mptyl/ThothII
1
0
Fork 0
Code Issues 25 Pull Requests Actions Packages Projects Releases 1 Wiki Activity

deck slide 08: split layout — current content (4 numbered points + quiet quality-gate box) contained in left column, new right panel 'The advantage: speed, precision, determinism' (machine speed / audited precision / deterministic by design); css v39

Browse Source
This commit is contained in:
Codex
2026-09-07 23:41:44 +02:00
parent 32ed2d1c14
commit 8f59da412c
2 changed files with 22 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files
presentation/deck/deck.css
+4
View File
@@ -229,6 +229,9 @@
.arit .cat .it { font: 400 11px/1 var(--sans); color: var(--ink-muted); } .arit .cat .it { font: 400 11px/1 var(--sans); color: var(--ink-muted); }
.arit .cat.fill { background: var(--bordeaux-tint); } .arit .cat.fill { background: var(--bordeaux-tint); }
.arit .onto-points { margin-top: 18px; } .arit .onto-points { margin-top: 18px; }
.arit .onto-left { display: flex; flex-direction: column; }
.arit .onto-left .roi { margin-top: auto; }
.arit .onto-left .onto-points { margin-bottom: 14px; }
.arit .oitem { display: flex; gap: 10px; padding: 8px 0; border-bottom: 1px solid var(--line); } .arit .oitem { display: flex; gap: 10px; padding: 8px 0; border-bottom: 1px solid var(--line); }
.arit .oitem:last-child { border-bottom: none; } .arit .oitem:last-child { border-bottom: none; }
.arit .oitem .n { flex: 0 0 22px; font: 700 11px/1.6 var(--mono); color: var(--bordeaux); } .arit .oitem .n { flex: 0 0 22px; font: 700 11px/1.6 var(--mono); color: var(--bordeaux); }
@@ -306,6 +309,7 @@ button.brain-btn:hover, button.brain-btn:focus-visible { transform: scale(1.18);
.ai-pop-list .n { color: var(--bordeaux-deep); font-weight: 800; } .ai-pop-list .n { color: var(--bordeaux-deep); font-weight: 800; }
.roi { margin-top: 14px; border: 1.5px solid var(--bordeaux); background: #faf4f5; border-radius: 8px; .roi { margin-top: 14px; border: 1.5px solid var(--bordeaux); background: #faf4f5; border-radius: 8px;
padding: 10px 14px; font: 400 13.5px/1.55 var(--sans); color: var(--ink); } padding: 10px 14px; font: 400 13.5px/1.55 var(--sans); color: var(--ink); }
.roi-quiet { border-color: var(--line-strong); border-width: 1px; background: var(--surface-alt); }
.roi .k { display: block; margin-bottom: 4px; font: 800 10.5px/1 var(--mono); letter-spacing: .14em; .roi .k { display: block; margin-bottom: 4px; font: 800 10.5px/1 var(--mono); letter-spacing: .14em;
text-transform: uppercase; color: var(--bordeaux-deep); } text-transform: uppercase; color: var(--bordeaux-deep); }
.ai-pop-x { position: absolute; top: 12px; right: 14px; width: 30px; height: 30px; border: none; .ai-pop-x { position: absolute; top: 12px; right: 14px; width: 30px; height: 30px; border: none;
presentation/deck/index.html
+18 -9
View File
@@ -5,7 +5,7 @@
<title>AritmoLab — Role of AI in the Analysis of Unstructured Clinical Databases</title> <title>AritmoLab — Role of AI in the Analysis of Unstructured Clinical Databases</title>
<link rel="stylesheet" href="vendor/reveal/dist/reveal.css"> <link rel="stylesheet" href="vendor/reveal/dist/reveal.css">
<link rel="stylesheet" href="fonts.css?v=8"> <link rel="stylesheet" href="fonts.css?v=8">
<link rel="stylesheet" href="deck.css?v=37"> <link rel="stylesheet" href="deck.css?v=39">
</head> </head>
<body> <body>
<div class="reveal"><div class="slides"> <div class="reveal"><div class="slides">
@@ -370,14 +370,23 @@
<div class="sbody"> <div class="sbody">
<div class="kicker">Validation</div> <div class="kicker">Validation</div>
<h2>Trusting the text is a process, not a promise</h2> <h2>Trusting the text is a process, not a promise</h2>
<div class="hgap"></div> <div class="onto-split">
<ul class="points"> <div class="onto-left">
<li><b>Quality gates</b> — ≥95% procedure classification, ≥85% pathology detection, validated on 100 manually reviewed records</li> <div class="onto-points">
<li><b>Clinical review</b> — the extracted Brugada cohort was validated by clinicians on Superset dashboards</li> <div class="oitem"><span class="n">01</span><div><b>Quality gates</b> <span class="d">— ≥95% procedure classification, ≥85% pathology detection, validated on 100 manually reviewed records</span></div></div>
<li><b>Feedback loop</b> — ~893 Brugada false positives found and removed in pattern-library v1.3.1</li> <div class="oitem"><span class="n">02</span><div><b>Clinical review</b> <span class="d">— the extracted Brugada cohort was validated by clinicians on Superset dashboards</span></div></div>
<li><b>Nothing is silently dropped</b> — negated and family-attributed findings are stored too, filtered only at the mart layer</li> <div class="oitem"><span class="n">03</span><div><b>Feedback loop</b> <span class="d">— ~893 Brugada false positives found and removed in pattern-library v1.3.1</span></div></div>
</ul> <div class="oitem"><span class="n">04</span><div><b>Nothing is silently dropped</b> <span class="d">— negated and family-attributed findings are stored too, filtered only at the mart layer</span></div></div>
<div class="roi"><b class="k">Quality gate</b>An extraction is accepted only after 100 manual reviews confirm the agreed accuracy — the same standard for every new pattern release.</div> </div>
<div class="roi roi-quiet"><b class="k">Quality gate</b>An extraction is accepted only after 100 manual reviews confirm the agreed accuracy — the same standard for every new pattern release.</div>
</div>
<div class="out-panel">
<div class="out-h">The advantage: speed, precision, determinism</div>
<div class="out-item"><span class="n">01</span><div><b>Machine speed</b><span class="d">twenty years of letters are read by rules, not by hand; when a pattern improves, the whole archive is rebuilt from the letters themselves</span></div></div>
<div class="out-item"><span class="n">02</span><div><b>Audited precision</b><span class="d">every release clears the same gate before production: ≥95% procedure classification, ≥85% pathology detection</span></div></div>
<div class="out-item"><span class="n">03</span><div><b>Deterministic by design</b><span class="d">the same letter always yields the same values: no LLM sampling, no randomness, every result traceable to a versioned pattern</span></div></div>
</div>
</div>
</div> </div>
<footer class="foot"><span class="g">Role of AI in the Analysis of Unstructured Clinical Databases</span><span class="g">Dr. Marco Pancotti - MultiPhysixLab</span><span class="g">Dr. Sara Paratico - Gruppo San Donato</span><span class="g">San Donato Milanese, Milan, Italy · 2–3 October 2026</span><span class="g num">08 / 14</span></footer> <footer class="foot"><span class="g">Role of AI in the Analysis of Unstructured Clinical Databases</span><span class="g">Dr. Marco Pancotti - MultiPhysixLab</span><span class="g">Dr. Sara Paratico - Gruppo San Donato</span><span class="g">San Donato Milanese, Milan, Italy · 2–3 October 2026</span><span class="g num">08 / 14</span></footer>
<aside class="notes"> <aside class="notes">