Fairness in Machine Learning: A Survey

Abstract

As Machine Learning technologies become increasingly used in contexts that affect citizens, companies as well as researchers need to be confident that their application of these methods will not have unexpected social implications, such as bias towards gender, ethnicity, and/or people with disabilities. There is significant literature on approaches to mitigate bias and promote fairness, yet the area is complex and hard to penetrate for newcomers to the domain. This article seeks to provide an overview of the different schools of thought and approaches to mitigating (social) biases and increase fairness in the Machine Learning literature. It organises approaches into the widely accepted framework of pre-processing, in-processing, and post-processing methods, subcategorizing into a further 11 method areas. Although much of the literature emphasizes binary classification, a discussion of fairness in regression, recommender systems, unsupervised learning, and natural language processing is also provided along with a selection of currently available open source libraries. The article concludes by summarising open challenges articulated as four dilemmas for fairness research.

References in corpus (98)

• A Unified Approach to Interpreting Model Predictions
• Semantics derived automatically from language corpora contain human-like biases
• Equality of Opportunity in Supervised Learning
• Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddings
• All Models are Wrong, but Many are Useful: Learning a Variable's Importance by Studying an Entire Class of Prediction Models Simultaneously
• Counterfactual Fairness
• Mitigating Bias in Algorithmic Hiring: Evaluating Claims and Practices
• Fairness Beyond Disparate Treatment & Disparate Impact: Learning Classification without Disparate Mistreatment
• Fairness of Exposure in Rankings
• FA*IR: A Fair Top-k Ranking Algorithm
• The Measure and Mismeasure of Fairness
• Fairness and Accountability Design Needs for Algorithmic Support in High-Stakes Public Sector Decision-Making
• Inherent Trade-Offs in the Fair Determination of Risk Scores
• Fairness Testing: Testing Software for Discrimination
• Preventing Fairness Gerrymandering: Auditing and Learning for Subgroup Fairness
• A Survey on the Fairness of Recommender Systems
• Equity of Attention: Amortizing Individual Fairness in Rankings
• 50 Years of Test (Un)fairness: Lessons for Machine Learning
• Data Decisions and Theoretical Implications when Adversarially Learning Fair Representations
• On Fairness and Calibration
• AI Fairness 360: An Extensible Toolkit for Detecting, Understanding, and Mitigating Unwanted Algorithmic Bias
• A Unified Approach to Quantifying Algorithmic Unfairness: Measuring Individual & Group Unfairness via Inequality Indices
• A Convex Framework for Fair Regression
• Learning Non-Discriminatory Predictors
• Why Is My Classifier Discriminatory?
• The Disparate Effects of Strategic Manipulation
• Fairness in Information Access Systems
• Satisfying Real-world Goals with Dataset Constraints
• Achieving Fairness through Adversarial Learning: an Application to Recidivism Prediction
• Identifying and Correcting Label Bias in Machine Learning
• Racial categories in machine learning
• The Dataset Nutrition Label: A Framework To Drive Higher Data Quality Standards
• Fighting Fire with Fire: Using Antidote Data to Improve Polarization and Fairness of Recommender Systems
• A Reductions Approach to Fair Classification
• Censoring Representations with an Adversary
• On the relation between accuracy and fairness in binary classification
• Fairness Through Computationally-Bounded Awareness
• Fairness in Ranking: A Survey
• Multisided Fairness for Recommendation
• Learning to Pivot with Adversarial Networks
• A Framework for Understanding Sources of Harm throughout the Machine Learning Life Cycle
• Residual Unfairness in Fair Machine Learning from Prejudiced Data
• Online Learning with an Unknown Fairness Metric
• Does mitigating ML's impact disparity require treatment disparity?
• Interventions over Predictions: Reframing the Ethical Debate for Actuarial Risk Assessment
• Penalizing Unfairness in Binary Classification
• Balanced Policy Evaluation and Learning
• Fair Regression: Quantitative Definitions and Reduction-based Algorithms
• Aequitas: A Bias and Fairness Audit Toolkit
• AdaFair: Cumulative Fairness Adaptive Boosting
• A statistical framework for fair predictive algorithms
• Scalable Fair Clustering
• Racial Disparity in Natural Language Processing: A Case Study of Social Media African-American English
• Obtaining fairness using optimal transport theory
• Calibration for the (Computationally-Identifiable) Masses
• Fairness Behind a Veil of Ignorance: A Welfare Analysis for Automated Decision Making
• Calibrated Fairness in Bandits
• Policy Learning for Fairness in Ranking
• Proportionally Fair Clustering
• Identifying Significant Predictive Bias in Classifiers
• Marrying Fairness and Explainability in Supervised Learning
• How to be Fair and Diverse?
• Fairer and more accurate, but for whom?
• The implicit fairness criterion of unconstrained learning
• Improved Adversarial Learning for Fair Classification
• Optimization with Non-Differentiable Constraints with Applications to Fairness, Recall, Churn, and Other Goals
• Repairing without Retraining: Avoiding Disparate Impact with Counterfactual Distributions
• A General Framework for Fair Regression
• Demographics and discussion influence views on algorithmic fairness
• Enhancing the Accuracy and Fairness of Human Decision Making
• The Authority of "Fair" in Machine Learning
• Achieving Fairness in the Stochastic Multi-armed Bandit Problem
• Mitigating Included- and Omitted-Variable Bias in Estimates of Disparate Impact
• Fair Algorithms for Multi-Agent Multi-Armed Bandits
• The cost of fairness in classification
• Algorithmic Decision Making in the Presence of Unmeasured Confounding
• Capuchin: Causal Database Repair for Algorithmic Fairness
• Unleashing Linear Optimizers for Group-Fair Learning and Optimization
• A Notion of Individual Fairness for Clustering
• Fair Algorithms for Infinite and Contextual Bandits
• FNNC: Achieving Fairness through Neural Networks
• Causal Interventions for Fairness
• Stable and Fair Classification
• Fair Decisions Despite Imperfect Predictions
• Counterfactual fairness: removing direct effects through regularization
• Survey on Fair Reinforcement Learning: Theory and Practice
• Fair Division Without Disparate Impact
• Adaptive Sampling for Minimax Fair Classification
• Can Active Learning Preemptively Mitigate Fairness Issues?
• FairNeuron: Improving Deep Neural Network Fairness with Adversary Games on Selective Neurons
• CPFair: Personalized Consumer and Producer Fairness Re-ranking for Recommender Systems
• FairGrad: Fairness Aware Gradient Descent
• Group Meritocratic Fairness in Linear Contextual Bandits
• Improved Approximation for Fair Correlation Clustering
• Measure Twice, Cut Once: Quantifying Bias and Fairness in Deep Neural Networks
• Generating Fair Universal Representations using Adversarial Models
• Mitigating Bias in Algorithmic Systems -- A Fish-Eye View
• Individual Fairness Revisited: Transferring Techniques from Adversarial Robustness

Cited by in corpus (55)

• Blockchain-based Digital Twins: Research Trends, Issues, and Future Challenges
• How to avoid machine learning pitfalls: a guide for academic researchers
• Debiasing Methods for Fairer Neural Models in Vision and Language Research: A Survey
• A survey on datasets for fairness-aware machine learning
• Socially Responsible AI Algorithms: Issues, Purposes, and Challenges
• Algorithmic Unfairness through the Lens of EU Non-Discrimination Law: Or Why the Law is not a Decision Tree
• Can Fairness be Automated? Guidelines and Opportunities for Fairness-aware AutoML
• Consumer-side Fairness in Recommender Systems: A Systematic Survey of Methods and Evaluation
• Enforcing Group Fairness in Algorithmic Decision Making: Utility Maximization Under Sufficiency
• Bias and Fairness in Computer Vision Applications of the Criminal Justice System
• Achieving Model Fairness in Vertical Federated Learning
• Group Fairness in Prediction-Based Decision Making: From Moral Assessment to Implementation
• Modeling Techniques for Machine Learning Fairness: A Survey
• Fairness in Algorithmic Profiling: A German Case Study
• On Prediction-Modelers and Decision-Makers: Why Fairness Requires More Than a Fair Prediction Model
• Properties of fairness measures in the context of varying class imbalance and protected group ratios
• FairIF: Boosting Fairness in Deep Learning via Influence Functions with Validation Set Sensitive Attributes
• FairSort: Learning to Fair Rank for Personalized Recommendations in Two-Sided Platforms
• A Fairness-Oriented Reinforcement Learning Approach for the Operation and Control of Shared Micromobility Services
• Generative AI in Health Economics and Outcomes Research: A Taxonomy of Key Definitions and Emerging Applications, an ISPOR Working Group Report
• Formalising Anti-Discrimination Law in Automated Decision Systems
• Can Active Learning Preemptively Mitigate Fairness Issues?
• Unveiling Group-Specific Distributed Concept Drift: A Fairness Imperative in Federated Learning
• Adversarial Reweighting for Speaker Verification Fairness
• Adaptive Sampling for Minimax Fair Classification
• Measuring Fairness in Generative Models
• Cooperation and Fairness in Multi-Agent Reinforcement Learning
• Towards Efficient and Explainable Hate Speech Detection via Model Distillation
• Fairness Definitions in Language Models Explained
• Mapping Stakeholder Needs to Multi-Sided Fairness in Candidate Recommendation for Algorithmic Hiring
• Facets of Disparate Impact: Evaluating Legally Consistent Bias in Machine Learning
• Fair Clustering Using Antidote Data
• FairGridSearch: A Framework to Compare Fairness-Enhancing Models
• On the Impact of Data Quality on Image Classification Fairness
• Computability, Complexity, Consistency and Controllability: A Four C's Framework for cross-disciplinary Ethical Algorithm Research
• Ethical Quantum Computing: A Roadmap
• Investigating Trade-offs in Utility, Fairness and Differential Privacy in Neural Networks
• Deep AUC Maximization for Medical Image Classification: Challenges and Opportunities
• Simulating counterfactuals
• Correct-By-Construction: Certified Individual Fairness through Neural Network Training
• The Fair Game: Auditing & Debiasing AI Algorithms Over Time
• Toward Substantive Intersectional Algorithmic Fairness: Desiderata for a Feminist Approach
• Adaptation and Generalization for Unknown Sensitive Factors of Variations
• A Hybrid 2-stage Neural Optimization for Pareto Front Extraction
• Fairness-aware Federated Minimax Optimization with Convergence Guarantee
• Designing Human-AI System for Legal Research: A Case Study of Precedent Search in Chinese Law
• Learning to Rank with Variable Result Presentation Lengths
• Introducing a Family of Synthetic Datasets for Research on Bias in Machine Learning
• Measure Twice, Cut Once: Quantifying Bias and Fairness in Deep Neural Networks
• Scalable Unidirectional Pareto Optimality for Multi-Task Learning with Constraints
• StackingNet: Collective Inference Across Independent AI Foundation Models
• Fair Overlap Number of Balls (Fair-ONB): A Data-Morphology-based Undersampling Method for Bias Reduction
• Deep Clustering based Fair Outlier Detection
• Contrastive Clustering: Toward Unsupervised Bias Reduction for Emotion and Sentiment Classification
• Fairness Degrading Adversarial Attacks Against Clustering Algorithms