Research map / Learning theory and algorithms / Fairness
Training Well-Generalizing Classifiers for Fairness Metrics and Other Data-Dependent Constraints
Andrew Cotter, Maya Gupta, Heinrich Jiang, Nathan Srebro, Karthik Sridharan, Serena Wang, Blake Woodworth, Seungil You
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 |
|---|---|---|---|
| Use an independent validation set to choose constraint multipliers instead of using the same data for model and constraint updates | PracticalTwoDataset on Communities and Crime · Communities and Crime | Maximum constraint violation: $.022$ -> $-.008$ | Improved |
| Use an independent validation set to choose constraint multipliers instead of using the same data for model and constraint updates | PracticalTwoDataset on Adult · Adult | Maximum constraint violation: $.011$ -> $.005$ | Improved |
| Use an independent validation set to choose constraint multipliers instead of using the same data for model and constraint updates | PracticalTwoDataset on COMPAS · COMPAS | Maximum constraint violation: $.038$ -> $.004$ | Improved |
| Use an independent validation set to choose constraint multipliers instead of using the same data for model and constraint updates | PracticalTwoDataset on Business Entity Resolution · Business Entity Resolution | Maximum constraint violation: $.073$ -> $.072$ | No clear effect |
| Use an independent validation set to choose constraint multipliers instead of using the same data for model and constraint updates | PracticalTwoDataset on Adult · Adult | Error: $.156$ -> $.169$ | Hurt |
| Use an independent validation set to choose constraint multipliers instead of using the same data for model and constraint updates | PracticalTwoDataset on COMPAS · COMPAS | Error: $.353$ -> $.378$ | Hurt |
| Use an independent validation set to choose constraint multipliers instead of using the same data for model and constraint updates | LagrangianTwoDataset on Communities and Crime · Communities and Crime | Maximum constraint violation: $.001$ -> $-.012$ | Improved |
| Use an independent validation set to choose constraint multipliers instead of using the same data for model and constraint updates | LagrangianTwoDataset on Adult · Adult | Maximum constraint violation: $.013$ -> $.008$ | Improved |
| Use an independent validation set to choose constraint multipliers instead of using the same data for model and constraint updates | LagrangianTwoDataset on COMPAS · COMPAS | Maximum constraint violation: $.029$ -> $.006$ | Improved |
| Use an independent validation set to choose constraint multipliers instead of using the same data for model and constraint updates | LagrangianTwoDataset on Business Entity Resolution · Business Entity Resolution | Maximum constraint violation: $.042$ -> $.041$ | No clear effect |
| Use an independent validation set to choose constraint multipliers instead of using the same data for model and constraint updates | LagrangianTwoDataset on COMPAS · COMPAS | Error: $.349$ -> $.378$ | Hurt |
Closest papers by meaning
- Optimizing Generalized Rate Metrics with Three Players (NeurIPS 2019 · similarity 0.90)
- Too Relaxed to Be Fair (ICML 2020 · similarity 0.88)
- Calibrated Data-Dependent Constraints with Exact Satisfaction Guarantees (NeurIPS 2022 · similarity 0.88)
- Learning with Statistical Equality Constraints (NeurIPS 2025 · similarity 0.88)
- Fairness via Independence: A General Regularization Framework for Machine Learning (ICLR 2026 · similarity 0.88)
- Fairness Overfitting in Machine Learning: An Information-Theoretic Perspective (ICML 2025 · similarity 0.85)
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.
