Counterfactual Fairness

Abstract

Machine learning can impact people with legal or ethical consequences when it is used to automate decisions in areas such as insurance, lending, hiring, and predictive policing. In many of these scenarios, previous decisions have been made that are unfairly biased against certain subpopulations, for example those of a particular race, gender, or sexual orientation. Since this past data may be biased, machine learning predictors must account for this to avoid perpetuating or creating discriminatory practices. In this paper, we develop a framework for modeling fairness using tools from causal inference. Our definition of counterfactual fairness captures the intuition that a decision is fair towards an individual if it is the same in (a) the actual world and (b) a counterfactual world where the individual belonged to a different demographic group. We demonstrate our framework on a real-world problem of fair prediction of success in law school.

References in corpus (4)

• Equality of Opportunity in Supervised Learning
• Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddings
• Fairness in Criminal Justice Risk Assessments: The State of the Art
• Causal Inference through a Witness Protection Program

Cited by in corpus (268)

• Shortcut Learning in Deep Neural Networks
• Explaining Machine Learning Classifiers through Diverse Counterfactual Explanations
• Prediction-Based Decisions and Fairness: A Catalogue of Choices, Assumptions, and Definitions
• Fairness of Exposure in Rankings
• Fairness in Machine Learning: A Survey
• The Measure and Mismeasure of Fairness
• Underspecification Presents Challenges for Credibility in Modern Machine Learning
• Fairness Testing: Testing Software for Discrimination
• Fairness in Machine Learning
• 50 Years of Test (Un)fairness: Lessons for Machine Learning
• AI Fairness 360: An Extensible Toolkit for Detecting, Understanding, and Mitigating Unwanted Algorithmic Bias
• Towards Out-Of-Distribution Generalization: A Survey
• A Clarification of the Nuances in the Fairness Metrics Landscape
• Fairness in Credit Scoring: Assessment, Implementation and Profit Implications
• An Empirical Characterization of Fair Machine Learning For Clinical Risk Prediction
• On Formalizing Fairness in Prediction with Machine Learning
• Personalized Counterfactual Fairness in Recommendation
• FairSight: Visual Analytics for Fairness in Decision Making
• EDITS: Modeling and Mitigating Data Bias for Graph Neural Networks
• Causality for Machine Learning
• Racial categories in machine learning
• Machine Learning Testing: Survey, Landscapes and Horizons
• Gender Bias in Coreference Resolution: Evaluation and Debiasing Methods
• Deep Structural Causal Models for Tractable Counterfactual Inference
• Operationalizing Individual Fairness with Pairwise Fair Representations
• Empirical Risk Minimization under Fairness Constraints
• A Framework for Understanding Sources of Harm throughout the Machine Learning Life Cycle
• Fairness in Algorithmic Decision Making: An Excursion Through the Lens of Causality
• Causal Reasoning for Algorithmic Fairness
• Addressing Bias in Generative AI: Challenges and Research Opportunities in Information Management
• Algorithmic recourse under imperfect causal knowledge: a probabilistic approach
• Algorithmic Recourse: from Counterfactual Explanations to Interventions
• What's Sex Got To Do With Fair Machine Learning?
• Human Perceptions of Fairness in Algorithmic Decision Making: A Case Study of Criminal Risk Prediction
• Causal Mediation Analysis for Interpreting Neural NLP: The Case of Gender Bias
• Ethics-Based Auditing of Automated Decision-Making Systems: Intervention Points and Policy Implications
• Machine learning fairness notions: Bridging the gap with real-world applications
• Visual Analysis of Discrimination in Machine Learning
• Fairness in Risk Assessment Instruments: Post-Processing to Achieve Counterfactual Equalized Odds
• Fairness Score and Process Standardization: Framework for Fairness Certification in Artificial Intelligence Systems
• Limitations of P-Values and $R^2$ for Stepwise Regression Building: A Fairness Demonstration in Health Policy Risk Adjustment
• A Model-Agnostic Causal Learning Framework for Recommendation using Search Data
• Policy Learning for Fairness in Ranking
• Robust Optimization for Fairness with Noisy Protected Groups
• Simpson's paradox in Covid-19 case fatality rates: a mediation analysis of age-related causal effects
• Getting a CLUE: A Method for Explaining Uncertainty Estimates
• Amazon SageMaker Clarify: Machine Learning Bias Detection and Explainability in the Cloud
• A survey on datasets for fairness-aware machine learning
• Unbiased Scene Graph Generation from Biased Training
• Survey on Causal-based Machine Learning Fairness Notions
• On Adversarial Bias and the Robustness of Fair Machine Learning
• FairBatch: Batch Selection for Model Fairness
• Asymmetric Shapley values: incorporating causal knowledge into model-agnostic explainability
• Mitigating Unfairness via Evolutionary Multi-objective Ensemble Learning
• Socially Responsible AI Algorithms: Issues, Purposes, and Challenges
• A Survey on Intersectional Fairness in Machine Learning: Notions, Mitigation, and Challenges
• Generative Counterfactual Introspection for Explainable Deep Learning
• The Sensitivity of Counterfactual Fairness to Unmeasured Confounding
• Balanced Datasets Are Not Enough: Estimating and Mitigating Gender Bias in Deep Image Representations
• Noise-tolerant fair classification
• A Human-Centered Review of Algorithms in Decision-Making in Higher Education
• On the Legal Compatibility of Fairness Definitions
• Causal Collaborative Filtering
• Algorithmic Unfairness through the Lens of EU Non-Discrimination Law: Or Why the Law is not a Decision Tree
• Fairness via Representation Neutralization
• Learning Certified Individually Fair Representations
• Debiasing Learning for Membership Inference Attacks Against Recommender Systems
• Algorithmic Decision Making with Conditional Fairness
• Explaining Deep Learning Models using Causal Inference
• Explainability Auditing for Intelligent Systems: A Rationale for Multi-Disciplinary Perspectives
• Adequate and fair explanations
• Multi-sided Exposure Bias in Recommendation
• Whither Bias Goes, I Will Go: An Integrative, Systematic Review of Algorithmic Bias Mitigation
• Fairness risk measures
• Towards Fair Graph Neural Networks via Graph Counterfactual
• Fairness Evaluation in Presence of Biased Noisy Labels
• Reducing Sentiment Bias in Language Models via Counterfactual Evaluation
• Learning Sparse Nonparametric DAGs
• Fairness without the sensitive attribute via Causal Variational Autoencoder
• Can Fairness be Automated? Guidelines and Opportunities for Fairness-aware AutoML
• When Causal Intervention Meets Adversarial Examples and Image Masking for Deep Neural Networks
• Fairness in Agreement With European Values: An Interdisciplinary Perspective on AI Regulation
• Bias in Data-driven AI Systems -- An Introductory Survey
• End-To-End Bias Mitigation: Removing Gender Bias in Deep Learning
• Directional Bias Amplification
• Towards a Unified Framework for Fair and Stable Graph Representation Learning
• On conditional parity as a notion of non-discrimination in machine learning
• Review of Mathematical frameworks for Fairness in Machine Learning
• Toward Operationalizing Pipeline-aware ML Fairness: A Research Agenda for Developing Practical Guidelines and Tools
• Ensuring Fairness Beyond the Training Data
• HARK Side of Deep Learning -- From Grad Student Descent to Automated Machine Learning
• Conditional Learning of Fair Representations
• Machine learning and AI research for Patient Benefit: 20 Critical Questions on Transparency, Replicability, Ethics and Effectiveness
• No computation without representation: Avoiding data and algorithm biases through diversity
• Fairness-Aware Explainable Recommendation over Knowledge Graphs
• Verifying Individual Fairness in Machine Learning Models
• Subverting machines, fluctuating identities: Re-learning human categorization
• Toward Fairness in Speech Recognition: Discovery and mitigation of performance disparities
• A fuzzy-rough uncertainty measure to discover bias encoded explicitly or implicitly in features of structured pattern classification datasets
• FairCanary: Rapid Continuous Explainable Fairness
• Artificial mental phenomena: Psychophysics as a framework to detect perception biases in AI models
• Fairness Through Robustness: Investigating Robustness Disparity in Deep Learning
• Tuning Fairness by Balancing Target Labels
• Man is to Person as Woman is to Location: Measuring Gender Bias in Named Entity Recognition
• Generative causal explanations of black-box classifiers
• BeFair: Addressing Fairness in the Banking Sector
• FairVis: Visual Analytics for Discovering Intersectional Bias in Machine Learning
• Counterfactual Vision-and-Language Navigation via Adversarial Path Sampling
• Causal Inference in Natural Language Processing: Estimation, Prediction, Interpretation and Beyond
• Understanding Agent Incentives using Causal Influence Diagrams. Part I: Single Action Settings
• ESR: Ethics and Society Review of Artificial Intelligence Research
• Capuchin: Causal Database Repair for Algorithmic Fairness
• A Local Method for Identifying Causal Relations under Markov Equivalence
• Geographic ratemaking with spatial embeddings
• Incentives for Responsiveness, Instrumental Control and Impact
• Causal Interpretability for Machine Learning -- Problems, Methods and Evaluation
• All of the Fairness for Edge Prediction with Optimal Transport
• Causal Discovery with General Non-Linear Relationships Using Non-Linear ICA
• MultiVerse: Causal Reasoning using Importance Sampling in Probabilistic Programming
• Causal Interventions for Fairness
• Personalized explanation in machine learning: A conceptualization
• Causal Modeling for Fairness in Dynamical Systems
• Privacy for All: Demystify Vulnerability Disparity of Differential Privacy against Membership Inference Attack
• Improvement-Focused Causal Recourse (ICR)
• Responsible AI: Gender bias assessment in emotion recognition
• Modeling Techniques for Machine Learning Fairness: A Survey
• Algorithms for Causal Reasoning in Probability Trees
• Regulatory Instruments for Fair Personalized Pricing
• Equal Opportunity and Affirmative Action via Counterfactual Predictions
• Fairness On The Ground: Applying Algorithmic Fairness Approaches to Production Systems
• Removing biased data to improve fairness and accuracy
• Counterfactual Reasoning for Fair Clinical Risk Prediction
• A survey of bias in Machine Learning through the prism of Statistical Parity for the Adult Data Set
• Biases in Generative Art -- A Causal Look from the Lens of Art History
• Characterizing Intersectional Group Fairness with Worst-Case Comparisons
• Deontological Ethics By Monotonicity Shape Constraints
• Dynamic fairness - Breaking vicious cycles in automatic decision making
• Counterfactual fairness: removing direct effects through regularization
• Principal Fairness for Human and Algorithmic Decision-Making
• Disentangled Representations from Non-Disentangled Models
• Revealing Neural Network Bias to Non-Experts Through Interactive Counterfactual Examples
• On the Fairness of Disentangled Representations
• Fairness-aware Model-agnostic Positive and Unlabeled Learning
• FairSample: Training Fair and Accurate Graph Convolutional Neural Networks Efficiently
• Fairkit, Fairkit, on the Wall, Who's the Fairest of Them All? Supporting Data Scientists in Training Fair Models
• Counterfactual Invariance to Spurious Correlations: Why and How to Pass Stress Tests
• Detection and Mitigation of Bias in Ted Talk Ratings
• DORO: Distributional and Outlier Robust Optimization
• Using Counterfactuals to Improve Causal Inferences from Visualizations
• Learning Fair Rule Lists
• Unifying local and global model explanations by functional decomposition of low dimensional structures
• To Split or Not to Split: The Impact of Disparate Treatment in Classification
• Machine learning and behavioral economics for personalized choice architecture
• Gender Slopes: Counterfactual Fairness for Computer Vision Models by Attribute Manipulation
• Statistical Equity: A Fairness Classification Objective
• Fairness in Forecasting and Learning Linear Dynamical Systems
• Disaggregated Interventions to Reduce Inequality
• Perfectly Parallel Fairness Certification of Neural Networks
• Fair and Unbiased Algorithmic Decision Making: Current State and Future Challenges
• The Fairness-Accuracy Pareto Front
• With Malice Towards None: Assessing Uncertainty via Equalized Coverage
• Stereotype-Free Classification of Fictitious Faces
• Pareto Efficient Fairness in Supervised Learning: From Extraction to Tracing
• Bandit Algorithms for Precision Medicine
• Counterfactuals Modulo Temporal Logics
• On the Privacy Risks of Algorithmic Fairness
• Fair Division Without Disparate Impact
• An Introduction to Algorithmic Fairness
• Fairness in Forecasting of Observations of Linear Dynamical Systems
• In-processing User Constrained Dominant Sets for User-Oriented Fairness in Recommender Systems
• Information Theoretic Measures for Fairness-aware Feature Selection
• Unintended Selection: Persistent Qualification Rate Disparities and Interventions
• Strategic Instrumental Variable Regression: Recovering Causal Relationships From Strategic Responses
• Explainable Deep Learning for Uncovering Actionable Scientific Insights for Materials Discovery and Design
• Dynamic Modeling and Equilibria in Fair Decision Making
• Cross-model Fairness: Empirical Study of Fairness and Ethics Under Model Multiplicity
• Fairness in generative modeling
• Identifying Biased Subgroups in Ranking and Classification
• FairMILE: Towards an Efficient Framework for Fair Graph Representation Learning
• The Authors Matter: Understanding and Mitigating Implicit Bias in Deep Text Classification
• Interpretable bias mitigation for textual data: Reducing gender bias in patient notes while maintaining classification performance
• Learning Individually Fair Classifier with Path-Specific Causal-Effect Constraint
• Optimal Training of Fair Predictive Models
• State of the Art in Fair ML: From Moral Philosophy and Legislation to Fair Classifiers
• Fair Classification and Social Welfare
• Fairness and Missing Values
• DIVINE: Diverse Influential Training Points for Data Visualization and Model Refinement
• Bias, Consistency, and Partisanship in U.S. Asylum Cases: A Machine Learning Analysis of Extraneous Factors in Immigration Court Decisions
• Intersectionality and Testimonial Injustice in Medical Records
• Technologies for Trustworthy Machine Learning: A Survey in a Socio-Technical Context
• A Causal Lens for Peeking into Black Box Predictive Models: Predictive Model Interpretation via Causal Attribution
• Beyond Individual and Group Fairness
• Pairwise Fairness for Ranking and Regression
• Fair Hate Speech Detection through Evaluation of Social Group Counterfactuals
• Fairness criteria through the lens of directed acyclic graphical models
• Abstracting Fairness: Oracles, Metrics, and Interpretability
• Effects of algorithmic flagging on fairness: quasi-experimental evidence from Wikipedia
• Fairness Through Regularization for Learning to Rank
• SSCR: Iterative Language-Based Image Editing via Self-Supervised Counterfactual Reasoning
• Priority-based Post-Processing Bias Mitigation for Individual and Group Fairness
• Metrics and methods for a systematic comparison of fairness-aware machine learning algorithms
• Individually Fair Ranking
• User Acceptance of Gender Stereotypes in Automated Career Recommendations
• Removing Spurious Features can Hurt Accuracy and Affect Groups Disproportionately
• Correspondences between Privacy and Nondiscrimination: Why They Should Be Studied Together
• Fairness Through Causal Awareness: Learning Latent-Variable Models for Biased Data
• Causal Inference with Deep Causal Graphs
• The Random Conditional Distribution for Higher-Order Probabilistic Inference
• Under the Radar -- Auditing Fairness in ML for Humanitarian Mapping
• Combining distributive ethics and causal Inference to make trade-offs between austerity and population health
• Fairness Through Counterfactual Utilities
• Linear Classifiers that Encourage Constructive Adaptation
• Statistical inference for individual fairness
• Robust Semantic Interpretability: Revisiting Concept Activation Vectors
• VERB: Visualizing and Interpreting Bias Mitigation Techniques for Word Representations
• Probably Approximately Correct Constrained Learning
• Actionable Attribution Maps for Scientific Machine Learning
• Oblivious Data for Fairness with Kernels
• DeDUCE: Generating Counterfactual Explanations Efficiently
• Improving Fair Predictions Using Variational Inference In Causal Models
• Fairness Constraints in Semi-supervised Learning
• Assessing Disparate Impacts of Personalized Interventions: Identifiability and Bounds
• Locating disparities in machine learning
• Alleviating Privacy Attacks via Causal Learning
• Identifying Counterfactual Queries with the R Package cfid
• Supervised Feature Compression based on Counterfactual Analysis
• MobilityMirror: Bias-Adjusted Transportation Datasets
• Simulating counterfactuals
• Equality of opportunity in travel behavior prediction with deep neural networks and discrete choice models
• Problem Learning: Towards the Free Will of Machines
• Fair inference on error-prone outcomes
• Model Mis-specification and Algorithmic Bias
• Probabilistic Verification of Fairness Properties via Concentration
• Improving Counterfactual Generation for Fair Hate Speech Detection
• Generative Models for Security: Attacks, Defenses, and Opportunities
• The invisible power of fairness. How machine learning shapes democracy
• Metric-Free Individual Fairness with Cooperative Contextual Bandits
• Towards Robust Classification Model by Counterfactual and Invariant Data Generation
• Causality in Neural Networks -- An Extended Abstract
• Fairness-guided SMT-based Rectification of Decision Trees and Random Forests
• Towards classification parity across cohorts
• CROCS: Clustering and Retrieval of Cardiac Signals Based on Patient Disease Class, Sex, and Age
• Inherent Trade-offs in the Fair Allocation of Treatments
• Independent Ethical Assessment of Text Classification Models: A Hate Speech Detection Case Study
• Causal Learning for Socially Responsible AI
• Learning Fair Representations for Recommendation: A Graph-based Perspective
• Hidden Technical Debts for Fair Machine Learning in Financial Services
• Fairness-aware Outlier Ensemble
• A Distributed Fair Machine Learning Framework with Private Demographic Data Protection
• Affirmative Action Policies for Top-k Candidates Selection, With an Application to the Design of Policies for University Admissions
• GroupMixNorm Layer for Learning Fair Models
• I Wish I Would Have Loved This One, But I Didn't -- A Multilingual Dataset for Counterfactual Detection in Product Reviews
• Non-Comparative Fairness for Human-Auditing and Its Relation to Traditional Fairness Notions
• Reducing Unintended Bias of ML Models on Tabular and Textual Data
• VACA: Design of Variational Graph Autoencoders for Interventional and Counterfactual Queries
• A Ladder of Causal Distances
• Debiasing Credit Scoring using Evolutionary Algorithms
• Fairness through Equality of Effort
• Logic Constraints to Feature Importances
• Local Justice and the Algorithmic Allocation of Societal Resources
• When black box algorithms are (not) appropriate: a principled prediction-problem ontology
• Data Management for Causal Algorithmic Fairness
• Comprehensible Counterfactual Explanation on Kolmogorov-Smirnov Test
• Counterfactually Fair Prediction Using Multiple Causal Models
• A Sociotechnical View of Algorithmic Fairness
• On the Identification of Fair Auditors to Evaluate Recommender Systems based on a Novel Non-Comparative Fairness Notion
• Identifying Best Fair Intervention
• Explaining Algorithmic Fairness Through Fairness-Aware Causal Path Decomposition