Matches in SemOpenAlex for { <https://semopenalex.org/work/W4200456370> ?p ?o ?g. }
- W4200456370 endingPage "10235" @default.
- W4200456370 startingPage "10222" @default.
- W4200456370 abstract "Deep reinforcement learning (RL) agents are becoming increasingly proficient in a range of complex control tasks. However, the agent's behavior is usually difficult to interpret due to the introduction of black-box function, making it difficult to acquire the trust of users. Although there have been some interesting interpretation methods for vision-based RL, most of them cannot uncover temporal causal information, raising questions about their reliability. To address this problem, we present a temporal-spatial causal interpretation (TSCI) model to understand the agent's long-term behavior, which is essential for sequential decision-making. TSCI model builds on the formulation of temporal causality, which reflects the temporal causal relations between sequential observations and decisions of RL agent. Then a separate causal discovery network is employed to identify temporal-spatial causal features, which are constrained to satisfy the temporal causality. TSCI model is applicable to recurrent agents and can be used to discover causal features with high efficiency once trained. The empirical results show that TSCI model can produce high-resolution and sharp attention masks to highlight task-relevant temporal-spatial information that constitutes most evidence about how vision-based RL agents make sequential decisions. In addition, we further demonstrate that our method is able to provide valuable causal interpretations for vision-based RL agents from the temporal perspective." @default.
- W4200456370 created "2021-12-31" @default.
- W4200456370 creator A5013240918 @default.
- W4200456370 creator A5055957441 @default.
- W4200456370 creator A5068624597 @default.
- W4200456370 creator A5074980160 @default.
- W4200456370 date "2022-12-01" @default.
- W4200456370 modified "2023-10-10" @default.
- W4200456370 title "Temporal-Spatial Causal Interpretations for Vision-Based Reinforcement Learning" @default.
- W4200456370 cites W2051228319 @default.
- W4200456370 cites W2065515536 @default.
- W4200456370 cites W2077177391 @default.
- W4200456370 cites W2101786389 @default.
- W4200456370 cites W2118418963 @default.
- W4200456370 cites W2145339207 @default.
- W4200456370 cites W2156803951 @default.
- W4200456370 cites W2158782408 @default.
- W4200456370 cites W2178225550 @default.
- W4200456370 cites W2282821441 @default.
- W4200456370 cites W2336525064 @default.
- W4200456370 cites W2487898712 @default.
- W4200456370 cites W2776207810 @default.
- W4200456370 cites W2891830784 @default.
- W4200456370 cites W2919115771 @default.
- W4200456370 cites W2951510753 @default.
- W4200456370 cites W2962702317 @default.
- W4200456370 cites W2962711930 @default.
- W4200456370 cites W2962843949 @default.
- W4200456370 cites W2962858109 @default.
- W4200456370 cites W2963081790 @default.
- W4200456370 cites W2963371637 @default.
- W4200456370 cites W2963749936 @default.
- W4200456370 cites W2964199361 @default.
- W4200456370 cites W2976187852 @default.
- W4200456370 cites W2980091417 @default.
- W4200456370 cites W2990810232 @default.
- W4200456370 cites W2997423597 @default.
- W4200456370 cites W2998004401 @default.
- W4200456370 cites W2999458807 @default.
- W4200456370 cites W3004058922 @default.
- W4200456370 cites W3012550998 @default.
- W4200456370 cites W3086764798 @default.
- W4200456370 cites W3098703396 @default.
- W4200456370 cites W3100388886 @default.
- W4200456370 cites W3101609372 @default.
- W4200456370 cites W3103780890 @default.
- W4200456370 doi "https://doi.org/10.1109/tpami.2021.3133717" @default.
- W4200456370 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/34882545" @default.
- W4200456370 hasPublicationYear "2022" @default.
- W4200456370 type Work @default.
- W4200456370 citedByCount "3" @default.
- W4200456370 countsByYear W42004563702022 @default.
- W4200456370 countsByYear W42004563702023 @default.
- W4200456370 crossrefType "journal-article" @default.
- W4200456370 hasAuthorship W4200456370A5013240918 @default.
- W4200456370 hasAuthorship W4200456370A5055957441 @default.
- W4200456370 hasAuthorship W4200456370A5068624597 @default.
- W4200456370 hasAuthorship W4200456370A5074980160 @default.
- W4200456370 hasBestOaLocation W42004563702 @default.
- W4200456370 hasConcept C11671645 @default.
- W4200456370 hasConcept C119857082 @default.
- W4200456370 hasConcept C121332964 @default.
- W4200456370 hasConcept C12713177 @default.
- W4200456370 hasConcept C142724271 @default.
- W4200456370 hasConcept C154945302 @default.
- W4200456370 hasConcept C163504300 @default.
- W4200456370 hasConcept C199360897 @default.
- W4200456370 hasConcept C41008148 @default.
- W4200456370 hasConcept C527412718 @default.
- W4200456370 hasConcept C62520636 @default.
- W4200456370 hasConcept C64357122 @default.
- W4200456370 hasConcept C71924100 @default.
- W4200456370 hasConcept C94966114 @default.
- W4200456370 hasConcept C97541855 @default.
- W4200456370 hasConceptScore W4200456370C11671645 @default.
- W4200456370 hasConceptScore W4200456370C119857082 @default.
- W4200456370 hasConceptScore W4200456370C121332964 @default.
- W4200456370 hasConceptScore W4200456370C12713177 @default.
- W4200456370 hasConceptScore W4200456370C142724271 @default.
- W4200456370 hasConceptScore W4200456370C154945302 @default.
- W4200456370 hasConceptScore W4200456370C163504300 @default.
- W4200456370 hasConceptScore W4200456370C199360897 @default.
- W4200456370 hasConceptScore W4200456370C41008148 @default.
- W4200456370 hasConceptScore W4200456370C527412718 @default.
- W4200456370 hasConceptScore W4200456370C62520636 @default.
- W4200456370 hasConceptScore W4200456370C64357122 @default.
- W4200456370 hasConceptScore W4200456370C71924100 @default.
- W4200456370 hasConceptScore W4200456370C94966114 @default.
- W4200456370 hasConceptScore W4200456370C97541855 @default.
- W4200456370 hasFunder F4320321001 @default.
- W4200456370 hasFunder F4320329860 @default.
- W4200456370 hasIssue "12" @default.
- W4200456370 hasLocation W42004563701 @default.
- W4200456370 hasLocation W42004563702 @default.
- W4200456370 hasLocation W42004563703 @default.
- W4200456370 hasOpenAccess W4200456370 @default.
- W4200456370 hasPrimaryLocation W42004563701 @default.
- W4200456370 hasRelatedWork W2018580387 @default.