Matches in SemOpenAlex for { <https://semopenalex.org/work/W4313303739> ?p ?o ?g. }
Showing items 1 to 93 of
93
with 100 items per page.
- W4313303739 abstract "Several self-supervised representation learning methods have been proposed for reinforcement learning (RL) with rich observations. For real-world applications of RL, recovering underlying latent states is crucial, particularly when sensory inputs contain irrelevant and exogenous information. In this work, we study how information bottlenecks can be used to construct latent states efficiently in the presence of task-irrelevant information. We propose architectures that utilize variational and discrete information bottlenecks, coined as RepDIB, to learn structured factorized representations. Exploiting the expressiveness bought by factorized representations, we introduce a simple, yet effective, bottleneck that can be integrated with any existing self-supervised objective for RL. We demonstrate this across several online and offline RL benchmarks, along with a real robot arm task, where we find that compressed representations with RepDIB can lead to strong performance improvements, as the learned bottlenecks help predict only the relevant state while ignoring irrelevant information." @default.
- W4313303739 created "2023-01-06" @default.
- W4313303739 creator A5000297059 @default.
- W4313303739 creator A5009971165 @default.
- W4313303739 creator A5016769717 @default.
- W4313303739 creator A5022526821 @default.
- W4313303739 creator A5030891292 @default.
- W4313303739 creator A5031311628 @default.
- W4313303739 creator A5031746269 @default.
- W4313303739 creator A5033636537 @default.
- W4313303739 creator A5062951341 @default.
- W4313303739 creator A5079926596 @default.
- W4313303739 creator A5080311165 @default.
- W4313303739 date "2022-12-28" @default.
- W4313303739 modified "2023-10-18" @default.
- W4313303739 title "Representation Learning in Deep RL via Discrete Information Bottleneck" @default.
- W4313303739 doi "https://doi.org/10.48550/arxiv.2212.13835" @default.
- W4313303739 hasPublicationYear "2022" @default.
- W4313303739 type Work @default.
- W4313303739 citedByCount "0" @default.
- W4313303739 crossrefType "posted-content" @default.
- W4313303739 hasAuthorship W4313303739A5000297059 @default.
- W4313303739 hasAuthorship W4313303739A5009971165 @default.
- W4313303739 hasAuthorship W4313303739A5016769717 @default.
- W4313303739 hasAuthorship W4313303739A5022526821 @default.
- W4313303739 hasAuthorship W4313303739A5030891292 @default.
- W4313303739 hasAuthorship W4313303739A5031311628 @default.
- W4313303739 hasAuthorship W4313303739A5031746269 @default.
- W4313303739 hasAuthorship W4313303739A5033636537 @default.
- W4313303739 hasAuthorship W4313303739A5062951341 @default.
- W4313303739 hasAuthorship W4313303739A5079926596 @default.
- W4313303739 hasAuthorship W4313303739A5080311165 @default.
- W4313303739 hasBestOaLocation W43133037391 @default.
- W4313303739 hasConcept C111472728 @default.
- W4313303739 hasConcept C119857082 @default.
- W4313303739 hasConcept C138885662 @default.
- W4313303739 hasConcept C149635348 @default.
- W4313303739 hasConcept C152139883 @default.
- W4313303739 hasConcept C154945302 @default.
- W4313303739 hasConcept C162324750 @default.
- W4313303739 hasConcept C17744445 @default.
- W4313303739 hasConcept C187736073 @default.
- W4313303739 hasConcept C199360897 @default.
- W4313303739 hasConcept C199539241 @default.
- W4313303739 hasConcept C2776359362 @default.
- W4313303739 hasConcept C2780451532 @default.
- W4313303739 hasConcept C2780513914 @default.
- W4313303739 hasConcept C2780586882 @default.
- W4313303739 hasConcept C2780801425 @default.
- W4313303739 hasConcept C41008148 @default.
- W4313303739 hasConcept C59404180 @default.
- W4313303739 hasConcept C60008888 @default.
- W4313303739 hasConcept C90509273 @default.
- W4313303739 hasConcept C94625758 @default.
- W4313303739 hasConcept C97541855 @default.
- W4313303739 hasConceptScore W4313303739C111472728 @default.
- W4313303739 hasConceptScore W4313303739C119857082 @default.
- W4313303739 hasConceptScore W4313303739C138885662 @default.
- W4313303739 hasConceptScore W4313303739C149635348 @default.
- W4313303739 hasConceptScore W4313303739C152139883 @default.
- W4313303739 hasConceptScore W4313303739C154945302 @default.
- W4313303739 hasConceptScore W4313303739C162324750 @default.
- W4313303739 hasConceptScore W4313303739C17744445 @default.
- W4313303739 hasConceptScore W4313303739C187736073 @default.
- W4313303739 hasConceptScore W4313303739C199360897 @default.
- W4313303739 hasConceptScore W4313303739C199539241 @default.
- W4313303739 hasConceptScore W4313303739C2776359362 @default.
- W4313303739 hasConceptScore W4313303739C2780451532 @default.
- W4313303739 hasConceptScore W4313303739C2780513914 @default.
- W4313303739 hasConceptScore W4313303739C2780586882 @default.
- W4313303739 hasConceptScore W4313303739C2780801425 @default.
- W4313303739 hasConceptScore W4313303739C41008148 @default.
- W4313303739 hasConceptScore W4313303739C59404180 @default.
- W4313303739 hasConceptScore W4313303739C60008888 @default.
- W4313303739 hasConceptScore W4313303739C90509273 @default.
- W4313303739 hasConceptScore W4313303739C94625758 @default.
- W4313303739 hasConceptScore W4313303739C97541855 @default.
- W4313303739 hasLocation W43133037391 @default.
- W4313303739 hasOpenAccess W4313303739 @default.
- W4313303739 hasPrimaryLocation W43133037391 @default.
- W4313303739 hasRelatedWork W175858725 @default.
- W4313303739 hasRelatedWork W2017279084 @default.
- W4313303739 hasRelatedWork W2985997154 @default.
- W4313303739 hasRelatedWork W3022038857 @default.
- W4313303739 hasRelatedWork W3161880184 @default.
- W4313303739 hasRelatedWork W3203309750 @default.
- W4313303739 hasRelatedWork W3206195449 @default.
- W4313303739 hasRelatedWork W4221157405 @default.
- W4313303739 hasRelatedWork W4313303739 @default.
- W4313303739 hasRelatedWork W4319083788 @default.
- W4313303739 isParatext "false" @default.
- W4313303739 isRetracted "false" @default.
- W4313303739 workType "article" @default.