Tony Cao

This paper introduces a framework to analyze the trustworthiness of near-optimal sparse decision trees from the Rashomon set, showing they can perform as well as models explicitly optimized for fairness, robustness, or privacy.