Matches in SemOpenAlex for { <https://semopenalex.org/work/W3207588081> ?p ?o ?g. }
- W3207588081 abstract "Incorporating prior knowledge in reinforcement learning algorithms is mainly an open question. Even when insights about the environment dynamics are available, reinforcement learning is traditionally used in a tabula rasa setting and must explore and learn everything from scratch. In this paper, we consider the problem of exploiting priors about action sequence equivalence: that is, when different sequences of actions produce the same effect. We propose a new local exploration strategy calibrated to minimize collisions and maximize new state visitations. We show that this strategy can be computed at little cost, by solving a convex optimization problem. By replacing the usual epsilon-greedy strategy in a DQN, we demonstrate its potential in several environments with various dynamic structures." @default.
- W3207588081 created "2021-10-25" @default.
- W3207588081 creator A5032092481 @default.
- W3207588081 creator A5071034635 @default.
- W3207588081 creator A5087706654 @default.
- W3207588081 creator A5087891858 @default.
- W3207588081 date "2021-10-20" @default.
- W3207588081 modified "2023-09-27" @default.
- W3207588081 title "More Efficient Exploration with Symbolic Priors on Action Sequence Equivalences" @default.
- W3207588081 cites W1714211023 @default.
- W3207588081 cites W172298727 @default.
- W3207588081 cites W1745373831 @default.
- W3207588081 cites W1988526405 @default.
- W3207588081 cites W2032558547 @default.
- W3207588081 cites W2052469048 @default.
- W3207588081 cites W2084792706 @default.
- W3207588081 cites W2119567691 @default.
- W3207588081 cites W2139612737 @default.
- W3207588081 cites W2140332127 @default.
- W3207588081 cites W2145339207 @default.
- W3207588081 cites W2153222087 @default.
- W3207588081 cites W2160589914 @default.
- W3207588081 cites W2215378786 @default.
- W3207588081 cites W2257979135 @default.
- W3207588081 cites W2626354230 @default.
- W3207588081 cites W2751973545 @default.
- W3207588081 cites W2902298341 @default.
- W3207588081 cites W2926028278 @default.
- W3207588081 cites W2953326529 @default.
- W3207588081 cites W2962832483 @default.
- W3207588081 cites W2963150186 @default.
- W3207588081 cites W2963160877 @default.
- W3207588081 cites W2963276097 @default.
- W3207588081 cites W2963966702 @default.
- W3207588081 cites W2964067469 @default.
- W3207588081 cites W2964158321 @default.
- W3207588081 cites W2964332824 @default.
- W3207588081 cites W2975835314 @default.
- W3207588081 cites W2996695841 @default.
- W3207588081 cites W3005701902 @default.
- W3207588081 cites W3035133122 @default.
- W3207588081 cites W3102762742 @default.
- W3207588081 cites W3107151975 @default.
- W3207588081 cites W3117497722 @default.
- W3207588081 cites W3120050142 @default.
- W3207588081 cites W3123169626 @default.
- W3207588081 cites W3213892249 @default.
- W3207588081 cites W2290452516 @default.
- W3207588081 hasPublicationYear "2021" @default.
- W3207588081 type Work @default.
- W3207588081 sameAs 3207588081 @default.
- W3207588081 citedByCount "0" @default.
- W3207588081 crossrefType "posted-content" @default.
- W3207588081 hasAuthorship W3207588081A5032092481 @default.
- W3207588081 hasAuthorship W3207588081A5071034635 @default.
- W3207588081 hasAuthorship W3207588081A5087706654 @default.
- W3207588081 hasAuthorship W3207588081A5087891858 @default.
- W3207588081 hasConcept C107673813 @default.
- W3207588081 hasConcept C111919701 @default.
- W3207588081 hasConcept C118615104 @default.
- W3207588081 hasConcept C121332964 @default.
- W3207588081 hasConcept C126255220 @default.
- W3207588081 hasConcept C154945302 @default.
- W3207588081 hasConcept C177769412 @default.
- W3207588081 hasConcept C2778112365 @default.
- W3207588081 hasConcept C2780069185 @default.
- W3207588081 hasConcept C2780791683 @default.
- W3207588081 hasConcept C2781235140 @default.
- W3207588081 hasConcept C33923547 @default.
- W3207588081 hasConcept C41008148 @default.
- W3207588081 hasConcept C54355233 @default.
- W3207588081 hasConcept C62520636 @default.
- W3207588081 hasConcept C86803240 @default.
- W3207588081 hasConcept C97541855 @default.
- W3207588081 hasConceptScore W3207588081C107673813 @default.
- W3207588081 hasConceptScore W3207588081C111919701 @default.
- W3207588081 hasConceptScore W3207588081C118615104 @default.
- W3207588081 hasConceptScore W3207588081C121332964 @default.
- W3207588081 hasConceptScore W3207588081C126255220 @default.
- W3207588081 hasConceptScore W3207588081C154945302 @default.
- W3207588081 hasConceptScore W3207588081C177769412 @default.
- W3207588081 hasConceptScore W3207588081C2778112365 @default.
- W3207588081 hasConceptScore W3207588081C2780069185 @default.
- W3207588081 hasConceptScore W3207588081C2780791683 @default.
- W3207588081 hasConceptScore W3207588081C2781235140 @default.
- W3207588081 hasConceptScore W3207588081C33923547 @default.
- W3207588081 hasConceptScore W3207588081C41008148 @default.
- W3207588081 hasConceptScore W3207588081C54355233 @default.
- W3207588081 hasConceptScore W3207588081C62520636 @default.
- W3207588081 hasConceptScore W3207588081C86803240 @default.
- W3207588081 hasConceptScore W3207588081C97541855 @default.
- W3207588081 hasLocation W32075880811 @default.
- W3207588081 hasOpenAccess W3207588081 @default.
- W3207588081 hasPrimaryLocation W32075880811 @default.
- W3207588081 hasRelatedWork W1640358294 @default.
- W3207588081 hasRelatedWork W1882226547 @default.
- W3207588081 hasRelatedWork W2128786740 @default.
- W3207588081 hasRelatedWork W2189395077 @default.
- W3207588081 hasRelatedWork W2215378786 @default.
- W3207588081 hasRelatedWork W2294805292 @default.