RecourseBench: Reproducible Counterfactual Explanation Evaluation

Published in Preprint, under review, 2026

Co-developed a scalable Python benchmarking platform integrating 22 heterogeneous ML methods across 137 method, dataset, and model configurations. Designed standardized interfaces for consistent evaluation of heterogeneous methods across ML frameworks, and built automated CI and reproducibility tests to validate execution reliability, output consistency, and fidelity to reported results.

Code: github.com/zkhotanlou/recourse_benchmarks

Recommended citation: Zahra Khotanlou, Hashir Ahmed, Chenghao Tan, Ahmed Abdelaal, and Amir-Hossein Karimi. "RecourseBench: Reproducible Counterfactual Explanation Evaluation." Preprint, under review.
Download Paper