Matches in SemOpenAlex for { <https://semopenalex.org/work/W4387185318> ?p ?o ?g. }
Showing items 1 to 65 of
65
with 100 items per page.
- W4387185318 abstract "Contrastive learning has been used to learn useful low-dimensional state representations in visual reinforcement learning (RL). Such state representations substantially improve the sample efficiency of visual RL. Nevertheless, existing contrastive learning-based RL methods have the problem of unstable training. Such instability comes from the fact that contrastive learning requires an extremely large batch size (e.g., 4096 or larger), while current contrastive learning-based RL methods typically set a small batch size (e.g., 512). In this paper, we propose an approach of discrete information bottleneck (DIB) to address this problem. DIB applies the technique of discretization and information bottleneck to contrastive learning in representing the state with concise discrete representation. Using this discrete representation for policy learning results in more stable algorithm training and higher sample efficiency with a small batch size. We demonstrate the advantage of discrete state representation of DIB on several continuous control tasks in the DeepMind Control suite. In the experiments, DIB outperforms prior visual RL methods, both model-based and model-free, in terms of performance and sample efficiency." @default.
- W4387185318 created "2023-09-30" @default.
- W4387185318 creator A5007930914 @default.
- W4387185318 creator A5038214042 @default.
- W4387185318 date "2023-09-28" @default.
- W4387185318 modified "2023-09-30" @default.
- W4387185318 title "Improving Visual Reinforcement Learning with Discrete Information Bottleneck Approach" @default.
- W4387185318 doi "https://doi.org/10.3233/faia230548" @default.
- W4387185318 hasPublicationYear "2023" @default.
- W4387185318 type Work @default.
- W4387185318 citedByCount "0" @default.
- W4387185318 crossrefType "book-chapter" @default.
- W4387185318 hasAuthorship W4387185318A5007930914 @default.
- W4387185318 hasAuthorship W4387185318A5038214042 @default.
- W4387185318 hasBestOaLocation W43871853181 @default.
- W4387185318 hasConcept C119857082 @default.
- W4387185318 hasConcept C134306372 @default.
- W4387185318 hasConcept C149635348 @default.
- W4387185318 hasConcept C152139883 @default.
- W4387185318 hasConcept C154945302 @default.
- W4387185318 hasConcept C177264268 @default.
- W4387185318 hasConcept C17744445 @default.
- W4387185318 hasConcept C199360897 @default.
- W4387185318 hasConcept C199539241 @default.
- W4387185318 hasConcept C2776359362 @default.
- W4387185318 hasConcept C2780513914 @default.
- W4387185318 hasConcept C33923547 @default.
- W4387185318 hasConcept C41008148 @default.
- W4387185318 hasConcept C60008888 @default.
- W4387185318 hasConcept C73000952 @default.
- W4387185318 hasConcept C94625758 @default.
- W4387185318 hasConcept C97541855 @default.
- W4387185318 hasConceptScore W4387185318C119857082 @default.
- W4387185318 hasConceptScore W4387185318C134306372 @default.
- W4387185318 hasConceptScore W4387185318C149635348 @default.
- W4387185318 hasConceptScore W4387185318C152139883 @default.
- W4387185318 hasConceptScore W4387185318C154945302 @default.
- W4387185318 hasConceptScore W4387185318C177264268 @default.
- W4387185318 hasConceptScore W4387185318C17744445 @default.
- W4387185318 hasConceptScore W4387185318C199360897 @default.
- W4387185318 hasConceptScore W4387185318C199539241 @default.
- W4387185318 hasConceptScore W4387185318C2776359362 @default.
- W4387185318 hasConceptScore W4387185318C2780513914 @default.
- W4387185318 hasConceptScore W4387185318C33923547 @default.
- W4387185318 hasConceptScore W4387185318C41008148 @default.
- W4387185318 hasConceptScore W4387185318C60008888 @default.
- W4387185318 hasConceptScore W4387185318C73000952 @default.
- W4387185318 hasConceptScore W4387185318C94625758 @default.
- W4387185318 hasConceptScore W4387185318C97541855 @default.
- W4387185318 hasLocation W43871853181 @default.
- W4387185318 hasOpenAccess W4387185318 @default.
- W4387185318 hasPrimaryLocation W43871853181 @default.
- W4387185318 hasRelatedWork W1504394672 @default.
- W4387185318 hasRelatedWork W2070945723 @default.
- W4387185318 hasRelatedWork W2381356463 @default.
- W4387185318 hasRelatedWork W2622284819 @default.
- W4387185318 hasRelatedWork W2950826591 @default.
- W4387185318 hasRelatedWork W2996506326 @default.
- W4387185318 hasRelatedWork W4300774107 @default.
- W4387185318 hasRelatedWork W4313898607 @default.
- W4387185318 hasRelatedWork W4319083788 @default.
- W4387185318 hasRelatedWork W4386721405 @default.
- W4387185318 isParatext "false" @default.
- W4387185318 isRetracted "false" @default.
- W4387185318 workType "book-chapter" @default.