Research map / Learning theory and algorithms / Fairness
Fair Classification with Noisy Protected Attributes: A Framework with Provable Guarantees
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 denoised statistical-rate constraints instead of unconstrained training | Logistic regression on Adult sex · Adult sex | statistical rate: .31 (.01) -> .85 (.15) | Helped |
| Use denoised false-positive-rate constraints instead of unconstrained training | Logistic regression on Adult sex · Adult sex | false positive rate: .45 (.03) -> .91 (.11) | Helped |
| Use denoised statistical-rate constraints instead of unconstrained training | Logistic regression on COMPAS sex · COMPAS sex | statistical rate: .78 (.04) -> .86 (.05) | Helped |
| Use denoised false-positive-rate constraints instead of unconstrained training | Logistic regression on COMPAS sex · COMPAS sex | false positive rate: .70 (.08) -> .82 (.10) | Helped |
| Use denoised statistical-rate constraints instead of unconstrained training | Logistic regression on COMPAS race · COMPAS race | SR_0: .66 (.02) -> .91 (.06) | Helped |
| Use denoised false-positive-rate constraints instead of unconstrained training | Logistic regression on COMPAS race · COMPAS race | FPR_0: .57 (.02) -> .93 (.08) | Helped |
| Set the fairness threshold to τ=.9 rather than τ=.7 | DLR-SR on Adult sex · Adult sex | statistical rate: .74 (.14) -> .85 (.15) | Helped |
| Set the fairness threshold to τ=.9 rather than τ=.7 | DLR-SR on Adult sex · Adult sex | accuracy: .77 (.01) -> .76 (.01) | Hurt |
| Use denoised false-discovery-rate constraints instead of noisy-attribute LR-FPR constraints | Logistic regression on Adult sex · Adult sex | false discovery rate: .54 (.45) -> .87 (.08) | Helped |
Closest papers by meaning
- Ensuring Fairness Beyond the Training Data (NeurIPS 2020 · similarity 0.91)
- Fair Classification with Adversarial Perturbations (NeurIPS 2021 · similarity 0.91)
- A Reductions Approach to Fair Classification (ICML 2018 · similarity 0.89)
- Efficient Fairness-Performance Pareto Front Computation (NeurIPS 2025 · similarity 0.88)
- Optimal Fair Learning Robust to Adversarial Distribution Shift (ICML 2025 · similarity 0.88)
- fairret: a Framework for Differentiable Fairness Regularization Terms (ICLR 2024 · similarity 0.87)
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.