Matches in SemOpenAlex for { <https://semopenalex.org/work/W2890105054> ?p ?o ?g. }
- W2890105054 abstract "This paper proposes an exploration method for deep reinforcement learning based on parameter space noise. Recent studies have experimentally shown that parameter space noise results in better exploration than the commonly used action space noise. Previous methods devised a way to update the diagonal covariance matrix of a noise distribution and did not consider the direction of the noise vector and its correlation. In addition, fast updates of the noise distribution are required to facilitate policy learning. We propose a method that deforms the noise distribution according to the accumulated returns and the noises that have led to the returns. Moreover, this method switches isotropic exploration and directional exploration in parameter space with regard to obtained rewards. We validate our exploration strategy in the OpenAI Gym continuous environments and modified environments with sparse rewards. The proposed method achieves results that are competitive with a previous method at baseline tasks. Moreover, our approach exhibits better performance in sparse reward environments by exploration with the switching strategy." @default.
- W2890105054 created "2018-09-27" @default.
- W2890105054 creator A5010543059 @default.
- W2890105054 creator A5045607922 @default.
- W2890105054 creator A5056875214 @default.
- W2890105054 creator A5059653874 @default.
- W2890105054 date "2018-09-18" @default.
- W2890105054 modified "2023-10-16" @default.
- W2890105054 title "Switching Isotropic and Directional Exploration with Parameter Space Noise in Deep Reinforcement Learning." @default.
- W2890105054 cites W2085627234 @default.
- W2890105054 cites W2101677491 @default.
- W2890105054 cites W2114537044 @default.
- W2890105054 cites W2119717200 @default.
- W2890105054 cites W2123327324 @default.
- W2890105054 cites W2145339207 @default.
- W2890105054 cites W2158782408 @default.
- W2890105054 cites W2165150801 @default.
- W2890105054 cites W2168405694 @default.
- W2890105054 cites W2173248099 @default.
- W2890105054 cites W2342662072 @default.
- W2890105054 cites W2417786368 @default.
- W2890105054 cites W2419612459 @default.
- W2890105054 cites W2596367596 @default.
- W2890105054 cites W2603088459 @default.
- W2890105054 cites W2614839826 @default.
- W2890105054 cites W2623491082 @default.
- W2890105054 cites W2724169821 @default.
- W2890105054 cites W2776365419 @default.
- W2890105054 cites W2778749116 @default.
- W2890105054 cites W2779171599 @default.
- W2890105054 cites W2785542505 @default.
- W2890105054 cites W2949608212 @default.
- W2890105054 cites W2951595529 @default.
- W2890105054 cites W2964043796 @default.
- W2890105054 cites W2964121744 @default.
- W2890105054 hasPublicationYear "2018" @default.
- W2890105054 type Work @default.
- W2890105054 sameAs 2890105054 @default.
- W2890105054 citedByCount "0" @default.
- W2890105054 crossrefType "posted-content" @default.
- W2890105054 hasAuthorship W2890105054A5010543059 @default.
- W2890105054 hasAuthorship W2890105054A5045607922 @default.
- W2890105054 hasAuthorship W2890105054A5056875214 @default.
- W2890105054 hasAuthorship W2890105054A5059653874 @default.
- W2890105054 hasConcept C105795698 @default.
- W2890105054 hasConcept C111919701 @default.
- W2890105054 hasConcept C11413529 @default.
- W2890105054 hasConcept C115961682 @default.
- W2890105054 hasConcept C121332964 @default.
- W2890105054 hasConcept C130367717 @default.
- W2890105054 hasConcept C154945302 @default.
- W2890105054 hasConcept C178650346 @default.
- W2890105054 hasConcept C184050105 @default.
- W2890105054 hasConcept C185142706 @default.
- W2890105054 hasConcept C2524010 @default.
- W2890105054 hasConcept C2778572836 @default.
- W2890105054 hasConcept C33923547 @default.
- W2890105054 hasConcept C41008148 @default.
- W2890105054 hasConcept C62520636 @default.
- W2890105054 hasConcept C97541855 @default.
- W2890105054 hasConcept C99498987 @default.
- W2890105054 hasConceptScore W2890105054C105795698 @default.
- W2890105054 hasConceptScore W2890105054C111919701 @default.
- W2890105054 hasConceptScore W2890105054C11413529 @default.
- W2890105054 hasConceptScore W2890105054C115961682 @default.
- W2890105054 hasConceptScore W2890105054C121332964 @default.
- W2890105054 hasConceptScore W2890105054C130367717 @default.
- W2890105054 hasConceptScore W2890105054C154945302 @default.
- W2890105054 hasConceptScore W2890105054C178650346 @default.
- W2890105054 hasConceptScore W2890105054C184050105 @default.
- W2890105054 hasConceptScore W2890105054C185142706 @default.
- W2890105054 hasConceptScore W2890105054C2524010 @default.
- W2890105054 hasConceptScore W2890105054C2778572836 @default.
- W2890105054 hasConceptScore W2890105054C33923547 @default.
- W2890105054 hasConceptScore W2890105054C41008148 @default.
- W2890105054 hasConceptScore W2890105054C62520636 @default.
- W2890105054 hasConceptScore W2890105054C97541855 @default.
- W2890105054 hasConceptScore W2890105054C99498987 @default.
- W2890105054 hasLocation W28901050541 @default.
- W2890105054 hasOpenAccess W2890105054 @default.
- W2890105054 hasPrimaryLocation W28901050541 @default.
- W2890105054 hasRelatedWork W117079632 @default.
- W2890105054 hasRelatedWork W1988465479 @default.
- W2890105054 hasRelatedWork W2034125839 @default.
- W2890105054 hasRelatedWork W2073486209 @default.
- W2890105054 hasRelatedWork W2486578931 @default.
- W2890105054 hasRelatedWork W2898319173 @default.
- W2890105054 hasRelatedWork W2911333147 @default.
- W2890105054 hasRelatedWork W2980722631 @default.
- W2890105054 hasRelatedWork W3008952013 @default.
- W2890105054 hasRelatedWork W3034327083 @default.
- W2890105054 hasRelatedWork W3039564194 @default.
- W2890105054 hasRelatedWork W3040126051 @default.
- W2890105054 hasRelatedWork W3041976821 @default.
- W2890105054 hasRelatedWork W3080082755 @default.
- W2890105054 hasRelatedWork W3095011812 @default.
- W2890105054 hasRelatedWork W3104100176 @default.
- W2890105054 hasRelatedWork W3127378822 @default.
- W2890105054 hasRelatedWork W3196059551 @default.
- W2890105054 hasRelatedWork W3198105516 @default.