Aviv A. Rosenberg

Jul 3, 2024 | arxiv.org | Aviv A. Rosenberg

arXiv:2407.03065 (cs) View PDF HTML (experimental) Subjects: Machine Learning (cs.LG); Machine Learning (stat.ML) Cite as: arXiv:2407.03065 [cs.LG] (or arXiv:2407.03065v1 [cs.LG] for this version) Submission history From: Aviv Rosenberg [ view email] [v1] Wed, 3 Jul 2024 12:36:24 UTC (123 KB) Bibliographic Tools Bibliographic Explorer Toggle Bibliographic Explorer () Litmaps Toggle Litmaps (What is Litmaps?) scite.ai Toggle scite Smart Citations (What are Smart Citations?) Code, Data, Media...

Jun 30, 2023 | amazon.science | Aviv A. Rosenberg |Dmitry Sotnikov |Raffay Hamid |Tal Lancewick

Policy Optimization (PO) is one of the most popular methods in Reinforcement Learning (RL). Thus, theoretical guarantees for PO algorithms have become especially important to the RL community. In this paper, we study PO in adversarial MDPs with a challenge that arises in almost every real-world application – delayed bandit feedback. We give the first near-optimal regret bounds for PO in tabular MDPs, and may even surpass state-of-the-art (which uses less efficient methods).