Each dot is one task's convergence score S. Below Lower → Reject · between → Gray (human) · at/above Upper → Auto-Resolve. Drag to reclassify live.
| Task | Name | Dimension | S(c,n) | Disposition |
|---|
Add evidence nodes and watch the real Heim engine compute the score live — W(v)=min(0.95, α·C+β·E+μ·R) and S = Σ W(validated) / Σ W_eff(all). Same engine that scores every task.
R(v): 0.50 = no edges · >0.50 corroborated · <0.50 contradicted. Disputed weight decays by γ^(n−n₀).
| # | Origin | C | E | R | W(v) | Status | W_eff |
|---|
Embedded demo AI-Vault BI Platform overview + Olympus test harness & Vault-Comply validation engine. Self-contained; opens in its own frame.
Embedded demo Cyber Security Operations Center console — live alert feed + three-agent adversarial triage. Click an alert to drill into the investigation view (loads inside this frame).
Embedded demo Vault-Comply — autonomous privacy/GRC dashboard with an embedded Olympus validation view.
W(v), W_eff, S = ΣW(validated)/ΣW_eff(all)) from that task's evidence profile — not scripted. Task names, domains and systems are illustrative reconstructions of real operations; the scoring and disposition are exact. The three embedded demos are the original product demos (fonts inlined for offline use).