Research map / Learning theory and algorithms / Fairness
Data preprocessing to mitigate bias: A maximum entropy based approach
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 the maximum-entropy distribution with the reweighted prior qCw and θw instead of raw data | Adult, gender · Adult | Data SR: 0.36 (0) -> 0.98 (0.02) | Helped |
| Use the maximum-entropy distribution with balanced expectation θb instead of θw (ablation) | Max-Entropy · Adult, gender | Data RR: 0.97 (0.02) -> 0.99 (0.01) | Helped |
| Use θb instead of θw (ablation) | Max-Entropy · Adult, gender | Classifier accuracy: 0.77 (0.02) -> 0.76 (0.01) | Hurt |
| Replace raw data with Max-Entropy using θb | Decision tree · COMPAS small, gender | Classifier SR: 0.72 (0.01) -> 0.96 (0.02) | Helped |
| Use θb instead of θw (ablation) | Max-Entropy · COMPAS small, gender | Data SR: 0.98 (0.02) -> 0.99 (0.01) | Helped |
| Use the reweighted prior qCw instead of uniform sample weights qCd (ablation) | Max-Entropy · COMPAS and Adult | Data SR | Helped |
| Use qCw and θb instead of qCw and empirical expectation θd (ablation) | Max-Entropy · COMPAS small | Data RR | Helped |
| Increase smoothing parameter C toward 1 (ablation) | Max-Entropy · COMPAS small and large | Classifier accuracy | Hurt |
| Use Max-Entropy preprocessing instead of raw data | Decision tree · COMPAS large, gender | Classifier SR: 0.73 (0.06) -> 0.88 (0.02) | Helped |
| Use Max-Entropy with θb instead of θw (ablation) | Max-Entropy · COMPAS large, gender | Classifier SR: 0.88 (0.02) -> 0.85 (0.06) | Hurt |
| Use Max-Entropy preprocessing instead of raw data | Decision tree · COMPAS small, race | Classifier SR: 0.75 (0.02) -> 0.96 (0.01) | Helped |
| Use Max-Entropy preprocessing instead of raw data | Decision tree · COMPAS large, race | Classifier SR: 0.72 (0.01) -> 0.91 (0.06) | Helped |
Closest papers by meaning
- Everybody Needs Good Neighbours: An Unsupervised Locality-based Method for Bias Mitigation (ICLR 2023 · similarity 0.88)
- Fairness without Demographics through Shared Latent Space-Based Debiasing (AAAI 2024 · similarity 0.87)
- Benchmarking Bias Mitigation Algorithms in Representation Learning through Fairness Metrics (NeurIPS 2021 · similarity 0.87)
- Fairness via Independence: A General Regularization Framework for Machine Learning (ICLR 2026 · similarity 0.87)
- Average Individual Fairness: Algorithms, Generalization and Experiments (NeurIPS 2019 · similarity 0.87)
- On Optimal Steering to Achieve Exact Fairness (NeurIPS 2025 · 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.