Matches in SemOpenAlex for { <https://semopenalex.org/work/W2923653485> ?p ?o ?g. }
Showing items 1 to 92 of
92
with 100 items per page.
- W2923653485 endingPage "117" @default.
- W2923653485 startingPage "107" @default.
- W2923653485 abstract "Deep reinforcement learning (DRL) algorithms with experience replays have been used to solve many sequential learning problems. However, in practice, DRL algorithms still suffer from the data inefficiency problem, which limits their applicability in many scenarios, and renders them inefficient in solving real-world problems. To improve the data efficiency of DRL, in this paper, a new multi-step method is proposed. Unlike traditional algorithms, the proposed method uses a new return function, which alters the discount of future rewards while decreasing the impact of the immediate reward when selecting the current state action. This approach has the potential to improve the efficiency of reward data. By combining the proposed method with classic DRL algorithms, deep Q-networks (DQN) and double deep Q-networks (DDQN), two novel algorithms are proposed for improving the efficiency of learning from experience replay. The performance of the proposed algorithms, expected n-step DQN (EnDQN) and expected n-step DDQN (EnDDQN), are validated using two simulation environments, CartPole and DeepTraffic. The experimental results demonstrate that the proposed multi-step methods greatly improve the data efficiency of DRL agents while further improving the performance of existing classic DRL algorithms when incorporated into their training." @default.
- W2923653485 created "2019-04-01" @default.
- W2923653485 creator A5004874964 @default.
- W2923653485 creator A5015297129 @default.
- W2923653485 creator A5035633072 @default.
- W2923653485 creator A5048899636 @default.
- W2923653485 creator A5059725501 @default.
- W2923653485 creator A5070234214 @default.
- W2923653485 creator A5077520782 @default.
- W2923653485 date "2019-07-01" @default.
- W2923653485 modified "2023-10-15" @default.
- W2923653485 title "A novel multi-step Q-learning method to improve data efficiency for deep reinforcement learning" @default.
- W2923653485 cites W1972063518 @default.
- W2923653485 cites W2017207054 @default.
- W2923653485 cites W2018705428 @default.
- W2923653485 cites W2076063813 @default.
- W2923653485 cites W2107726111 @default.
- W2923653485 cites W2132411323 @default.
- W2923653485 cites W2139418546 @default.
- W2923653485 cites W2145339207 @default.
- W2923653485 cites W2257979135 @default.
- W2923653485 cites W2460299708 @default.
- W2923653485 cites W2462003052 @default.
- W2923653485 cites W2610686804 @default.
- W2923653485 cites W2767753530 @default.
- W2923653485 cites W2774719390 @default.
- W2923653485 cites W2919115771 @default.
- W2923653485 cites W2962834855 @default.
- W2923653485 cites W3041202696 @default.
- W2923653485 cites W3100789280 @default.
- W2923653485 cites W2301408779 @default.
- W2923653485 doi "https://doi.org/10.1016/j.knosys.2019.03.018" @default.
- W2923653485 hasPublicationYear "2019" @default.
- W2923653485 type Work @default.
- W2923653485 sameAs 2923653485 @default.
- W2923653485 citedByCount "32" @default.
- W2923653485 countsByYear W29236534852019 @default.
- W2923653485 countsByYear W29236534852020 @default.
- W2923653485 countsByYear W29236534852021 @default.
- W2923653485 countsByYear W29236534852022 @default.
- W2923653485 countsByYear W29236534852023 @default.
- W2923653485 crossrefType "journal-article" @default.
- W2923653485 hasAuthorship W2923653485A5004874964 @default.
- W2923653485 hasAuthorship W2923653485A5015297129 @default.
- W2923653485 hasAuthorship W2923653485A5035633072 @default.
- W2923653485 hasAuthorship W2923653485A5048899636 @default.
- W2923653485 hasAuthorship W2923653485A5059725501 @default.
- W2923653485 hasAuthorship W2923653485A5070234214 @default.
- W2923653485 hasAuthorship W2923653485A5077520782 @default.
- W2923653485 hasConcept C108583219 @default.
- W2923653485 hasConcept C119857082 @default.
- W2923653485 hasConcept C14036430 @default.
- W2923653485 hasConcept C154945302 @default.
- W2923653485 hasConcept C162324750 @default.
- W2923653485 hasConcept C175444787 @default.
- W2923653485 hasConcept C2778869765 @default.
- W2923653485 hasConcept C41008148 @default.
- W2923653485 hasConcept C78458016 @default.
- W2923653485 hasConcept C86803240 @default.
- W2923653485 hasConcept C97541855 @default.
- W2923653485 hasConceptScore W2923653485C108583219 @default.
- W2923653485 hasConceptScore W2923653485C119857082 @default.
- W2923653485 hasConceptScore W2923653485C14036430 @default.
- W2923653485 hasConceptScore W2923653485C154945302 @default.
- W2923653485 hasConceptScore W2923653485C162324750 @default.
- W2923653485 hasConceptScore W2923653485C175444787 @default.
- W2923653485 hasConceptScore W2923653485C2778869765 @default.
- W2923653485 hasConceptScore W2923653485C41008148 @default.
- W2923653485 hasConceptScore W2923653485C78458016 @default.
- W2923653485 hasConceptScore W2923653485C86803240 @default.
- W2923653485 hasConceptScore W2923653485C97541855 @default.
- W2923653485 hasFunder F4320321001 @default.
- W2923653485 hasLocation W29236534851 @default.
- W2923653485 hasOpenAccess W2923653485 @default.
- W2923653485 hasPrimaryLocation W29236534851 @default.
- W2923653485 hasRelatedWork W2795261237 @default.
- W2923653485 hasRelatedWork W3014300295 @default.
- W2923653485 hasRelatedWork W3164822677 @default.
- W2923653485 hasRelatedWork W4223943233 @default.
- W2923653485 hasRelatedWork W4225161397 @default.
- W2923653485 hasRelatedWork W4312200629 @default.
- W2923653485 hasRelatedWork W4360585206 @default.
- W2923653485 hasRelatedWork W4364306694 @default.
- W2923653485 hasRelatedWork W4380075502 @default.
- W2923653485 hasRelatedWork W4380086463 @default.
- W2923653485 hasVolume "175" @default.
- W2923653485 isParatext "false" @default.
- W2923653485 isRetracted "false" @default.
- W2923653485 magId "2923653485" @default.
- W2923653485 workType "article" @default.