ElliCE: Efficient and Provably Robust Algorithmic Recourse via the Rashomon Sets

TL;DR

ElliCE is a new framework for generating robust, reliable, and flexible algorithmic recourse explanations that are valid across a set of near-optimal models, with theoretical guarantees and improved efficiency.

Contribution

It introduces ElliCE, which optimizes counterfactuals over an ellipsoidal approximation of the Rashomon set, providing provably valid, stable, and adaptable recourse explanations.

Findings

ElliCE produces more robust counterfactuals than existing methods.

The explanations are faster to compute and adaptable to user constraints.

ElliCE offers theoretical guarantees on explanation stability and validity.

Abstract

Machine learning models now influence decisions that directly affect people's lives, making it important to understand not only their predictions, but also how individuals could act to obtain better results. Algorithmic recourse provides actionable input modifications to achieve more favorable outcomes, typically relying on counterfactual explanations to suggest such changes. However, when the Rashomon set - the set of near-optimal models - is large, standard counterfactual explanations can become unreliable, as a recourse action valid for one model may fail under another. We introduce ElliCE, a novel framework for robust algorithmic recourse that optimizes counterfactuals over an ellipsoidal approximation of the Rashomon set. The resulting explanations are provably valid over this ellipsoid, with theoretical guarantees on uniqueness, stability, and alignment with key feature directions. Empirically, ElliCE generates counterfactuals that are not only more robust but also more flexible, adapting to user-specified feature constraints while being substantially faster than existing baselines. This provides a principled and practical solution for reliable recourse under model uncertainty, ensuring stable recommendations for users even as models evolve.