Matches in SemOpenAlex for { <https://semopenalex.org/work/W2235081654> ?p ?o ?g. }
- W2235081654 endingPage "3351" @default.
- W2235081654 startingPage "3345" @default.
- W2235081654 abstract "Online multi-task learning is an important capability for lifelong learning agents, enabling them to acquire models for diverse tasks over time and rapidly learn new tasks by building upon prior experience. However, recent progress toward lifelong reinforcement learning (RL) has been limited to learning from within a single task domain. For truly versatile lifelong learning, the agent must be able to autonomously transfer knowledge between different task domains. A few methods for cross-domain transfer have been developed, but these methods are computationally inefficient for scenarios where the agent must learn tasks consecutively.In this paper, we develop the first cross-domain lifelong RL framework. Our approach efficiently optimizes a shared repository of transferable knowledge and learns projection matrices that specialize that knowledge to different task domains. We provide rigorous theoretical guarantees on the stability of this approach, and empirically evaluate its performance on diverse dynamical systems. Our results show that the proposed method can learn effectively from interleaved task domains and rapidly acquire high performance in new domains." @default.
- W2235081654 created "2016-06-24" @default.
- W2235081654 creator A5005962804 @default.
- W2235081654 creator A5020691490 @default.
- W2235081654 creator A5034334762 @default.
- W2235081654 creator A5059480804 @default.
- W2235081654 date "2015-07-25" @default.
- W2235081654 modified "2023-09-26" @default.
- W2235081654 title "Autonomous cross-domain knowledge transfer in lifelong policy gradient reinforcement learning" @default.
- W2235081654 cites W110451278 @default.
- W2235081654 cites W1519626139 @default.
- W2235081654 cites W1560550898 @default.
- W2235081654 cites W1942758450 @default.
- W2235081654 cites W1969074599 @default.
- W2235081654 cites W1996342763 @default.
- W2235081654 cites W2012392077 @default.
- W2235081654 cites W2097381042 @default.
- W2235081654 cites W2098432798 @default.
- W2235081654 cites W2099618002 @default.
- W2235081654 cites W2103285838 @default.
- W2235081654 cites W2106008664 @default.
- W2235081654 cites W2114235770 @default.
- W2235081654 cites W2114537044 @default.
- W2235081654 cites W2119717200 @default.
- W2235081654 cites W2123327324 @default.
- W2235081654 cites W2126635028 @default.
- W2235081654 cites W2134197408 @default.
- W2235081654 cites W2135046866 @default.
- W2235081654 cites W2135316123 @default.
- W2235081654 cites W2153988437 @default.
- W2235081654 cites W2154328025 @default.
- W2235081654 cites W2155027007 @default.
- W2235081654 cites W2158150115 @default.
- W2235081654 cites W2169743339 @default.
- W2235081654 cites W2181867278 @default.
- W2235081654 cites W2186054958 @default.
- W2235081654 cites W2903077804 @default.
- W2235081654 hasPublicationYear "2015" @default.
- W2235081654 type Work @default.
- W2235081654 sameAs 2235081654 @default.
- W2235081654 citedByCount "25" @default.
- W2235081654 countsByYear W22350816542015 @default.
- W2235081654 countsByYear W22350816542016 @default.
- W2235081654 countsByYear W22350816542017 @default.
- W2235081654 countsByYear W22350816542019 @default.
- W2235081654 countsByYear W22350816542020 @default.
- W2235081654 countsByYear W22350816542021 @default.
- W2235081654 crossrefType "proceedings-article" @default.
- W2235081654 hasAuthorship W2235081654A5005962804 @default.
- W2235081654 hasAuthorship W2235081654A5020691490 @default.
- W2235081654 hasAuthorship W2235081654A5034334762 @default.
- W2235081654 hasAuthorship W2235081654A5059480804 @default.
- W2235081654 hasConcept C107457646 @default.
- W2235081654 hasConcept C108771440 @default.
- W2235081654 hasConcept C112972136 @default.
- W2235081654 hasConcept C11413529 @default.
- W2235081654 hasConcept C119857082 @default.
- W2235081654 hasConcept C127413603 @default.
- W2235081654 hasConcept C134306372 @default.
- W2235081654 hasConcept C150899416 @default.
- W2235081654 hasConcept C154945302 @default.
- W2235081654 hasConcept C15744967 @default.
- W2235081654 hasConcept C19417346 @default.
- W2235081654 hasConcept C201995342 @default.
- W2235081654 hasConcept C207685749 @default.
- W2235081654 hasConcept C2776960227 @default.
- W2235081654 hasConcept C2780451532 @default.
- W2235081654 hasConcept C33923547 @default.
- W2235081654 hasConcept C36503486 @default.
- W2235081654 hasConcept C41008148 @default.
- W2235081654 hasConcept C56739046 @default.
- W2235081654 hasConcept C57493831 @default.
- W2235081654 hasConcept C97541855 @default.
- W2235081654 hasConceptScore W2235081654C107457646 @default.
- W2235081654 hasConceptScore W2235081654C108771440 @default.
- W2235081654 hasConceptScore W2235081654C112972136 @default.
- W2235081654 hasConceptScore W2235081654C11413529 @default.
- W2235081654 hasConceptScore W2235081654C119857082 @default.
- W2235081654 hasConceptScore W2235081654C127413603 @default.
- W2235081654 hasConceptScore W2235081654C134306372 @default.
- W2235081654 hasConceptScore W2235081654C150899416 @default.
- W2235081654 hasConceptScore W2235081654C154945302 @default.
- W2235081654 hasConceptScore W2235081654C15744967 @default.
- W2235081654 hasConceptScore W2235081654C19417346 @default.
- W2235081654 hasConceptScore W2235081654C201995342 @default.
- W2235081654 hasConceptScore W2235081654C207685749 @default.
- W2235081654 hasConceptScore W2235081654C2776960227 @default.
- W2235081654 hasConceptScore W2235081654C2780451532 @default.
- W2235081654 hasConceptScore W2235081654C33923547 @default.
- W2235081654 hasConceptScore W2235081654C36503486 @default.
- W2235081654 hasConceptScore W2235081654C41008148 @default.
- W2235081654 hasConceptScore W2235081654C56739046 @default.
- W2235081654 hasConceptScore W2235081654C57493831 @default.
- W2235081654 hasConceptScore W2235081654C97541855 @default.
- W2235081654 hasLocation W22350816541 @default.
- W2235081654 hasOpenAccess W2235081654 @default.
- W2235081654 hasPrimaryLocation W22350816541 @default.
- W2235081654 hasRelatedWork W1519626139 @default.