Oh, D.D. et al. 2026. Provably Optimal Reinforcement Learning Under Safety Filtering. Proceedings of IASEAI Conference. 2, 1 (Jul. 2026), 488–500.