From 7177c346ab42144256affa6de930952066e97e9e Mon Sep 17 00:00:00 2001 From: Codex Date: Mon, 7 Sep 2026 21:38:13 +0200 Subject: [PATCH] deck slide 05: numbers as slim top strip; analysis components on ellipse vertices with arc arrows; stat note explained in plain language --- presentation/deck/deck.css | 30 +++++++++++++++------------- presentation/deck/index.html | 38 +++++++++++++++++++++--------------- 2 files changed, 38 insertions(+), 30 deletions(-) diff --git a/presentation/deck/deck.css b/presentation/deck/deck.css index 1979ca67..9b5bdc12 100644 --- a/presentation/deck/deck.css +++ b/presentation/deck/deck.css @@ -185,20 +185,22 @@ .arit .stat .sub { font: 400 12px/1.45 var(--sans); color: var(--ink-muted); } .arit .stat-note { margin-top: 26px; font: 400 13px/1.5 var(--sans); color: var(--ink-muted); } -/* slide "What the text miner reads": compact numeral band on top + analysis components */ -.arit.stats .sbody { padding-top: 24px; } -.arit.stats .stat-grid { grid-template-columns: repeat(5, 1fr); } -.arit.stats .stat { padding: 4px 18px; gap: 5px; } -.arit.stats .stat .num { font-size: 40px; } -.arit.stats .stat .lbl { font-size: 12.5px; } -.arit.stats .stat .sub { font-size: 10.5px; } -.arit.stats .stat.month .num { font-size: 32px; } -.arit.stats .comps { display: flex; align-items: stretch; gap: 8px; margin-top: 34px; } -.arit.stats .comp { flex: 1; text-align: center; padding: 4px 8px; } -.arit.stats .comp .cicon { width: 42px; height: 42px; color: var(--bordeaux); display: block; margin: 0 auto 9px; } -.arit.stats .comp b { display: block; font: 700 13.5px/1.25 var(--sans); text-transform: uppercase; - letter-spacing: .03em; color: var(--ink); } -.arit.stats .comp span.d { display: block; margin-top: 5px; font: 400 11.5px/1.4 var(--sans); color: var(--ink-muted); } +/* slide "What the text miner reads": slim numbers strip + elliptical component cycle */ +.arit.stats .sbody { padding-top: 22px; } +.arit.stats .strip { display: flex; justify-content: space-between; align-items: baseline; + border-top: 1px solid var(--line); border-bottom: 1px solid var(--line); padding: 9px 2px; } +.arit.stats .sstat { display: flex; align-items: baseline; gap: 9px; } +.arit.stats .sstat .n { font: 700 25px/1 var(--sans); color: var(--bordeaux); letter-spacing: -.02em; } +.arit.stats .sstat .l { font: 600 11.5px/1.2 var(--sans); color: var(--ink); } +.arit.stats .cycle { position: relative; width: 1184px; height: 280px; margin: 6px auto 0; } +.arit.stats .cycle-svg { position: absolute; inset: 0; width: 100%; height: 100%; } +.arit.stats .vcomp { position: absolute; transform: translate(-50%, -50%); width: 250px; + text-align: center; background: var(--surface); padding: 4px 6px; } +.arit.stats .vcomp .cicon { width: 46px; height: 46px; color: var(--bordeaux); display: block; margin: 0 auto 7px; } +.arit.stats .vcomp b { display: block; font: 700 14.5px/1.25 var(--sans); text-transform: uppercase; + letter-spacing: .04em; color: var(--ink); } +.arit.stats .vcomp span.d { display: block; margin-top: 4px; font: 400 11.5px/1.4 var(--sans); color: var(--ink-muted); } +.arit.stats .stat-note { margin-top: 8px; font-size: 12.5px; line-height: 1.55; max-width: 1100px; } /* divider */ .arit.divider .sbody { justify-content: center; } diff --git a/presentation/deck/index.html b/presentation/deck/index.html index 2cd3cef0..c5869921 100644 --- a/presentation/deck/index.html +++ b/presentation/deck/index.html @@ -5,7 +5,7 @@ AritmoLab — Role of AI in the Analysis of Unstructured Clinical Databases - +
@@ -226,23 +226,29 @@
What the text mining produced

Deterministic AI, audited numbers

-
-
58,438discharge letters readfree-text fields, bilingual IT/EN patterns
-
73,389pathologies extractedtwo-tier clinical ontology
-
10,908drug-challenge tests parsedflecainide · ajmaline · adrenaline · isoprenaline
-
2,307Brugada patients identifiedconfirmed by clinical review on dashboards
-
≈ 240letters processed monthlynightly batch on twenty years of history
+
+
58,438letters read
+
73,389pathologies extracted
+
10,908tests parsed
+
2,307Brugada patients
+
≈ 240letters / month
-
-
Letter readerevery discharge letter, Italian & English
-
→
-
Clinical matcherrecognises diagnoses, procedures, test outcomes
-
→
-
Context guardnegations & family history kept apart
-
→
-
Ontology sorterevery finding into its clinical category
+
+ + + + + + + + + +
Letter readerevery discharge letter, Italian & English
+
Clinical matcherrecognises diagnoses, procedures, test outcomes
+
Context guardnegations & family history kept apart
+
Ontology sorterevery finding into its clinical category
-

Every row carries its pattern-library semver (pattern_version) · quality gate: ≥95% accuracy on 100 manual reviews.

+

No black box: every extracted finding carries the version of the rules that produced it (pattern_version) — any number in this talk can be rebuilt years from now. And no rule set reaches production until it proves ≥95% accuracy on 100 records reviewed by hand.

Role of AI in the Analysis of Unstructured Clinical DatabasesDr. Marco Pancotti - MultiPhysixLabDr. Sara Paratico - Gruppo San DonatoSan Donato Milanese, Milan, Italy · 2–3 October 202605 / 14