A survey of algorithmic recourse: definitions, formulations, solutions, and prospects

Abstract

Machine learning is increasingly used to inform decision-making in sensitive situations where decisions have consequential effects on individuals' lives. In these settings, in addition to requiring models to be accurate and robust, socially relevant values such as fairness, privacy, accountability, and explainability play an important role for the adoption and impact of said technologies. In this work, we focus on algorithmic recourse, which is concerned with providing explanations and recommendations to individuals who are unfavourably treated by automated decision-making systems. We first perform an extensive literature review, and align the efforts of many authors by presenting unified definitions, formulations, and solutions to recourse. Then, we provide an overview of the prospective research directions towards which the community may engage, challenging existing assumptions and making explicit connections to other ethical challenges such as security, privacy, and fairness.

References in corpus (58)

• Towards A Rigorous Science of Interpretable Machine Learning
• Methods for Interpreting and Understanding Deep Neural Networks
• Explaining Machine Learning Classifiers through Diverse Counterfactual Explanations
• Explaining Explanations in AI
• Certified Adversarial Robustness via Randomized Smoothing
• The What-If Tool: Interactive Probing of Machine Learning Models
• Actionable Recourse in Linear Classification
• The Measure and Mismeasure of Fairness
• Towards Robust Interpretability with Self-Explaining Neural Networks
• Black-box Adversarial Attacks with Limited Queries and Information
• Fairness in Machine Learning
• The Hidden Assumptions Behind Counterfactual Explanations and Principal Reasons
• Local Rule-Based Explanations of Black Box Decision Systems
• Contrastive Explanation: A Structural-Model Approach
• Causal Discovery from a Mixture of Experimental and Observational Data
• Learning Model-Agnostic Counterfactual Explanations for Tabular Data
• Explanations can be manipulated and geometry is to blame
• The Disparate Effects of Strategic Manipulation
• An Evaluation of the Human-Interpretability of Explanation
• Causality for Machine Learning
• Towards Realistic Individual Recourse and Actionable Explanations in Black-Box Decision Making Systems
• Interpretability via Model Extraction
• PRINCE: Provider-side Interpretability with Counterfactual Explanations in Recommender Systems
• Counterfactual Explanations for Machine Learning on Multivariate Time Series Data
• On the Robustness of Interpretability Methods
• Causal Discovery from Changes
• Algorithmic recourse under imperfect causal knowledge: a probabilistic approach
• Algorithmic Recourse: from Counterfactual Explanations to Interventions
• Unlabeled Data Improves Adversarial Robustness
• Generating Counterfactual and Contrastive Explanations using SHAP
• Explaining Data-Driven Decisions made by AI Systems: The Counterfactual Approach
• Measurable Counterfactual Local Explanations for Any Classifier
• Imperceptible Adversarial Attacks on Tabular Data
• Inverse Classification for Comparison-based Interpretability in Machine Learning
• Machine Learning Explainability for External Stakeholders
• Recommendations and User Agency: The Reachability of Collaboratively-Filtered Information
• Counterfactual Visual Explanations
• Interpreting Neural Network Judgments via Minimal, Stable, and Symbolic Corrections
• PermuteAttack: Counterfactual Explanation of Machine Learning Credit Scorecards
• Counterfactual Explanation Based on Gradual Construction for Deep Networks
• Latent Space Oddity: on the Curvature of Deep Generative Models
• Strategic Classification is Causal Modeling in Disguise
• Model extraction from counterfactual explanations
• On Counterfactual Explanations under Predictive Multiplicity
• Issues with post-hoc counterfactual explanations: a discussion
• Fairwashing: the risk of rationalization
• Privacy-preserving Prediction
• On the computation of counterfactual explanations -- A survey
• Equalizing Recourse across Groups
• DECE: Decision Explorer with Counterfactual Explanations for Machine Learning Models
• Efficient computation of counterfactual explanations of LVQ models
• Fast Real-time Counterfactual Explanations
• Algorithmic Recourse in the Wild: Understanding the Impact of Data and Model Shifts
• Synthesizing Action Sequences for Modifying Model Decisions
• Distal Explanations for Model-free Explainable Reinforcement Learning
• From Predictions to Decisions: Using Lookahead Regularization
• EMAP: Explanation by Minimal Adversarial Perturbation
• Prophit: Causal inverse classification for multiple continuously valued treatment policies

Cited by in corpus (24)

• Algorithmic recourse under imperfect causal knowledge: a probabilistic approach
• On Tackling Explanation Redundancy in Decision Trees
• Towards Robust and Reliable Algorithmic Recourse
• Model-Based Counterfactual Synthesizer for Interpretation
• Rational Shapley Values
• Consequence-aware Sequential Counterfactual Generation
• CARLA: A Python Library to Benchmark Algorithmic Recourse and Counterfactual Explanation Algorithms
• Optimal Counterfactual Explanations in Tree Ensembles
• Conditional Generative Models for Counterfactual Explanations
• Understanding Consumer Preferences for Explanations Generated by XAI Algorithms
• A Survey on the Robustness of Feature Importance and Counterfactual Explanations
• Counterfactual Explanations in Sequential Decision Making Under Uncertainty
• Exploring Counterfactual Explanations Through the Lens of Adversarial Examples: A Theoretical and Empirical Analysis
• Feature Attributions and Counterfactual Explanations Can Be Manipulated
• Multi-Agent Algorithmic Recourse
• Counterfactual Explanations for Machine Learning: Challenges Revisited
• Model-agnostic and Scalable Counterfactual Explanations via Reinforcement Learning
• Counterfactual Explanations Can Be Manipulated
• Linear Classifiers that Encourage Constructive Adaptation
• MCCE: Monte Carlo sampling of realistic counterfactual explanations
• Strategic Classification Made Practical
• Everything is Relative: Understanding Fairness with Optimal Transport
• Algorithmic Recourse in Partially and Fully Confounded Settings Through Bounding Counterfactual Effects
• Low-Cost Algorithmic Recourse for Users With Uncertain Cost Functions