Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Can I Trust My Fairness Metric? Assessing Fairness with Unlabeled Data and Bayesian Inference
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 labeled and unlabeled data with Bayesian calibration instead of labeled data alone | Fairness metric estimation · Accuracy difference; multiple datasets and classifiers, n_L=10 | MAE: 18.5 -> 3.9 | Helped |
| Use labeled and unlabeled data with Bayesian calibration instead of labeled data alone | Fairness metric estimation · Accuracy difference; multiple datasets and classifiers, n_L=10 | MAE: 15.9 -> 2.5 | Helped |
| Use labeled and unlabeled data with Bayesian calibration instead of labeled data alone | Fairness metric estimation · Accuracy difference; multiple datasets and classifiers, n_L=10 | MAE: 34.6 -> 5.0 | Helped |
| Use labeled and unlabeled data with Bayesian calibration instead of labeled data alone | Fairness metric estimation · Accuracy difference; multiple datasets and classifiers, n_L=10 | MAE: 21.0 -> 4.2 | Helped |
| Use LLO calibration instead of beta calibration (ablation) | Bayesian calibration fairness estimator · Multiple datasets and metrics | MAE | No clear effect |
| Use Bayesian calibration credible intervals as labeled sample size increases | Bayesian calibration fairness estimator · Multiple dataset-group combinations | 95% credible interval coverage | Mixed |
| Use Bayesian calibration instead of beta-binomial estimation | Fairness metric estimation · German, Gender (n_L=200), MLP | TPR difference MAE: 1.0 -> 1.6 | Hurt |
Closest papers by meaning
- Size-adaptive Hypothesis Testing for Fairness (NeurIPS 2025 · similarity 0.90)
- Understanding challenges to the interpretation of disaggregated evaluations of algorithmic fairness (NeurIPS 2025 · similarity 0.88)
- Fairness without Demographics through Shared Latent Space-Based Debiasing (AAAI 2024 · similarity 0.88)
- The Fragility of Fairness: Causal Sensitivity Analysis for Fair Machine Learning (NeurIPS 2024 · similarity 0.87)
- Multiaccuracy and Multicalibration via Proxy Groups (ICML 2025 · similarity 0.87)
- Fairness without Harm: An Influence-Guided Active Sampling Approach (NeurIPS 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.