Matches in SemOpenAlex for { <https://semopenalex.org/work/W3200762456> ?p ?o ?g. }
Showing items 1 to 67 of
67
with 100 items per page.
- W3200762456 endingPage "012061" @default.
- W3200762456 startingPage "012061" @default.
- W3200762456 abstract "The development of artificial intelligence is becoming more intelligent, and Reinforcement learning (RL) plays an important role. RL has been widely used in recommendation systems, smart cars, game AI, and investment transactions. Since the emergence of DQN, DRL improvement methods have been proposed continuously. Soft Actor-Critic (SAC) is the current state-of-the-art RL algorithm. In this paper, some tricks are used to improve SAC. Firstly, the baseline is added to the policy gradient algorithm to reduce variance and speed up training. This paper adopts a new method to implement it in SAC. Because the temperature coefficient is a super parameter, which is difficult to adjust, so I propose a novel method to set it. These tricks proved to be effective on the MUJOCO." @default.
- W3200762456 created "2021-09-27" @default.
- W3200762456 creator A5012877244 @default.
- W3200762456 creator A5017701414 @default.
- W3200762456 creator A5027770821 @default.
- W3200762456 creator A5054482111 @default.
- W3200762456 creator A5058117645 @default.
- W3200762456 date "2021-09-01" @default.
- W3200762456 modified "2023-09-25" @default.
- W3200762456 title "Some effective tricks are used to improve Soft Actor Critic" @default.
- W3200762456 doi "https://doi.org/10.1088/1742-6596/2010/1/012061" @default.
- W3200762456 hasPublicationYear "2021" @default.
- W3200762456 type Work @default.
- W3200762456 sameAs 3200762456 @default.
- W3200762456 citedByCount "0" @default.
- W3200762456 crossrefType "journal-article" @default.
- W3200762456 hasAuthorship W3200762456A5012877244 @default.
- W3200762456 hasAuthorship W3200762456A5017701414 @default.
- W3200762456 hasAuthorship W3200762456A5027770821 @default.
- W3200762456 hasAuthorship W3200762456A5054482111 @default.
- W3200762456 hasAuthorship W3200762456A5058117645 @default.
- W3200762456 hasBestOaLocation W32007624561 @default.
- W3200762456 hasConcept C11413529 @default.
- W3200762456 hasConcept C119857082 @default.
- W3200762456 hasConcept C121955636 @default.
- W3200762456 hasConcept C144133560 @default.
- W3200762456 hasConcept C154945302 @default.
- W3200762456 hasConcept C177264268 @default.
- W3200762456 hasConcept C196083921 @default.
- W3200762456 hasConcept C196340769 @default.
- W3200762456 hasConcept C199360897 @default.
- W3200762456 hasConcept C41008148 @default.
- W3200762456 hasConcept C48103436 @default.
- W3200762456 hasConcept C97541855 @default.
- W3200762456 hasConceptScore W3200762456C11413529 @default.
- W3200762456 hasConceptScore W3200762456C119857082 @default.
- W3200762456 hasConceptScore W3200762456C121955636 @default.
- W3200762456 hasConceptScore W3200762456C144133560 @default.
- W3200762456 hasConceptScore W3200762456C154945302 @default.
- W3200762456 hasConceptScore W3200762456C177264268 @default.
- W3200762456 hasConceptScore W3200762456C196083921 @default.
- W3200762456 hasConceptScore W3200762456C196340769 @default.
- W3200762456 hasConceptScore W3200762456C199360897 @default.
- W3200762456 hasConceptScore W3200762456C41008148 @default.
- W3200762456 hasConceptScore W3200762456C48103436 @default.
- W3200762456 hasConceptScore W3200762456C97541855 @default.
- W3200762456 hasIssue "1" @default.
- W3200762456 hasLocation W32007624561 @default.
- W3200762456 hasOpenAccess W3200762456 @default.
- W3200762456 hasPrimaryLocation W32007624561 @default.
- W3200762456 hasRelatedWork W2759757439 @default.
- W3200762456 hasRelatedWork W2923653485 @default.
- W3200762456 hasRelatedWork W2957776456 @default.
- W3200762456 hasRelatedWork W3022038857 @default.
- W3200762456 hasRelatedWork W3025133396 @default.
- W3200762456 hasRelatedWork W3088315509 @default.
- W3200762456 hasRelatedWork W4285324524 @default.
- W3200762456 hasRelatedWork W4319083788 @default.
- W3200762456 hasRelatedWork W4321146100 @default.
- W3200762456 hasRelatedWork W4300772058 @default.
- W3200762456 hasVolume "2010" @default.
- W3200762456 isParatext "false" @default.
- W3200762456 isRetracted "false" @default.
- W3200762456 magId "3200762456" @default.
- W3200762456 workType "article" @default.