Matches in SemOpenAlex for { <https://semopenalex.org/work/W4377161554> ?p ?o ?g. }
Showing items 1 to 83 of
83
with 100 items per page.
- W4377161554 abstract "This paper studies tabular reinforcement learning (RL) in the hybrid setting, which assumes access to both an offline dataset and online interactions with the unknown environment. A central question boils down to how to efficiently utilize online data collection to strengthen and complement the offline dataset and enable effective policy fine-tuning. Leveraging recent advances in reward-agnostic exploration and model-based offline RL, we design a three-stage hybrid RL algorithm that beats the best of both worlds -- pure offline RL and pure online RL -- in terms of sample complexities. The proposed algorithm does not require any reward information during data collection. Our theory is developed based on a new notion called single-policy partial concentrability, which captures the trade-off between distribution mismatch and miscoverage and guides the interplay between offline and online data." @default.
- W4377161554 created "2023-05-21" @default.
- W4377161554 creator A5024124240 @default.
- W4377161554 creator A5053809095 @default.
- W4377161554 creator A5059740024 @default.
- W4377161554 creator A5061140388 @default.
- W4377161554 creator A5072190520 @default.
- W4377161554 date "2023-05-17" @default.
- W4377161554 modified "2023-09-29" @default.
- W4377161554 title "Reward-agnostic Fine-tuning: Provable Statistical Benefits of Hybrid Reinforcement Learning" @default.
- W4377161554 doi "https://doi.org/10.48550/arxiv.2305.10282" @default.
- W4377161554 hasPublicationYear "2023" @default.
- W4377161554 type Work @default.
- W4377161554 citedByCount "0" @default.
- W4377161554 crossrefType "posted-content" @default.
- W4377161554 hasAuthorship W4377161554A5024124240 @default.
- W4377161554 hasAuthorship W4377161554A5053809095 @default.
- W4377161554 hasAuthorship W4377161554A5059740024 @default.
- W4377161554 hasAuthorship W4377161554A5061140388 @default.
- W4377161554 hasAuthorship W4377161554A5072190520 @default.
- W4377161554 hasBestOaLocation W43771615541 @default.
- W4377161554 hasConcept C104317684 @default.
- W4377161554 hasConcept C105795698 @default.
- W4377161554 hasConcept C111919701 @default.
- W4377161554 hasConcept C112313634 @default.
- W4377161554 hasConcept C11413529 @default.
- W4377161554 hasConcept C119857082 @default.
- W4377161554 hasConcept C127716648 @default.
- W4377161554 hasConcept C133462117 @default.
- W4377161554 hasConcept C136764020 @default.
- W4377161554 hasConcept C154945302 @default.
- W4377161554 hasConcept C185592680 @default.
- W4377161554 hasConcept C188082640 @default.
- W4377161554 hasConcept C196921405 @default.
- W4377161554 hasConcept C198531522 @default.
- W4377161554 hasConcept C2778445095 @default.
- W4377161554 hasConcept C2780102126 @default.
- W4377161554 hasConcept C2780490138 @default.
- W4377161554 hasConcept C2986087404 @default.
- W4377161554 hasConcept C33923547 @default.
- W4377161554 hasConcept C41008148 @default.
- W4377161554 hasConcept C43617362 @default.
- W4377161554 hasConcept C55493867 @default.
- W4377161554 hasConcept C97541855 @default.
- W4377161554 hasConceptScore W4377161554C104317684 @default.
- W4377161554 hasConceptScore W4377161554C105795698 @default.
- W4377161554 hasConceptScore W4377161554C111919701 @default.
- W4377161554 hasConceptScore W4377161554C112313634 @default.
- W4377161554 hasConceptScore W4377161554C11413529 @default.
- W4377161554 hasConceptScore W4377161554C119857082 @default.
- W4377161554 hasConceptScore W4377161554C127716648 @default.
- W4377161554 hasConceptScore W4377161554C133462117 @default.
- W4377161554 hasConceptScore W4377161554C136764020 @default.
- W4377161554 hasConceptScore W4377161554C154945302 @default.
- W4377161554 hasConceptScore W4377161554C185592680 @default.
- W4377161554 hasConceptScore W4377161554C188082640 @default.
- W4377161554 hasConceptScore W4377161554C196921405 @default.
- W4377161554 hasConceptScore W4377161554C198531522 @default.
- W4377161554 hasConceptScore W4377161554C2778445095 @default.
- W4377161554 hasConceptScore W4377161554C2780102126 @default.
- W4377161554 hasConceptScore W4377161554C2780490138 @default.
- W4377161554 hasConceptScore W4377161554C2986087404 @default.
- W4377161554 hasConceptScore W4377161554C33923547 @default.
- W4377161554 hasConceptScore W4377161554C41008148 @default.
- W4377161554 hasConceptScore W4377161554C43617362 @default.
- W4377161554 hasConceptScore W4377161554C55493867 @default.
- W4377161554 hasConceptScore W4377161554C97541855 @default.
- W4377161554 hasLocation W43771615541 @default.
- W4377161554 hasOpenAccess W4377161554 @default.
- W4377161554 hasPrimaryLocation W43771615541 @default.
- W4377161554 hasRelatedWork W2548307888 @default.
- W4377161554 hasRelatedWork W4225619808 @default.
- W4377161554 hasRelatedWork W4226283576 @default.
- W4377161554 hasRelatedWork W4306295215 @default.
- W4377161554 hasRelatedWork W4308759076 @default.
- W4377161554 hasRelatedWork W4311991951 @default.
- W4377161554 hasRelatedWork W4312614522 @default.
- W4377161554 hasRelatedWork W4313571701 @default.
- W4377161554 hasRelatedWork W4327524442 @default.
- W4377161554 hasRelatedWork W4361865079 @default.
- W4377161554 isParatext "false" @default.
- W4377161554 isRetracted "false" @default.
- W4377161554 workType "article" @default.