Fairness in Machine Learning: A Survey

Published in

ACM Computing Surveys(2020)

External Links:

Generate Graph DownloadPDF

TL;DR

An overview of the different schools of thought and approaches that aim to increase the fairness of Machine Learning is provided, organizes approaches into the widely accepted framework of pre-processing, in-processing, and post-processing methods, subcategorizing into a further 11 method areas.

Abstract

When Machine Learning technologies are used in contexts that affect citizens, companies as well as researchers need to be confident that there will not be any 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 that aim to increase the fairness of Machine Learning. It organizes 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, and unsupervised learning is also provided along with a selection of currently available open source libraries. The article concludes by summarizing open challenges articulated as five dilemmas for fairness research.

Authors

Simon Caton

1 papers

C. Haas

1 papers

References360 items

Learning fair representations via an adversarial framework

An Ontology for Fairness Metrics

Improving fairness generalization through a sample-robust optimization method

Improved Approximation for Fair Correlation Clustering

A Survey on the Fairness of Recommender Systems

Fairness in Machine Learning: A Survey

Published in

ACM Computing Surveys(2020)

External Links:

Generate Graph DownloadPDF

TL;DR

Abstract

Authors

Simon Caton

1 papers

C. Haas

1 papers

References360 items

Learning fair representations via an adversarial framework

An Ontology for Fairness Metrics

Improving fairness generalization through a sample-robust optimization method

Improved Approximation for Fair Correlation Clustering

A Survey on the Fairness of Recommender Systems

Group Meritocratic Fairness in Linear Contextual Bandits

Fairness in Ranking, Part I: Score-Based Ranking

Fairness in Ranking, Part II: Learning-to-Rank and Recommender Systems

CPFair: Personalized Consumer and Producer Fairness Re-ranking for Recommender Systems

Fairneuron: Improving Deep Neural Network Fairness with Adversary Games on Selective Neurons

Marrying Fairness and Explainability in Supervised Learning

Search-based fairness testing for regression-based machine learning systems

Ethics Guidelines for Trustworthy AI

An Interactive Approach to Bias Mitigation in Machine Learning

Measure Twice, Cut Once: Quantifying Bias and Fairness in Deep Neural Networks

A survey on datasets for fairness‐aware machine learning

Fairness metrics and bias mitigation strategies for rating predictions

Trustworthy AI: A Computational Perspective

Deep Clustering based Fair Outlier Detection

Using Pareto simulated annealing to address algorithmic bias in machine learning

Analyzing the impact of missing values and selection bias on fairness

Can Active Learning Preemptively Mitigate Fairness Issues?

Mitigating Bias in Algorithmic Systems—A Fish-eye View

Missing the missing values: The ugly duckling of fairness in machine learning

Fairness of Exposure in Stochastic Bandits

Adaptive Sampling for Minimax Fair Classification

TARA: Training and Representation Alteration for AI Fairness and Domain Generalization

FairBatch: Batch Selection for Model Fairness

FAIR: Fair Adversarial Instance Re-weighting

Achieving User-Side Fairness in Contextual Bandits

Fair Algorithms for Multi-Agent Multi-Armed Bandits

A Notion of Individual Fairness for Clustering

Counterfactual fairness: removing direct effects through regularization

Fair Prediction with Endogenous Behavior

Individual Fairness Revisited: Transferring Techniques from Adversarial Robustness

Fair Correlation Clustering

The false promise of risk assessments: epistemic reform and the limits of fairness

Algorithmic realism: expanding the boundaries of algorithmic thought

Algorithmic Fairness

FAE: A Fairness-Aware Ensemble Framework

Preventing undesirable behavior of intelligent machines

Predictive Biases in Natural Language Processing Models: A Conceptual Framework and Overview

Censored and Fair Universal Representations using Generative Adversarial Models

AdaFair: Cumulative Fairness Adaptive Boosting

Fair Coresets and Streaming Algorithms for Fair k-means

A Survey on Bias and Fairness in Machine Learning

Data Management for Causal Algorithmic Fairness

Integrating Ethics within Machine Learning Courses

Achieving Causal Fairness through Generative Adversarial Networks

Achieving Fairness in the Stochastic Multi-armed Bandit Problem

One-Network Adversarial Fairness

Allocating Interventions Based on Predicted Outcomes: A Case Study on Homelessness Services

Wasserstein Fair Classification

Interventional Fairness: Causal Database Repair for Algorithmic Fairness

Mitigating bias in algorithmic hiring: evaluating claims and practices

Mitigating Gender Bias in Natural Language Processing: Literature Review

Mitigating Bias in Algorithmic Employment Screening: Evaluating Claims and Practices

Why Machine Learning May Lead to Unfairness: Evidence from Risk Assessment for Juvenile Justice in Catalonia

Fair Division Without Disparate Impact

Democratizing Algorithmic Fairness

Fair Regression: Quantitative Definitions and Reduction-based Algorithms

Fairness without Harm: Decoupled Classifiers with Preference Guarantees

Proportionally Fair Clustering

Learning Fair Representations via an Adversarial Framework

Matroids, Matchings, and Fairness

Lessons for artificial intelligence from the study of natural stupidity

Learning Optimal and Fair Decision Trees for Non-Discriminative Decision-Making

Fairness in Recommendation Ranking through Pairwise Comparisons

Stable and Fair Classification

Capuchin: Causal Database Repair for Algorithmic Fairness

Accuracy and Fairness for Juvenile Justice Risk Assessments

Mathematical Notions vs. Human Perception of Fairness: A Descriptive Approach to Fairness for Machine Learning

Policy Learning for Fairness in Ranking

Scalable Fair Clustering

Fair Decisions Despite Imperfect Predictions

Improved Adversarial Learning for Fair Classification

Measuring the Biases that Matter: The Ethical and Casual Foundations for Measures of Fairness in Algorithms

Repairing without Retraining: Avoiding Disparate Impact with Counterfactual Distributions

A Framework for Understanding Unintended Consequences of Machine Learning

Fair Transfer Learning with Missing Protected Attributes

Bayesian Fairness

Identifying and Correcting Label Bias in Machine Learning

Putting Fairness Principles into Practice: Challenges, Metrics, and Improvements

Fair Algorithms for Clustering

Measuring and Mitigating Unintended Bias in Text Classification

Meritocratic Fairness for Infinite and Contextual Bandits

Algorithmic Pollution: Understanding and Responding to Negative Consequences of Algorithmic Decision-Making

Probabilistic verification of fairness properties via concentration

Fighting Fire with Fire: Using Antidote Data to Improve Polarization and Fairness of Recommender Systems

Data Statements for Natural Language Processing: Toward Mitigating System Bias and Enabling Better Science

Racial categories in machine learning

On the cost of essentially fair clusterings

50 Years of Test (Un)fairness: Lessons for Machine Learning

Equality of Voice: Towards Fair Representation in Crowdsourced Top-K Recommendations

Prediction-Based Decisions and Fairness: A Catalogue of Choices, Assumptions, and Definitions

Aequitas: A Bias and Fairness Audit Toolkit

A Neural Network Framework for Fair Classifier

Taking Advantage of Multitask Learning for Fair Classification

100

Fairness-Aware Tensor-Based Recommendation

101

A General Framework for Fair Regression

102

AI Fairness 360: An Extensible Toolkit for Detecting, Understanding, and Mitigating Unwanted Algorithmic Bias

103

Active Fairness in Algorithmic Decision Making

104

Calibrated recommendations

105

Optimization with Non-Differentiable Constraints with Applications to Fairness, Recall, Churn, and Other Goals

106

Learning Optimal Fair Policies

107

Fair Algorithms for Learning in Allocation Problems

108

The Implicit Fairness Criterion of Unconstrained Learning

109

The Disparate Effects of Strategic Manipulation

110

The Social Cost of Strategic Classification

111

Exploring author gender in book rating and recommendation

112

Reducing Gender Bias in Abusive Language Detection

113

A Causal Bayesian Networks Viewpoint on Fairness

114

Data Pre-Processing for Discrimination Prevention: Information-Theoretic Optimization and Analysis

115

Detecting racial bias in algorithms and machine learning

116

The Measure and Mismeasure of Fairness: A Critical Review of Fair Machine Learning

117

Nonconvex Optimization for Regression with Fairness Constraints

118

A Unified Approach to Quantifying Algorithmic Unfairness: Measuring Individual &Group Unfairness via Inequality Indices

119

Fairness In Reciprocal Recommendations: A Speed-Dating Study

120

Preventing Disparate Treatment in Sequential Decision Making

121

Achieving Fairness through Adversarial Learning: an Application to Recidivism Prediction

123

Classification with Fairness Constraints: A Meta-Algorithm with Provable Guarantees

124

Fairness Behind a Veil of Ignorance: A Welfare Analysis for Automated Decision Making

125

Obtaining Fairness using Optimal Transport Theory

126

Residual Unfairness in Fair Machine Learning from Prejudiced Data

127

Causal Interventions for Fairness

128

iFair: Learning Individually Fair Data Representations for Algorithmic Decision Making

129

Multiaccuracy: Black-Box Post-Processing for Fairness in Classification

130

Why Is My Classifier Discriminatory?

131

Fairness Definitions Explained

132

FairGAN: Fairness-aware Generative Adversarial Networks

134

Integrating Technical and Legal Concepts of Privacy

135

Examining Gender and Race Bias in Two Hundred Sentiment Analysis Systems

136

The Dataset Nutrition Label: A Framework To Drive Higher Data Quality Standards

137

Equity of Attention: Amortizing Individual Fairness in Rankings

138

Algorithmic Decision Making in the Presence of Unmeasured Confounding

139

Enhancing the Accuracy and Fairness of Human Decision Making

140

Non-Discriminatory Machine Learning through Convex Fairness Criteria

141

Adaptive Sensitive Reweighting to Mitigate Bias in Fairness-aware Classification

142

Addressing Age-Related Bias in Sentiment Analysis

143

Unleashing Linear Optimizers for Group-Fair Learning and Optimization

144

Learning with Complex Loss Functions and Constraints

145

Delayed Impact of Fair Machine Learning

146

Fairness Through Computationally-Bounded Awareness

147

A Reductions Approach to Fair Classification

148

Human Perceptions of Fairness in Algorithmic Decision Making: A Case Study of Criminal Risk Prediction

149

Path-Specific Counterfactual Fairness

150

Fairness of Exposure in Rankings

151

Learning Adversarially Fair and Transferable Representations

152

Fair Clustering Through Fairlets

153

A comparative study of fairness-enhancing interventions in machine learning

154

Preventing Disparities: Bayesian and Frequentist Methods for Assessing Fairness in Machine-Learning Decision-Support Models

155

Privacy preserving clustering with constraints

156

Fairness and Accountability Design Needs for Algorithmic Support in High-Stakes Public Sector Decision-Making

157

Online Learning with an Unknown Fairness Metric

158

Empirical Risk Minimization under Fairness Constraints

159

Mitigating Unwanted Biases with Adversarial Learning

160

Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification

161

All The Cool Kids, How Do They Fit In?: Popularity and Demographic Biases in Recommender Evaluation and Effectiveness

162

Balanced Neighborhoods for Multi-sided Fairness in Recommendation

163

All Models are Wrong but many are Useful: Variable Importance for Black-Box, Proprietary, or Misspecified Prediction Models, using Model Class Reliance

164

Demographics and discussion influence views on algorithmic fairness

165

Gender differences in beliefs about algorithmic fairness

166

Matching code and law: achieving algorithmic fairness with optimal transport

167

Interventions over Predictions: Reframing the Ethical Debate for Actuarial Risk Assessment

168

Fairness in Machine Learning: Lessons from Political Philosophy

169

When Worlds Collide: Integrating Different Counterfactual Assumptions in Fairness

170

Calibration for the (Computationally-Identifiable) Masses

171

Does mitigating ML's impact disparity require treatment disparity?

172

Preventing Fairness Gerrymandering: Auditing and Learning for Subgroup Fairness

173

Towards a Code of Ethics for Artificial Intelligence

174

Fairer machine learning in the real world: Mitigating discrimination without collecting sensitive data

175

FairSquare: probabilistic verification of program fairness

176

Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy

177

Fair Kernel Learning

178

On Fairness and Calibration

179

Fairness testing: testing software for discrimination

180

Fair, Transparent, and Accountable Algorithmic Decision-making Processes

181

The Selective Labels Problem: Evaluating Algorithmic Predictions in the Presence of Unobservables

183

Decoupled Classifiers for Group-Fair and Efficient Machine Learning

184

Data Decisions and Theoretical Implications when Adversarially Learning Fair Representations

185

Multisided Fairness for Recommendation

186

Fairer and more accurate, but for whom?

187

Penalizing Unfairness in Binary Classification

188

Racial Disparity in Natural Language Processing: A Case Study of Social Media African-American English

189

Runaway Feedback Loops in Predictive Policing

190

The Authority of "Fair" in Machine Learning

191

FA*IR: A Fair Top-k Ranking Algorithm

192

Avoiding Discrimination through Causal Reasoning

193

A Convex Framework for Fair Regression

194

Fair Inference on Outcomes

195

The cost of fairness in classification

196

Beyond Parity: Fairness Objectives for Collaborative Filtering

197

Algorithmic Regulation: A Critical Interrogation

198

A Unified Approach to Interpreting Model Predictions

199

Balanced Policy Evaluation and Learning

200

Semantic Similarity Metrics for Evaluating Source Code Summarization

201

On Measuring Bias in Online Information

202

Gender and Dialect Bias in YouTube’s Automatic Captions

203

Fairness in Criminal Justice Risk Assessments: The State of the Art

204

Counterfactual Fairness

205

TL;DR

Abstract

Authors

References360 items

Learning fair representations via an adversarial framework

An Ontology for Fairness Metrics

Improving fairness generalization through a sample-robust optimization method

Improved Approximation for Fair Correlation Clustering

A Survey on the Fairness of Recommender Systems

TL;DR

Abstract

Authors

References360 items

Learning fair representations via an adversarial framework

An Ontology for Fairness Metrics

Improving fairness generalization through a sample-robust optimization method

Improved Approximation for Fair Correlation Clustering

A Survey on the Fairness of Recommender Systems

Group Meritocratic Fairness in Linear Contextual Bandits

Fairness in Ranking, Part I: Score-Based Ranking

Fairness in Ranking, Part II: Learning-to-Rank and Recommender Systems

CPFair: Personalized Consumer and Producer Fairness Re-ranking for Recommender Systems

Fairneuron: Improving Deep Neural Network Fairness with Adversary Games on Selective Neurons

Marrying Fairness and Explainability in Supervised Learning

Search-based fairness testing for regression-based machine learning systems

Ethics Guidelines for Trustworthy AI

An Interactive Approach to Bias Mitigation in Machine Learning

Measure Twice, Cut Once: Quantifying Bias and Fairness in Deep Neural Networks

A survey on datasets for fairness‐aware machine learning

Fairness metrics and bias mitigation strategies for rating predictions

Trustworthy AI: A Computational Perspective

Deep Clustering based Fair Outlier Detection

Using Pareto simulated annealing to address algorithmic bias in machine learning

Analyzing the impact of missing values and selection bias on fairness

Can Active Learning Preemptively Mitigate Fairness Issues?

Mitigating Bias in Algorithmic Systems—A Fish-eye View

Missing the missing values: The ugly duckling of fairness in machine learning

Fairness of Exposure in Stochastic Bandits

Adaptive Sampling for Minimax Fair Classification

TARA: Training and Representation Alteration for AI Fairness and Domain Generalization

FairBatch: Batch Selection for Model Fairness

FAIR: Fair Adversarial Instance Re-weighting

Reuben

Achieving User-Side Fairness in Contextual Bandits

Fair Algorithms for Multi-Agent Multi-Armed Bandits

A Notion of Individual Fairness for Clustering

Counterfactual fairness: removing direct effects through regularization

Fair Prediction with Endogenous Behavior

Individual Fairness Revisited: Transferring Techniques from Adversarial Robustness

Fair Correlation Clustering

The false promise of risk assessments: epistemic reform and the limits of fairness

Algorithmic realism: expanding the boundaries of algorithmic thought

Algorithmic Fairness

FAE: A Fairness-Aware Ensemble Framework

Preventing undesirable behavior of intelligent machines

Predictive Biases in Natural Language Processing Models: A Conceptual Framework and Overview

Censored and Fair Universal Representations using Generative Adversarial Models

AdaFair: Cumulative Fairness Adaptive Boosting

Fair Coresets and Streaming Algorithms for Fair k-means

A Survey on Bias and Fairness in Machine Learning

Data Management for Causal Algorithmic Fairness

Integrating Ethics within Machine Learning Courses

Achieving Causal Fairness through Generative Adversarial Networks

Achieving Fairness in the Stochastic Multi-armed Bandit Problem

One-Network Adversarial Fairness

Allocating Interventions Based on Predicted Outcomes: A Case Study on Homelessness Services

Wasserstein Fair Classification

Interventional Fairness: Causal Database Repair for Algorithmic Fairness

Mitigating bias in algorithmic hiring: evaluating claims and practices

Mitigating Gender Bias in Natural Language Processing: Literature Review

Mitigating Bias in Algorithmic Employment Screening: Evaluating Claims and Practices

Why Machine Learning May Lead to Unfairness: Evidence from Risk Assessment for Juvenile Justice in Catalonia

Fair Division Without Disparate Impact

Democratizing Algorithmic Fairness

Fair Regression: Quantitative Definitions and Reduction-based Algorithms

Fairness without Harm: Decoupled Classifiers with Preference Guarantees

Proportionally Fair Clustering

Learning Fair Representations via an Adversarial Framework

Matroids, Matchings, and Fairness

Lessons for artificial intelligence from the study of natural stupidity