Research map / Learning theory and algorithms / Prediction and decision losses
On Human-Aligned Risk Minimization
Liu Leqi, Adarsh Prasad, Pradeep K Ravikumar
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| minimize human risk (CPT-weighted, b=.3) instead of expected loss | ERM · synthetic regression, minority group | minority risk | Improved |
| use EHRM instead of ERM | ERM · synthetic regression, population/majority | majority risk | Hurt |
| increase b toward 1 (weighting becomes identity) (ablation) | EHRM · synthetic regression | minority risk | Hurt |
| use EHRM instead of ERM | ERM logistic regression · COMPAS recidivism | misclassification rate / worst-subgroup gap | Mixed |
| use EHRM instead of ERM | ERM · UTKFace gender classification | accuracy: .8767 ±.0060 -> .8751 ±.0052 | Hurt |
| use EHRM instead of ERM | ERM · UTKFace gender classification | statistical parity difference: -.0881 ±.0208 -> -.0825 ±.0220 | Improved |
| use EHRM instead of ERM | ERM · UTKFace gender classification | Theil index: .0855 ±.0071 -> .0859 ±.0058 | No clear effect |
| use small CVaR alpha instead of EHRM | EHRM · synthetic regression | minority vs majority risk | Mixed |
Closest papers by meaning
- Model Alignment as Prospect Theoretic Optimization (ICML 2024 · similarity 0.90)
- Learning to Decide with AI Assistance under Human-Alignment (NeurIPS 2026 · similarity 0.87)
- Decision Theoretic Foundations for Conformal Prediction: Optimal Uncertainty Quantification for Risk-Averse Agents (ICML 2025 · similarity 0.85)
- Controlling Counterfactual Harm in Decision Support Systems Based on Prediction Sets (NeurIPS 2024 · similarity 0.85)
- Risk-aware Direct Preference Optimization under Nested Risk Measure (NeurIPS 2025 · similarity 0.84)
- Direct Alignment with Heterogeneous Preferences (NeurIPS 2025 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
