Research map / Label-efficient and robust learning / Domain adaptation and generalization
Does enforcing fairness mitigate biases caused by subpopulation shift?
Subha Maity, Debarghya Mukherjee, Mikhail Yurochkin, Yuekai Sun
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 |
|---|---|---|---|
| Enforce tight Equalized Odds during training instead of using a loose fairness constraint | Reductions with logistic regression · COMPAS, target distribution P* | accuracy: 0.634±0.011 -> 0.652±0.013; 1.8% higher | Improved |
| Enforce tight Equalized Odds during training instead of using a loose fairness constraint | Reductions with logistic regression · COMPAS, training distribution P̃ | accuracy: 0.668±0.010 -> 0.660±0.009; 0.8% | Hurt |
| Enforce tight Equalized Odds during training instead of using a loose fairness constraint | Reductions with logistic regression · Adult, target distribution P* | accuracy: 0.848±0.005 -> 0.852±0.004 | Improved |
| Enforce tight Equalized Odds during training instead of using a loose fairness constraint | Reductions with logistic regression · Adult, training distribution P̃ | accuracy: 0.847±0.003 -> 0.843±0.003 | Hurt |
| Enforce Equalized Odds during training on biased simulated data | Reductions with logistic regression · Simulated data, target distribution P* | accuracy: +3% | Improved |
| Enforce Equalized Odds during training on biased simulated data | Reductions with logistic regression · Simulated data, training distribution P̃ | accuracy: 1% less accurate | Hurt |
| Enforce risk parity rather than minimize risk without a fairness constraint | Risk minimization under subpopulation shift · Target domain; outcome depends on target group proportions | overall risk | Mixed |
| Enforce fairness when the target-domain Bayes decision rule is fair | Fair risk minimization under subpopulation shift · Target domain | overall risk | Mixed |
Closest papers by meaning
- Fairness without Harm: An Influence-Guided Active Sampling Approach (NeurIPS 2024 · similarity 0.88)
- Understanding challenges to the interpretation of disaggregated evaluations of algorithmic fairness (NeurIPS 2025 · similarity 0.87)
- Fairness under Competition (NeurIPS 2025 · similarity 0.87)
- AIM-Fair: Advancing Algorithmic Fairness via Selectively Fine-Tuning Biased Models with Contextual Synthetic Data (CVPR 2025 · similarity 0.87)
- A Comprehensive View of Fairness through Distributional Stability (NeurIPS 2026 · similarity 0.86)
- Stay Fair! Ensuring Group Fairness in Diffusion Models Across Guidance Scales (NeurIPS 2026 · similarity 0.86)
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.
