SUPREME-1 v2.0: Cryptographic Determinism and Auditable Governance for High-Risk AI Systems under EU AI Act (Reg. 2024/1689)
Fuente:
Zenodo
Enregistré dans:
| Auteur principal: | |
|---|---|
| Format: | Recurso digital |
| Publié: |
Zenodo
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866902090407739392 |
|---|---|
| author | Valente, Stefano |
| author_facet | Valente, Stefano |
| contents | <p>SUPREME-1 v2.0 delivers production-grade governance for High-Risk AI Systems (HRAIS), operationalizing EU AI Act requirements (Art. 6, 9–15, 27, 61) through Reality Drift Index (Dᵣ), predictive entropy bounds, variance stabilization, and SHA-256 cryptographic audit trails. Validated on MIMIC-IV (n=50k) via ROC-optimized thresholds (Youden’s J=0.92) and ablation studies (92% risk reduction), the framework minimizes Free Energy Principle variational bounds under deterministic control. Includes FRIA (Art. 27), post-market KPIs (SLA 99.9%), and cross-language (Python/C++) implementations for regulatory non-repudiation. Deployable template for healthcare/justice compliance.</p> |
| format | Recurso digital |
| id | zenodo_https___doi_org_10_5281_zenodo_18306194 |
| institution | Zenodo |
| language | |
| publishDate | 2026 |
| publisher | Zenodo |
| record_format | zenodo |
| spellingShingle | SUPREME-1 v2.0: Cryptographic Determinism and Auditable Governance for High-Risk AI Systems under EU AI Act (Reg. 2024/1689) Valente, Stefano • SUPREME-1 v2.0 • High-Risk AI Systems (HRAIS) • EU AI Act 2024/1689 • Reality Drift Index (Dᵣ) • Free Energy Principle • Cryptographic governance • SHA-256 audit trails • FRIA Article 27 • Post-market monitoring • Deterministic AI safety <p>SUPREME-1 v2.0 delivers production-grade governance for High-Risk AI Systems (HRAIS), operationalizing EU AI Act requirements (Art. 6, 9–15, 27, 61) through Reality Drift Index (Dᵣ), predictive entropy bounds, variance stabilization, and SHA-256 cryptographic audit trails. Validated on MIMIC-IV (n=50k) via ROC-optimized thresholds (Youden’s J=0.92) and ablation studies (92% risk reduction), the framework minimizes Free Energy Principle variational bounds under deterministic control. Includes FRIA (Art. 27), post-market KPIs (SLA 99.9%), and cross-language (Python/C++) implementations for regulatory non-repudiation. Deployable template for healthcare/justice compliance.</p> |
| title | SUPREME-1 v2.0: Cryptographic Determinism and Auditable Governance for High-Risk AI Systems under EU AI Act (Reg. 2024/1689) |
| topic | • SUPREME-1 v2.0 • High-Risk AI Systems (HRAIS) • EU AI Act 2024/1689 • Reality Drift Index (Dᵣ) • Free Energy Principle • Cryptographic governance • SHA-256 audit trails • FRIA Article 27 • Post-market monitoring • Deterministic AI safety |
| url | https://doi.org/10.5281/zenodo.18306194 |