Research map / Learning theory and algorithms / Fairness
Learning Certified Individually Fair Representations
Anian Ruoss, Mislav Balunovic, Marc Fischer, Martin Vechev
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 |
|---|---|---|---|
| learn the representation with the fairness constraint (LCIFR) instead of γ=0 baseline | LCIFR · Cat + Noise across Adult, Compas, Crime, German, Health, Law School | certified (%): 47.5 -> 97.6; 30.9 -> 75.6; 6.2 -> 63.3; 68.0 -> 95.5; 24.7 -> 97.3; 11.6 -> 28.9 | Improved |
| learn the representation with the fairness constraint (LCIFR) instead of γ=0 baseline | LCIFR · Cat + Noise across Adult, Compas, Crime, German, Health, Law School | accuracy (%): 83.3 -> 81.3; 65.6 -> 63.7; 84.4 -> 81.5; 76.0 -> 70.0; 80.7 -> 80.7; 84.4 -> 84.5 | Mixed |
| increase fairness-loss weight γ from 0 to 0.1 (ablation) | LCIFR · Crime, Cat + Noise | certified (%): 6.15 -> 9.23 | Improved |
| increase fairness-loss weight γ from 0 to 0.2 (ablation) | LCIFR · Crime, Cat + Noise | certified (%): 6.15 -> 12.05 | Improved |
| increase fairness-loss weight γ from 0 to 0.5 (ablation) | LCIFR · Crime, Cat + Noise | certified (%): 6.15 -> 18.46 | Improved |
| increase fairness-loss weight γ from 0 to 1 (ablation) | LCIFR · Crime, Cat + Noise | certified (%): 6.15 -> 33.08 | Improved |
| increase fairness-loss weight γ from 0 to 2 (ablation) | LCIFR · Crime, Cat + Noise | certified (%): 6.15 -> 52.31 | Improved |
| increase fairness-loss weight γ from 0 to 5 (ablation) | LCIFR · Crime, Cat + Noise | certified (%): 6.15 -> 61.28 | Improved |
| increase fairness-loss weight γ from 0 to 10 (ablation) | LCIFR · Crime, Cat + Noise | certified (%): 6.15 -> 62.82 | Improved |
| increase fairness-loss weight γ from 0 to 20 (ablation) | LCIFR · Crime, Cat + Noise | accuracy (%): 84.36 -> 50.77 | Hurt |
| increase fairness-loss weight γ from 0 to 50 (ablation) | LCIFR · Crime, Cat + Noise | accuracy (%): 84.36 -> 50.77 | Hurt |
| add adversarial training to a 2-layer neural-network consumer classifier (ablation) | LCIFR · Health, Cat + Noise, γ=1 | certified (%): 34% -> 70.8% | Improved |
Closest papers by meaning
- Fair Representation: Guaranteeing Approximate Multiple Group Fairness for Unknown Tasks (TPAMI 2022 · similarity 0.90)
- Two Simple Ways to Learn Individual Fairness Metrics from Data (ICML 2020 · similarity 0.89)
- Kernel-based Equalized Odds: A Quantification of Accuracy-Fairness Trade-off in Fair Representation Learning (NeurIPS 2025 · similarity 0.87)
- Fair Representation Learning with Controllable High Confidence Guarantees via Adversarial Inference (NeurIPS 2025 · similarity 0.87)
- Individual Fairness In Strategic Classification (NeurIPS 2025 · similarity 0.87)
- Monotone Individual Fairness (ICML 2024 · 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.
