Verified Training for Counterfactual Explanation Robustness under Data Shift

arXiv:2403.03773v1 Announce Type: new
Abstract: Counterfactual explanations (CEs) enhance the interpretability of machine learning models by describing what changes to an input are necessary to change its prediction to a desired class. These explanations are commonly used to guide users' actions, e.g., by describing how a user whose loan application was denied can be approved for a loan in the future. Existing approaches generate CEs by focusing on a single, fixed model, and do not provide any formal guarantees on …

abstract application arxiv ces change class counterfactual cs.lg data guide interpretability machine machine learning machine learning models prediction robustness shift training type