Research map / Learning theory and algorithms / Fairness

An Information-theoretic Framework for Auditing Unfairness in Training Data

NeurIPS 2026 · Learning theory and algorithms › Fairness · approach: Fairness-constrained classification

Read the paperCheck your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Derives information-theoretic measures for auditing unfairness bias directly in training data, independent of any downstream model.

Problem

Quantifying bias in training data independently of any downstream model is critical for fair machine learning, since most fairness work focuses on model-level mitigation and leaves the data side underexplored.

Approach

Derives multiple information-theoretic measures that audit unfairness in a dataset without requiring access to a trained model.

Closest papers by meaning

• The Fragility of Fairness: Causal Sensitivity Analysis for Fair Machine Learning (NeurIPS 2024 · similarity 0.91)
• Reconciling Predictive and Statistical Parity: A Causal Approach (AAAI 2024 · similarity 0.90)
• Causal Fairness under Unobserved Confounding: A Neural Sensitivity Framework (ICLR 2024 · similarity 0.89)
• Achievable Fairness on Your Data With Utility Guarantees (NeurIPS 2024 · similarity 0.89)
• Understanding Fairness and Prediction Error through Subspace Decomposition and Influence Analysis (NeurIPS 2025 · similarity 0.89)
• Understanding challenges to the interpretation of disaggregated evaluations of algorithmic fairness (NeurIPS 2025 · similarity 0.88)