Research map / Learning theory and algorithms / Fairness
A Game-Theoretic Framework for Measuring and Explaining Metric Compatibility in Fair Machine Learning
Lingfeng Zhang, Jingran Yang, Zhaohui Wang, Min Zhang, Zhang Qing
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 |
|---|---|---|---|
| apply Fair-SMOTE debiasing | Vanilla models · Compatibility under debiasing | Overall mean compatibility: 0.2518 -> 0.1405 | Hurt |
| apply MAAT debiasing | Vanilla models · Compatibility under debiasing | Overall mean compatibility: 0.2518 -> 0.2219 | Hurt |
| apply Fair-SMOTE debiasing | Vanilla models · Compatibility under debiasing | GF-U compatibility: -0.0380 -> -0.0507 | Hurt |
| apply Fair-SMOTE debiasing | Vanilla models · Compatibility under debiasing | GF-IF compatibility: 0.41 -> 0.1271 | Hurt |
| apply Fair-SMOTE debiasing | Vanilla models · Compatibility under debiasing | U-U compatibility: 0.7454 -> 0.5683 | Hurt |
| apply MAAT debiasing | Vanilla models · Compatibility under debiasing | U-U compatibility: 0.7454 -> 0.7823 | Improved |
| apply Blindness debiasing | Vanilla models · Compatibility under debiasing | IF-IF compatibility: 0.4154 -> 0.0000 | Hurt |
| apply MAAT debiasing | Vanilla models · Compatibility under debiasing | IF-U compatibility: -0.0340 -> -0.0352 | No clear effect |
| apply Fair-SMOTE debiasing | Vanilla models · Compatibility under debiasing | IF-U compatibility: -0.0340 -> -0.0241 | Improved |
| increase HIFI regularization strength eta beyond about 1000 | HIFI, LR on Census Income · Census Income | U-U compatibility: from ~1.0 to ~0.6 | Hurt |
| use HIFI with eta below 1000 | HIFI, LR on Census Income · Census Income | AOD: AOD reduces by ~40% | Improved |
Closest papers by meaning
- The Fragility of Fairness: Causal Sensitivity Analysis for Fair Machine Learning (NeurIPS 2024 · similarity 0.87)
- Reconciling Predictive and Statistical Parity: A Causal Approach (AAAI 2024 · similarity 0.86)
- Fairness-Aware Meta-Learning via Nash Bargaining (NeurIPS 2024 · similarity 0.85)
- Fairness under Competition (NeurIPS 2025 · similarity 0.85)
- Constructing Fair Latent Space for Intersection of Fairness and Explainability (AAAI 2025 · similarity 0.85)
- Understanding challenges to the interpretation of disaggregated evaluations of algorithmic fairness (NeurIPS 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.
