Kartikeya Chitranshi

Training on Plausible Counterfactuals Removes Spurious Correlations

2025 · arXiv (Cornell University)

Plausible counterfactual explanations (p-CFEs) are perturbations that minimally modify inputs to change classifier decisions while remaining plausible under the data distribution. In this study, we demonstrate that classifiers can be trained on p-CFEs labeled with …