Mechanical Governance for LLM Decisions — model-agnostic governance regimes (R1/R2/R3), hard gates, entropy commit-reveal and governance metrics for high-stakes LLM decision systems.
-
Updated
Oct 1, 2026 - Python
Mechanical Governance for LLM Decisions — model-agnostic governance regimes (R1/R2/R3), hard gates, entropy commit-reveal and governance metrics for high-stakes LLM decision systems.
Code for "Learning to Defer with Limited Expert Predictions" (AAAI 2023)
Code for "Forming Effective Human-AI Teams: Building Machine Learning Models that Complement the Capabilities of Multiple Experts" (IJCAI-ECAI 2022)
Learning to defer between a small and a large LLM — and measuring honestly whether it pays. Headline: at this price ratio it does not, and the identity says when it would.
Research framework for AI assurance and Human–AI routing in financial fraud review using multi-evidence AI-reliance risk and auditable escalation decisions.
Reproducible MEDAI deferral simulation (AIRI 2026). Synthetic research code.
MSc thesis: a risk-sensitive Learning-to-Defer model that sends high-value fraud cases to humans. Live in-browser demo, notebooks, models and the full report.
Official implementation of DeferredSeg, published in Pattern Recognition (2026).
To associate your repository with the learning-to-defer topic, visit your repo's landing page and select "manage topics."