Preserving Causal Constraints in Counterfactual Explanations for Machine Learning Classifiers

Abstract

To construct interpretable explanations that are consistent with the original ML model, counterfactual examples---showing how the model's output changes with small perturbations to the input---have been proposed. This paper extends the work in counterfactual explanations by addressing the challenge of feasibility of such examples. For explanations of ML models in critical domains such as healthcare and finance, counterfactual examples are useful for an end-user only to the extent that perturbation of feature inputs is feasible in the real world. We formulate the problem of feasibility as preserving causal relationships among input features and present a method that uses (partial) structural causal models to generate actionable counterfactuals. When feasibility constraints cannot be easily expressed, we consider an alternative mechanism where people can label generated CF examples on feasibility: whether it is feasible to intervene and realize the candidate CF example from the original input. To learn from this labelled feasibility data, we propose a modified variational auto encoder loss for generating CF examples that optimizes for feasibility as people interact with its output. Our experiments on Bayesian networks and the widely used ''Adult-Income'' dataset show that our proposed methods can generate counterfactual explanations that better satisfy feasibility constraints than existing methods.. Code repository can be accessed here: \textit{https://github.com/divyat09/cf-feasibility}

arXiv comment2019 NeurIPS Workshop on Do the right thing: Machine learning and Causal Inference for improved decision making

References in corpus (6)

• Explaining Machine Learning Classifiers through Diverse Counterfactual Explanations
• Invariant Risk Minimization
• The Hidden Assumptions Behind Counterfactual Explanations and Principal Reasons
• A Meta-Transfer Objective for Learning to Disentangle Causal Mechanisms
• Algorithmic Recourse: from Counterfactual Explanations to Interventions
• Explaining Visual Models by Causal Attribution

Cited by in corpus (32)

• Deep Neural Networks and Tabular Data: A Survey
• A Survey on the Explainability of Supervised Machine Learning
• Algorithmic recourse under imperfect causal knowledge: a probabilistic approach
• Towards Robust and Reliable Algorithmic Recourse
• Asymmetric Shapley values: incorporating causal knowledge into model-agnostic explainability
• A Framework and Benchmarking Study for Counterfactual Generating Methods on Tabular Data
• Explainable Machine Learning with Prior Knowledge: An Overview
• Model-Based Counterfactual Synthesizer for Interpretation
• On Counterfactual Explanations under Predictive Multiplicity
• Rational Shapley Values
• How Well do Feature Visualizations Support Causal Understanding of CNN Activations?
• GAM Coach: Towards Interactive and User-centered Algorithmic Recourse
• Consequence-aware Sequential Counterfactual Generation
• Achieving Diversity in Counterfactual Explanations: a Review and Discussion
• Uncertainty Estimation and Out-of-Distribution Detection for Counterfactual Explanations: Pitfalls and Solutions
• Counterfactual Explanations for Arbitrary Regression Models
• Rationalizing Predictions by Adversarial Information Calibration
• Optimal Counterfactual Explanations in Tree Ensembles
• Conditional Generative Models for Counterfactual Explanations
• On Quantitative Evaluations of Counterfactuals
• Framing Algorithmic Recourse for Anomaly Detection
• Counterfactual Explanations for Machine Learning: Challenges Revisited
• Explaining the Behavior of Black-Box Prediction Algorithms with Causal Learning
• Towards User Guided Actionable Recourse
• Towards Feasible Counterfactual Explanations: A Taxonomy Guided Template-based NLG Method
• Algorithmic Recourse in Partially and Fully Confounded Settings Through Bounding Counterfactual Effects
• SafeAR: Safe Algorithmic Recourse by Risk-Aware Policies
• CounterNet: End-to-End Training of Prediction Aware Counterfactual Explanations
• A Series of Unfortunate Counterfactual Events: the Role of Time in Counterfactual Explanations
• Counterfactual Explanations via Latent Space Projection and Interpolation
• DA-DGCEx: Ensuring Validity of Deep Guided Counterfactual Explanations With Distribution-Aware Autoencoder Loss
• Longitudinal Distance: Towards Accountable Instance Attribution