Feature Responsiveness Scores: Model-Agnostic Explanations for Recourse
Keywords: explainability, feature attribution, algorithmic recourse, regulation
TL;DR: We introduce *feature responsiveness scores*, the probability that an individual can change their model prediction by altering a feature.
Abstract: Machine learning models are often used to automate or support decisions in applications such as lending and hiring. In such settings, consumer protection rules mandate that we provide consumers who receive adverse decisions with a list of "principal reasons." In practice, lenders and employers identify principal reasons as the top-scoring features from a *feature attribution* method. In this work, we study how such practices align with one of the underlying goals of consumer protection -- *recourse* -- i.e., educating individuals on how to achieve a desired outcome. We show that standard attribution methods can highlight features that will not lead to recourse -- providing individuals with *reasons without recourse*. We propose to score features on the basis of *responsiveness*, i.e., the proportion of interventions that can lead to a desired outcome. We develop efficient methods to compute responsiveness scores for any model and any dataset under complex actionability constraints. We present an empirical study on the responsiveness of explanations in lending, and demonstrate how responsiveness scores can highlight features that support recourse and mitigate harm by flagging instances with fixed predictions.
Primary Area: interpretability and explainable AI
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.
Reciprocal Reviewing: I understand the reciprocal reviewing requirement as described on https://iclr.cc/Conferences/2025/CallForPapers. If none of the authors are registered as a reviewer, it may result in a desk rejection at the discretion of the program chairs. To request an exception, please complete this form at https://forms.gle/Huojr6VjkFxiQsUp6.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 12358
Loading