Research map / Learning theory and algorithms / Fairness
A General Framework for Fair and Robust Regression
Wenhai Cui, Xiaoting Ji, Wen Su, Xingqiu Zhao
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 |
|---|---|---|---|
| learn optimal quantile transform instead of Wasserstein barycenter (FRWB) | fair robust regression · Simulation (Cauchy loss) | empirical risk: 1.2354 -> 1.0469 | Improved |
| learn optimal quantile transform instead of Wasserstein barycenter (FRWB) | fair robust regression · Simulation (LAD loss) | empirical risk: 1.7207 -> 1.5718 | Improved |
| learn optimal quantile transform instead of Wasserstein barycenter (FRWB) | fair robust regression · Simulation (Quantile loss) | empirical risk: 0.8029 -> 0.7950 | Improved |
| learn optimal quantile transform instead of Wasserstein barycenter (FRWB) | fair robust regression · Simulation (Cauchy loss) | unfairness: 0.0125 -> 0.0048 | Improved |
| learn optimal quantile transform instead of Wasserstein barycenter (FRWB) | fair robust regression · CRIME (LAD loss) | unfairness: 0.0615 -> 0.0093 | Improved |
| learn optimal quantile transform instead of Wasserstein barycenter (FRWB) | fair robust regression · CRIME (Cauchy loss) | empirical risk: 0.0339 -> 0.0352 | No clear effect |
| learn optimal quantile transform instead of Wasserstein barycenter (FRWB) | fair robust regression · CRIME (LAD loss) | empirical risk: 0.1330 -> 0.1294 | Improved |
| add demographic parity transformation to robust DNN (RDNN) | robust regression · Simulation (Cauchy loss) | unfairness: 2.4669 -> 0.0048 | Improved |
| add demographic parity transformation to robust DNN (RDNN) | robust regression · Simulation (Cauchy loss) | empirical risk: 0.7788 -> 1.0469 | Hurt |
| increase interpolation parameter lambda toward unconstrained predictor | interpolated fair predictor · Simulation (Cauchy loss) | empirical risk and unfairness | Mixed |
| use Q(u)=u instead of Wasserstein barycenter quantile | DP transformation · Toy example (LAD) | LAD risk: 0.362 -> 0.272 | Improved |
Closest papers by meaning
- Meta Optimality for Demographic Parity Constrained Regression via Post-Processing (ICML 2025 · similarity 0.90)
- Geometry of Relaxed Fair Regression: A Unified Framework for Aware and Unaware Settings (NeurIPS 2026 · similarity 0.88)
- Demographic parity in regression and classification within the unawareness framework (NeurIPS 2026 · similarity 0.87)
- Auditing and Enforcing Conditional Fairness via Optimal Transport (AAAI 2025 · similarity 0.87)
- Regression under demographic parity constraints via unlabeled post-processing (NeurIPS 2024 · similarity 0.85)
- A Sequentially Fair Mechanism for Multiple Sensitive Attributes (AAAI 2024 · 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.
