Offline Reinforcement Learning in Large State Spaces: Algorithms and Guarantees
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866908577202962432 |
|---|---|
| author | Jiang, Nan Xie, Tengyang |
| author_facet | Jiang, Nan Xie, Tengyang |
| contents | This article introduces the theory of offline reinforcement learning in large state spaces, where good policies are learned from historical data without online interactions with the environment. Key concepts introduced include expressivity assumptions on function approximation (e.g., Bellman completeness vs. realizability) and data coverage (e.g., all-policy vs. single-policy coverage). A rich landscape of algorithms and results is described, depending on the assumptions one is willing to make and the sample and computational complexity guarantees one wishes to achieve. We also discuss open questions and connections to adjacent areas. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_04088 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Offline Reinforcement Learning in Large State Spaces: Algorithms and Guarantees Jiang, Nan Xie, Tengyang Machine Learning Artificial Intelligence This article introduces the theory of offline reinforcement learning in large state spaces, where good policies are learned from historical data without online interactions with the environment. Key concepts introduced include expressivity assumptions on function approximation (e.g., Bellman completeness vs. realizability) and data coverage (e.g., all-policy vs. single-policy coverage). A rich landscape of algorithms and results is described, depending on the assumptions one is willing to make and the sample and computational complexity guarantees one wishes to achieve. We also discuss open questions and connections to adjacent areas. |
| title | Offline Reinforcement Learning in Large State Spaces: Algorithms and Guarantees |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2510.04088 |