Assistant Professor, University of Southern California
1 paper at NeurIPS 2025
We propose priori knowledge-free algorithms for distributionally robust average-reward estimation, with both finite-sample theoretical guarantees and numerical validation.
