Research map / Learning theory and algorithms / Fairness
Identifying, Mitigating, and Anticipating Bias in Algorithmic Decisions
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
A doctoral-consortium style research summary outlining the author's work on identifying, mitigating, and anticipating bias in algorithmic decision-making, with no new experiments.
Problem
Algorithmic decision-making in consequential domains can learn and reinforce social biases, and standard ML optimizes objectives that neglect fairness; the paper summarizes how to identify, mitigate, and anticipate such bias.
Approach
Summarizes a research agenda linking group fairness metrics to distributive justice theories, synthetic bias data generation, optimal constrained mitigation for predictive parity, and categorization of feedback loops, and proposes moving toward dynamic long-term fairness and algorithmic collective action.
Key results
- Baumann, Hannák and Heitz (2022) show that for predictive parity it may be optimal for a decision maker to leave out the most promising individuals of a group, in contrast to threshold rules for statistical parity and equality of opportunity.
- Baumann et al. (2023b) propose a bias identification framework linking group fairness metrics to theories of distributive justice and using Pareto efficiency to compare decision rules.
- Baumann et al. (2023a) present a modeling framework for generating synthetic datasets with different forms of bias and post-processing mitigation insights.
Surprising findings
- For predictive parity, the optimal decision may involve leaving out the most promising individuals of a group to achieve globally optimal parity.
Limitations
- Summary of prior work only; no new experiments reported in this paper.
- Future directions are described at a high level without concrete methods.
Experimental setup
Builds on: Barocas et al. fairness textbook Corbett-Davies et al. 2017 Hardt et al. 2016 equality of opportunity
Closest papers by meaning
- Mind the Gap: A Causal Perspective on Bias Amplification in Prediction & Decision-Making (NeurIPS 2024 · similarity 0.83)
- Counterfactual Fairness by Combining Factual and Counterfactual Predictions (NeurIPS 2024 · similarity 0.82)
- Identifying Reasons for Bias: An Argumentation-Based Approach (AAAI 2024 · similarity 0.82)
- Intersectional Fairness via Mixed-Integer Optimization (NeurIPS 2026 · similarity 0.81)
- The Relative Value of Prediction in Algorithmic Decision Making (ICML 2024 · similarity 0.81)
- Understanding challenges to the interpretation of disaggregated evaluations of algorithmic fairness (NeurIPS 2025 · similarity 0.81)
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.