row (was inline color:#888). */ | |
| .result-table tbody tr.reference td { color: var(--text-muted); } | |
| .result-table .method { text-align: left; padding-left: calc(8 * var(--u)); } | |
| .result-table .best { color: var(--accent); font-weight: 700; } | |
| /* 3-up stat box */ | |
| .keybox { | |
| display: grid; | |
| grid-template-columns: repeat(3, 1fr); | |
| gap: calc(4 * var(--u)); | |
| margin: calc(4 * var(--u)) 0 0; | |
| } | |
| .keybox .kb-item { | |
| background: var(--bg-emphasis); | |
| border-top: calc(2 * var(--u)) solid var(--accent); | |
| padding: calc(3 * var(--u)); | |
| text-align: center; | |
| /* The grid stretches every tile to the tallest one; center the content | |
| vertically so a 1-line tile's number aligns with a 2-line neighbour's | |
| instead of top-ragged (SKILL Layout pitfall — stat-tile alignment). */ | |
| display: flex; flex-direction: column; justify-content: center; | |
| border-radius: 0 0 calc(3 * var(--u)) calc(3 * var(--u)); | |
| } | |
| .kb-item .kb-num { | |
| font-family: var(--font-sans); | |
| font-weight: 800; | |
| font-size: var(--fs-6); | |
| color: var(--accent); | |
| line-height: 1; | |
| } | |
| .kb-item .kb-label { | |
| font-family: var(--font-sans); | |
| font-size: var(--fs-1); | |
| color: var(--text-secondary); | |
| margin-top: calc(2 * var(--u)); | |
| line-height: 1.1; | |
| } | |
| /* Optional takeaways strip (delete the section if not needed) */ | |
| .takeaways-strip { | |
| display: grid; | |
| grid-template-columns: auto repeat(4, 1fr); | |
| align-items: center; | |
| gap: calc(10 * var(--u)); | |
| /* Flat solid emphasis fill (de-gradient). */ | |
| background: var(--bg-emphasis); | |
| border: calc(1 * var(--u)) solid var(--border-soft); | |
| border-radius: calc(5 * var(--u)); | |
| padding: calc(8 * var(--u)) calc(12 * var(--u)); | |
| } | |
| .takeaways-strip .ts-title { | |
| font-family: var(--font-sans); | |
| font-size: var(--fs-6); | |
| font-weight: 800; | |
| color: var(--accent-deep); | |
| display: flex; align-items: center; gap: calc(6 * var(--u)); | |
| } | |
| .takeaways-strip .ts-title .num { | |
| display: inline-flex; align-items: center; justify-content: center; | |
| width: calc(22 * var(--u)); height: calc(22 * var(--u)); | |
| background: var(--accent); color: white; | |
| border-radius: 50%; | |
| font-size: var(--fs-5); font-weight: 700; | |
| } | |
| .takeaways-strip .ts-item { | |
| border-left: calc(3 * var(--u)) solid var(--accent); | |
| padding-left: calc(8 * var(--u)); | |
| line-height: 1.3; | |
| text-align: center; | |
| text-wrap: balance; | |
| } | |
| .takeaways-strip .ts-key { | |
| font-family: var(--font-sans); | |
| font-size: var(--fs-3); | |
| font-weight: 700; | |
| color: var(--accent); | |
| text-transform: uppercase; | |
| letter-spacing: 1px; | |
| } | |
| .takeaways-strip .ts-text { | |
| font-family: var(--font-serif); | |
| font-size: var(--fs-4); | |
| margin-left: calc(4 * var(--u)); | |
| } | |
| /* Footer */ | |
| .footer { | |
| grid-column: 1 / -1; | |
| display: flex; justify-content: space-between; align-items: baseline; | |
| /* When the two blocks can't sit side by side they stack (wrap) instead of | |
| overflowing; row-gap keeps the stacked blocks legible. */ | |
| flex-wrap: wrap; gap: calc(2 * var(--u)) calc(10 * var(--u)); | |
| padding-top: calc(8 * var(--u)); | |
| border-top: calc(1 * var(--u)) solid var(--border-soft); | |
| font-family: var(--font-sans); | |
| font-size: var(--fs-5); | |
| color: var(--text-muted); | |
| } | |
| .footer .method-name { color: var(--accent-deep); } | |
| /* Long repo URL / email breaks mid-token rather than overflowing the edge. */ | |
| .footer .repo { color: var(--accent); font-weight: 600; overflow-wrap: anywhere; } | |
| /* Optional watermark */ | |
| .ornament { | |
| position: absolute; | |
| right: calc(20 * var(--u)); | |
| bottom: calc(20 * var(--u)); | |
| font-family: var(--font-sans); | |
| font-size: var(--fs-9); | |
| color: var(--ornament-ink); | |
| font-weight: 900; | |
| letter-spacing: 4px; | |
| pointer-events: none; | |
| user-select: none; | |
| } | |
| /* ========================================================= | |
| DENSITY COMPONENTS (catalogued — see COMPONENTS.md). | |
| Guardrail: no component-local color semantics. Distinction is | |
| carried by labels, order and typography — never by new hues. | |
| ========================================================= */ | |
| /* equation-stack: compact multi-row formula stack (denser than one big .eqn) */ | |
| .equation-stack { margin: calc(2 * var(--u)) 0; } | |
| .equation-stack .eqn { margin: calc(2 * var(--u)) 0; padding: calc(2 * var(--u)) calc(8 * var(--u)); } | |
| /* eqn-anatomy: term-by-term anatomy grid (2x2 default; --row variant = 1x4) */ | |
| .eqn-anatomy { | |
| display: grid; grid-template-columns: 1fr 1fr; | |
| gap: calc(2.5 * var(--u)); margin: calc(3 * var(--u)) 0; | |
| } | |
| .eqn-anatomy.eqn-anatomy--row { grid-template-columns: repeat(4, 1fr); } | |
| .eqn-anatomy .ea-item { | |
| background: var(--bg-card-tint); | |
| border: 1px solid var(--border-soft); | |
| border-left: calc(2 * var(--u)) solid var(--accent); | |
| border-radius: calc(2 * var(--u)); | |
| padding: calc(2 * var(--u)) calc(4 * var(--u)); | |
| font-size: var(--fs-2); line-height: 1.3; | |
| } | |
| .eqn-anatomy .ea-tag { | |
| display: inline-block; font-family: var(--font-sans); font-weight: 700; | |
| font-size: var(--fs-1); color: var(--accent-deep); | |
| background: var(--accent-light); border-radius: calc(1.5 * var(--u)); | |
| padding: calc(0.5 * var(--u)) calc(2.5 * var(--u)); | |
| margin-right: calc(1.5 * var(--u)); | |
| } | |
| /* flow-strip: labeled pipeline, ALL steps the same accent; ONLY the | |
| final step may carry the gold top bar. Not an "algorithm" unless the | |
| paper has one — see COMPONENTS.md. */ | |
| .flow-strip { display: flex; align-items: stretch; gap: calc(1.5 * var(--u)); margin: calc(3 * var(--u)) 0; } | |
| .flow-strip .step { | |
| flex: 1; background: var(--accent-light); | |
| border: 1px solid var(--accent-soft); border-radius: calc(2 * var(--u)); | |
| padding: calc(2 * var(--u)) calc(2.5 * var(--u)); | |
| text-align: center; font-size: var(--fs-2); line-height: 1.25; | |
| } | |
| .flow-strip .step .step-name { | |
| display: block; font-family: var(--font-sans); font-weight: 700; | |
| font-size: var(--fs-1); color: var(--accent-deep); | |
| letter-spacing: 0.5px; text-transform: uppercase; | |
| margin-bottom: calc(1 * var(--u)); | |
| } | |
| .flow-strip .step--final { border-top: calc(1.5 * var(--u)) solid var(--gold); background: var(--gold-soft); } | |
| .flow-strip .arrow { | |
| align-self: center; color: var(--accent); | |
| font-family: var(--font-sans); font-weight: 700; font-size: var(--fs-4); | |
| flex: 0 0 auto; | |
| } | |
| /* figure--duo: two paper figures sharing one caption; each img MUST be | |
| a data-source="paper" asset and carry .w-45/.w-50 (42-48% each). */ | |
| .figure--duo { display: flex; gap: calc(3 * var(--u)); align-items: flex-start; justify-content: center; } | |
| .figure--duo img { max-width: 48%; min-width: 0; } /* hard cap: duo contract is 42-48% each; an oversized w-NN cannot overflow the card */ | |
| /* result-table derived column (gold-soft = DERIVED arithmetic; label it) */ | |
| .result-table th.derived, .result-table td.derived { | |
| background: var(--gold-soft); font-family: var(--font-sans); font-weight: 700; | |
| } | |
| /* keybox 4-up variant */ | |
| .keybox.keybox--4 { grid-template-columns: repeat(4, 1fr); } | |
| /* algo: compact numbered procedure — ONLY when the paper itself states | |
| an explicit algorithm/procedure; never invent steps. */ | |
| ol.algo { padding-left: calc(16 * var(--u)); } | |
| ol.algo li { margin-bottom: calc(1.5 * var(--u)); font-size: var(--fs-3); line-height: 1.3; } | |
| /* claim-pills: provenance mini-table (numeric-heavy posters only) */ | |
| .claim-pills { width: 100%; border-collapse: collapse; font-family: var(--font-sans); font-size: var(--fs-2); } | |
| .claim-pills td { | |
| border-bottom: 1px solid var(--border-soft); | |
| padding: calc(1.5 * var(--u)) calc(3 * var(--u)); | |
| } | |
| .claim-pills .cp-id { | |
| font-weight: 700; color: var(--accent-deep); background: var(--accent-light); | |
| border-radius: calc(1.5 * var(--u)); padding: 0 calc(2.5 * var(--u)); white-space: nowrap; | |
| } | |
| .claim-pills .cp-fact { color: var(--accent-deep); font-weight: 700; } | |
| .claim-pills .cp-derived { color: var(--gold); font-weight: 700; } | |
| /* Grid-blowout guards: (1) columns may shrink below min-content so a | |
| wide equation can never widen its track and squeeze siblings; | |
| (2) over-wide display math scales down to fit its box instead of | |
| blowing out the card (SVG keeps aspect via height:auto). */ | |
| .body-grid > .column { min-width: 0; } | |
| .card { min-width: 0; } | |
| .eqn mjx-container > svg { max-width: 100%; height: auto; } | |
| /* ── Utility classes — defined LAST so equal-specificity ties | |
| resolve in the utility's favor (source order). ── */ | |
| .fs-1 { font-size: var(--fs-1); } | |
| .fs-2 { font-size: var(--fs-2); } | |
| .fs-3 { font-size: var(--fs-3); } | |
| .fs-4 { font-size: var(--fs-4); } | |
| .fs-5 { font-size: var(--fs-5); } | |
| .fs-6 { font-size: var(--fs-6); } | |
| .fs-7 { font-size: var(--fs-7); } | |
| .fs-8 { font-size: var(--fs-8); } | |
| .fs-9 { font-size: var(--fs-9); } | |
| .mt-1 { margin-top: calc(1 * var(--u)); } | |
| .mt-2 { margin-top: calc(2 * var(--u)); } | |
| .mt-3 { margin-top: calc(3 * var(--u)); } | |
| .mt-4 { margin-top: calc(4 * var(--u)); } | |
| .mt-5 { margin-top: calc(5 * var(--u)); } | |
| .mt-6 { margin-top: calc(6 * var(--u)); } | |
| .mb-1 { margin-bottom: calc(1 * var(--u)); } | |
| .mb-2 { margin-bottom: calc(2 * var(--u)); } | |
| .mb-3 { margin-bottom: calc(3 * var(--u)); } | |
| .mb-4 { margin-bottom: calc(4 * var(--u)); } | |
| .w-45 { width: 45%; } | |
| .w-50 { width: 50%; } | |
| .w-55 { width: 55%; } | |
| .w-60 { width: 60%; } | |
| .w-65 { width: 65%; } | |
| .w-70 { width: 70%; } | |
| .w-75 { width: 75%; } | |
| .w-80 { width: 80%; } | |
| .w-85 { width: 85%; } | |
| .w-90 { width: 90%; } | |
| .w-95 { width: 95%; } | |
| .w-100 { width: 100%; } | |
| /* Sub-line column baseline alignment (posterly SKILL: padding-bottom on a | |
| column's last normal-flow card is the continuous lever of last resort). */ | |
| .pb-col-a { padding-bottom: calc(22.3 * var(--u)); } | |
| .pb-col-b { padding-bottom: calc(11.3 * var(--u)); } | |
| .pb-col-c { padding-bottom: calc(11.9 * var(--u)); } | |
| .pb-col-d { padding-bottom: calc(28.4 * var(--u)); } | |
| .text-secondary { color: var(--text-secondary); } | |
| .text-muted { color: var(--text-muted); } | |
| .nowrap { white-space: nowrap; } | |
| .text-center { text-align: center; } | |
| /* ========================================================= | |
| PRINT OVERRIDE — KEEP LAST so it wins source-order ties. | |
| ========================================================= */ | |
| @media print { | |
| html, body { background: white; } | |
| .poster { margin: 0; box-shadow: none; width: 60in; height: 36in; page-break-after: avoid; } | |
| :root { --u: 1mm; } | |
| } | |
| </style> | |
| </head> | |
| <body> | |
| <div class="poster" data-measure-role="poster"> | |
| <!-- =========================== HEADER =========================== --> | |
| <header class="header" data-measure-role="header"> | |
| <!-- LEFT: venue badge (venue is named: ICML 2026, spotlight). --> | |
| <div class="venue-badge"> | |
| <div class="vb-venue">ICML</div> | |
| <div class="vb-year">2026</div> | |
| <div class="vb-tag">SPOTLIGHT · REPRO</div> | |
| </div> | |
| <!-- CENTER: title block --> | |
| <div class="title-block"> | |
| <h1 class="title">Stable Deep RL via Isotropic Gaussian Representations <span class="accent fs-8 nowrap">[ Independent Reproduction ]</span></h1> | |
| <div class="subtitle">Does shaping representations toward an isotropic Gaussian stabilize learning under non-stationarity? We re-derive the theory, re-run the CIFAR-10 study, and audit the RL pipeline.</div> | |
| <div class="authors-line"> | |
| <span class="author">Paper: Saheb Pasand, Obando-Ceron, Courville, Bashivan, Castro</span> <span class="aff">Mila · McGill · UdeM · Google DeepMind</span> · | |
| <span class="author">Reproduction: ProCreations<sup>✉</sup></span> <span class="aff">HF Agent Reproduction Challenge</span> | |
| </div> | |
| </div> | |
| <!-- RIGHT: empty spacer keeps the title track centred (no logo / QR assets provided). --> | |
| <div class="right-block"></div> | |
| </header> | |
| <!-- ===================== FRAMEWORK BANNER ===================== --> | |
| <section class="framework-banner" data-measure-role="banner"> | |
| <div class="fb-text"> | |
| <span class="fb-label">Reproduction</span> | |
| <strong>SIGReg</strong> shapes deep-RL representations toward an isotropic Gaussian — claimed to stabilize learning under drifting targets. We verified the theory, re-ran the non-stationary CIFAR-10 grid on one T4, and audited the released RL pipeline line-by-line; all runs exit-0 and logged. Every deviation is documented, and the paper's ambiguous Table-1 units were resolved. | |
| </div> | |
| <div class="banner-stats"> | |
| <div class="bs-item"><div class="bs-num">24/24</div><div class="bs-label">directional effects<br>reproduce (C5)</div></div> | |
| <div class="bs-item"><div class="bs-num">1e-6</div><div class="bs-label">median rel. err,<br>Thm 3.1 check</div></div> | |
| <div class="bs-item"><div class="bs-num">5/6</div><div class="bs-label">claims verified<br>(1 partial)</div></div> | |
| <div class="bs-item"><div class="bs-num">$0.57</div><div class="bs-label">total paid<br>compute</div></div> | |
| </div> | |
| </section> | |
| <!-- ============================ BODY ============================ --> | |
| <div class="body-grid" data-measure-role="body"> | |
| <!-- ============ COLUMN 1 ============ --> | |
| <div class="column" data-measure-role="column"> | |
| <div class="card highlight" data-measure-role="card" data-logbook-target="feasibility-audit-and-evidence-plan" data-logbook-label="Feasibility audit & evidence plan"> | |
| <div class="section-title"><span class="num">1</span><span class="st-text">The Paper & Our Verdicts</span></div> | |
| <p class="body-text"> | |
| Deep RL targets drift as the policy improves. The paper argues <span class="keyword">isotropic Gaussian embeddings</span> make tracking drifting targets provably stable, enforced by a sliced characteristic-function regularizer (<span class="highlight-text">SIGReg</span>, from LeJEPA). | |
| </p> | |
| <table class="claim-pills mt-3"> | |
| <tr><td class="cp-id">C1</td><td>Thm 3.1 tracking-error dynamics</td><td class="cp-fact">VERIFIED</td></tr> | |
| <tr><td class="cp-id">C2</td><td>Isotropy → uniform contraction, bounded drift</td><td class="cp-fact">VERIFIED</td></tr> | |
| <tr><td class="cp-id">C3</td><td>Gaussianity controls higher moments & tails</td><td class="cp-fact">VERIFIED</td></tr> | |
| <tr><td class="cp-id">C4</td><td>SIGReg shapes reps toward isotropic Gaussian</td><td class="cp-fact">VERIFIED*</td></tr> | |
| <tr><td class="cp-id">C5</td><td>Non-stationary CIFAR-10 Table 1</td><td class="cp-fact">VERIFIED (2/5 seeds)</td></tr> | |
| <tr><td class="cp-id">C6</td><td>Gains across RL domains (Atari, Isaac)</td><td class="cp-derived">PARTIAL</td></tr> | |
| </table> | |
| <p class="body-text mt-2 fs-2 text-secondary">*limits quantified · C6: mechanism verified, RL outcomes audited only.</p> | |
| </div> | |
| <div class="card pb-col-a" data-measure-role="card" data-logbook-target="claim-1-theorem-3-1-tracking-error-dynamics" data-logbook-label="Claim 1 evidence"> | |
| <div class="section-title"><span class="num">2</span><span class="st-text">Theory Check: Tracking-Error Identity</span></div> | |
| <p class="body-text">Gradient-flow critic, drifting $b_t$, fixed $\Sigma_\phi \succ 0$:</p> | |
| <div class="eqn"> | |
| <span class="label">Theorem 3.1</span> | |
| $$\dot\Gamma \;=\; \underbrace{-4\,e^\top \Sigma_\phi\, e}_{\text{contraction}} \;\underbrace{-\;2\,e^\top \Sigma_\phi^{-1}\dot b_t}_{\text{drift}}, \qquad \Gamma = \|e\|_2^2$$ | |
| </div> | |
| <div class="figure mt-3"> | |
| <img src="assets/claim1_identity.png" data-repro-generated="true" data-asset-id="claim1" class="w-75" alt="Analytic identity vs finite-difference simulation"> | |
| <div class="caption">RK4, d=8: analytic RHS (dots) vs finite differences (line); median rel. err ≤ 1.5e-6, incl. Eq. 26's dropped third term.</div> | |
| </div> | |
| <div class="callout mt-3"> | |
| <strong>Precision:</strong> Eq. 33's κ bound is loose; the sharp $1/\lambda_{\min}$ bound is minimized iff isotropic. | |
| </div> | |
| </div> | |
| </div> | |
| <!-- ============ COLUMN 2 ============ --> | |
| <div class="column" data-measure-role="column"> | |
| <div class="card" data-measure-role="card" data-logbook-target="claim-2-isotropy-equalizes-contraction-and-reduces-drift-amplification" data-logbook-label="Claim 2 evidence"> | |
| <div class="section-title"><span class="num">3</span><span class="st-text">Isotropy in Simulation (App. F)</span></div> | |
| <p class="body-text">Online SGD tracks targets switching at steps 100/200; features share one variance budget $\mathrm{Tr}(\Sigma) = d$ and one lr.</p> | |
| <div class="figure mt-3"> | |
| <img src="assets/claim2_row.png" data-repro-generated="true" data-asset-id="claim2" class="w-100" alt="Tracking stability across condition numbers"> | |
| <div class="caption">Isotropic Gaussian (κ=1) re-converges fastest after each switch: mean recovery 3.9 vs 86.4 steps for κ=100, with 3× fewer transient divergences (d=2, 50 seeds; holds for d=8, 32).</div> | |
| </div> | |
| </div> | |
| <div class="card" data-measure-role="card" data-logbook-target="claim-3-gaussianity-controls-higher-order-moments-and-tails" data-logbook-label="Claim 3 evidence"> | |
| <div class="section-title"><span class="num">4</span><span class="st-text">Gaussianity Controls the Tails</span></div> | |
| <p class="body-text">Three checks at <em>identical covariance</em> (4M samples each):</p> | |
| <ul class="mt-3 fs-4"> | |
| <li><strong>Isserlis:</strong> Gaussian 4th moments match covariance to ≤0.4%; same-covariance Laplace deviates up to 100%.</li> | |
| <li><strong>Stein residual</strong> (drives drift variance, Eq. 55): ≤0.1% Gaussian vs 8–55% Laplace / Student-t.</li> | |
| <li><strong>Tails:</strong> isotropic Laplace features cause 2.4× more error spikes, 2× larger p99 updates.</li> | |
| </ul> | |
| <div class="callout mt-3"> | |
| Covariance can't control higher moments without Gaussianity. | |
| </div> | |
| </div> | |
| <div class="card highlight pb-col-b" data-measure-role="card" data-logbook-target="claim-4-sigreg-shapes-representations-toward-isotropic-gaussian" data-logbook-label="Claim 4 audit"> | |
| <div class="section-title"><span class="num">5</span><span class="st-text">SIGReg: Verbatim Audit <span class="key-mark">★ KEY</span></span></div> | |
| <p class="body-text">Identical regularizer in cifar.py / pqn.py / ppo.py: 16 random unit slices per step, empirical CF on 8 frequencies matched to $\mathcal N(0,1)$ — one additive loss term, policy/value objectives untouched.</p> | |
| <div class="eqn"> | |
| <span class="label">Sliced CF loss (per slice)</span> | |
| $$\mathcal{L} = \tfrac{1}{8}\textstyle\sum_{t}\big[(\widehat{\mathrm{Re}}\,\varphi(t) - e^{-t^2/2})^2 + \widehat{\mathrm{Im}}\,\varphi(t)^2\big],\;\; t \in [-5, 5]$$ | |
| </div> | |
| <p class="body-text mt-2 fs-3">Matches Table 8; baseline = λ 0.</p> | |
| </div> | |
| </div> | |
| <!-- ============ COLUMN 3 ============ --> | |
| <div class="column" data-measure-role="column"> | |
| <div class="card" data-measure-role="card" data-logbook-target="claim-4-sigreg-shapes-representations-toward-isotropic-gaussian" data-logbook-label="Claim 4 mechanism evidence"> | |
| <div class="section-title"><span class="num">6</span><span class="st-text">The Mechanism, In Real Training</span></div> | |
| <p class="body-text fs-3 text-secondary mb-1">Probe-batch diagnostics from the 12 GPU runs (Adam shown; RAdam/Kron alike).</p> | |
| <div class="figure"> | |
| <img src="assets/claim4_network.png" data-repro-generated="true" data-asset-id="claim4" class="w-100" alt="Network-in-the-loop SIGReg evidence"> | |
| <div class="caption">CF distance to $\mathcal N(0, I)$: 0.75→0.06; top-eigenvalue share 1.00→0.28; RankMe 1.0→12.6 — baselines collapse to near-1-D; SIGReg keeps variance spread.</div> | |
| </div> | |
| <div class="callout gold mt-3"> | |
| <strong>Quantified limits:</strong> at D=512 the sliced test is near-blind per batch to moderate anisotropy and coordinate-wise heavy tails (CLT along random dense directions; axis-aligned probes detect both) — yet descent still drives κ 279→35, excess kurtosis 2.8→0.16, and recovers rank 8→366/512 from collapse. | |
| </div> | |
| </div> | |
| <div class="card pb-col-c" data-measure-role="card" data-logbook-target="claim-5-non-stationary-cifar-10-table-1-adam-radam-kron" data-logbook-label="Table 1 comparison"> | |
| <div class="section-title"><span class="num">7</span><span class="st-text">Table 1: Measured vs Reported</span></div> | |
| <p class="body-text fs-2 text-secondary mb-1">AUC = 100 × time-averaged metric (units resolved here, card 9); ours = 2-seed mean, paper in grey.</p> | |
| <table class="result-table"> | |
| <thead> | |
| <tr> | |
| <th class="method">Method</th> | |
| <th>Acc ↑</th> | |
| <th>SIGReg ↓</th> | |
| <th>Rank ↑</th> | |
| <th>Dorm% ↓</th> | |
| </tr> | |
| </thead> | |
| <tbody> | |
| <tr><td class="method">Adam</td><td>21.3 <span class="text-muted">(33.0)</span></td><td>64.2 <span class="text-muted">(49.1)</span></td><td>196 <span class="text-muted">(149)</span></td><td>28.6 <span class="text-muted">(18.5)</span></td></tr> | |
| <tr class="ours"><td class="method">+SIGReg</td><td class="best">29.2 <span class="text-muted">(50.0)</span></td><td class="best">9.5 <span class="text-muted">(7.6)</span></td><td class="best">1029 <span class="text-muted">(263)</span></td><td class="best">1.7 <span class="text-muted">(0.4)</span></td></tr> | |
| <tr><td class="method">RAdam</td><td>22.7 <span class="text-muted">(29.8)</span></td><td>58.5 <span class="text-muted">(42.3)</span></td><td>278 <span class="text-muted">(149)</span></td><td>14.1 <span class="text-muted">(12.6)</span></td></tr> | |
| <tr class="ours"><td class="method">+SIGReg</td><td class="best">29.9 <span class="text-muted">(42.8)</span></td><td class="best">8.3 <span class="text-muted">(8.4)</span></td><td class="best">1320 <span class="text-muted">(252)</span></td><td class="best">0.4 <span class="text-muted">(0.6)</span></td></tr> | |
| <tr><td class="method">Kron</td><td>39.6 <span class="text-muted">(62.6)</span></td><td>51.5 <span class="text-muted">(28.5)</span></td><td>413 <span class="text-muted">(167)</span></td><td>24.2 <span class="text-muted">(7.9)</span></td></tr> | |
| <tr class="ours"><td class="method">+SIGReg</td><td class="best">66.0 <span class="text-muted">(72.7)</span></td><td class="best">6.7 <span class="text-muted">(7.2)</span></td><td class="best">2370 <span class="text-muted">(252)</span></td><td class="best">1.1 <span class="text-muted">(0.9)</span></td></tr> | |
| </tbody> | |
| </table> | |
| <p class="body-text mt-2 fs-3"> | |
| SIGReg-arm loss and dormancy land within ~2 units of the paper (Kron+SIGReg: 6.7 vs 7.2; 1.1 vs 0.9). Cross-optimizer orderings reproduce, including Kron's “implicit isotropy” (lowest baseline SIGReg loss). Kron runs cost ~4× Adam's wall-clock. Per-seed CSVs and covariance spectra are in the logbook artifact bundle. | |
| </p> | |
| </div> | |
| </div> | |
| <!-- ============ COLUMN 4 ============ --> | |
| <div class="column" data-measure-role="column"> | |
| <div class="card highlight" data-measure-role="card" data-logbook-target="claim-5-non-stationary-cifar-10-table-1-adam-radam-kron" data-logbook-label="Claim 5 evidence"> | |
| <div class="section-title"><span class="num">8</span><span class="st-text">Non-Stationary CIFAR-10 <span class="key-mark">★</span></span></div> | |
| <p class="body-text fs-3 text-secondary mb-1"> | |
| Authors' protocol on 1×T4: first 10k CIFAR-10 images (canonical order, md5-verified), labels re-permuted every 20 of 100 epochs; 3 optimizers × {baseline, +SIGReg} × 2 seeds. | |
| </p> | |
| <div class="figure"> | |
| <img src="assets/claim5_curves.png" data-repro-generated="true" data-asset-id="claim5" class="w-100" alt="Reproduced non-stationary CIFAR-10 curves"> | |
| <div class="caption fs-2"> | |
| Seed-averaged curves (dotted = baseline, solid = +SIGReg; grey dashes = reshuffles): faster post-shuffle recovery, near-zero SIGReg loss and dormancy, preserved rank. | |
| </div> | |
| </div> | |
| <div class="callout mt-2"> | |
| <strong>24 / 24 directional comparisons reproduce</strong> (4 metrics × 3 optimizers × 2 seeds, each seed independently). | |
| </div> | |
| <div class="keybox mt-3"> | |
| <div class="kb-item"><div class="kb-num">24/24</div><div class="kb-label">SIGReg effect<br>directions hold</div></div> | |
| <div class="kb-item"><div class="kb-num">≤2</div><div class="kb-label">units off paper<br>(SIGReg loss & dorm)</div></div> | |
| <div class="kb-item"><div class="kb-num">84min</div><div class="kb-label">one T4,<br>≈$0.56</div></div> | |
| </div> | |
| </div> | |
| <div class="card pb-col-d" data-measure-role="card" data-logbook-target="conclusion" data-logbook-label="Conclusion & deviations"> | |
| <div class="section-title"><span class="num">9</span><span class="st-text">What Doesn't Match & What We Learned</span></div> | |
| <ul class="mt-2 fs-4"> | |
| <li><strong>Units:</strong> Table 1's “Feature Rank 149–263” is raw RankMe ≈ 1.5–2.6 (columns are 100× time-averages) — the paper's own baselines are near-fully collapsed; our SIGReg rank effect is <em>larger</em>.</li> | |
| <li><strong>Baseline accuracies</strong> run 8–23 points below the paper's (trunk config & probe protocol unspecified; all deviations logged).</li> | |
| <li><strong>Kron-baseline dormancy</strong> reduction vs Adam is weak here (24.2 vs 28.6) where the paper's is strong (7.9 vs 18.5); Kron baseline is also the highest-variance cell (acc range 14.9 over 2 seeds).</li> | |
| <li><strong>RL outcomes (C6):</strong> 51/57 Atari games improved / +889% mean AUC are audited at code level, <em>not rerun</em> — one faithful game-pair alone needs ~10× our compute ceiling.</li> | |
| </ul> | |
| </div> | |
| </div> | |
| </div> | |
| <!-- ============================ FOOTER ============================ --> | |
| <div class="footer" data-measure-role="footer"> | |
| <div> | |
| <strong class="method-name">SIGReg REPRODUCTION</strong> · ICML 2026 Agent Reproduction Challenge · | |
| Paper: arXiv:2602.19373 (OpenReview gc7Gg18ejz). | |
| </div> | |
| <div> | |
| Authors' code: <span class="repo">github.com/asahebpa/IsoGaussian-DRL</span> · | |
| Artifacts: <span class="repo">hf.co/datasets/ProCreations/isogaussian-drl-repro</span> | |
| </div> | |
| </div> | |
| <!-- Watermark. --> | |
| <div class="ornament">HF AGENT REPRODUCTION CHALLENGE · ICML 2026</div> | |
| </div> | |
| </body> | |
| </html> | |
Xet Storage Details
- Size:
- 53.9 kB
- Xet hash:
- b860a01b77c3389e291bce28ee4d74b0f84962a7c02a7e98f2c74ed2893edaca
·
Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.