Matches in SemOpenAlex for { <https://semopenalex.org/work/W4366766578> ?p ?o ?g. }
Showing items 1 to 67 of
67
with 100 items per page.
- W4366766578 abstract "While deep reinforcement learning has shown important empirical success, it tends to learn relatively slow due to slow propagation of rewards information and slow update of parametric neural networks. Non-parametric episodic memory, on the other hand, provides a faster learning alternative that does not require representation learning and uses maximum episodic return as state-action values for action selection. Episodic memory and reinforcement learning both have their own strengths and weaknesses. Notably, humans can leverage multiple memory systems concurrently during learning and benefit from all of them. In this work, we propose a method called Two-Memory reinforcement learning agent (2M) that combines episodic memory and reinforcement learning to distill both of their strengths. The 2M agent exploits the speed of the episodic memory part and the optimality and the generalization capacity of the reinforcement learning part to complement each other. Our experiments demonstrate that the 2M agent is more data efficient and outperforms both pure episodic memory and pure reinforcement learning, as well as a state-of-the-art memory-augmented RL agent. Moreover, the proposed approach provides a general framework that can be used to combine any episodic memory agent with other off-policy reinforcement learning algorithms." @default.
- W4366766578 created "2023-04-25" @default.
- W4366766578 creator A5028127027 @default.
- W4366766578 creator A5062774048 @default.
- W4366766578 creator A5065869031 @default.
- W4366766578 creator A5085542421 @default.
- W4366766578 date "2023-04-20" @default.
- W4366766578 modified "2023-10-14" @default.
- W4366766578 title "Two-Memory Reinforcement Learning" @default.
- W4366766578 doi "https://doi.org/10.48550/arxiv.2304.10098" @default.
- W4366766578 hasPublicationYear "2023" @default.
- W4366766578 type Work @default.
- W4366766578 citedByCount "0" @default.
- W4366766578 crossrefType "posted-content" @default.
- W4366766578 hasAuthorship W4366766578A5028127027 @default.
- W4366766578 hasAuthorship W4366766578A5062774048 @default.
- W4366766578 hasAuthorship W4366766578A5065869031 @default.
- W4366766578 hasAuthorship W4366766578A5085542421 @default.
- W4366766578 hasBestOaLocation W43667665781 @default.
- W4366766578 hasConcept C119857082 @default.
- W4366766578 hasConcept C134306372 @default.
- W4366766578 hasConcept C153083717 @default.
- W4366766578 hasConcept C154945302 @default.
- W4366766578 hasConcept C15744967 @default.
- W4366766578 hasConcept C169760540 @default.
- W4366766578 hasConcept C169900460 @default.
- W4366766578 hasConcept C177148314 @default.
- W4366766578 hasConcept C180747234 @default.
- W4366766578 hasConcept C33923547 @default.
- W4366766578 hasConcept C41008148 @default.
- W4366766578 hasConcept C67203356 @default.
- W4366766578 hasConcept C7149132 @default.
- W4366766578 hasConcept C77805123 @default.
- W4366766578 hasConcept C88576662 @default.
- W4366766578 hasConcept C97541855 @default.
- W4366766578 hasConceptScore W4366766578C119857082 @default.
- W4366766578 hasConceptScore W4366766578C134306372 @default.
- W4366766578 hasConceptScore W4366766578C153083717 @default.
- W4366766578 hasConceptScore W4366766578C154945302 @default.
- W4366766578 hasConceptScore W4366766578C15744967 @default.
- W4366766578 hasConceptScore W4366766578C169760540 @default.
- W4366766578 hasConceptScore W4366766578C169900460 @default.
- W4366766578 hasConceptScore W4366766578C177148314 @default.
- W4366766578 hasConceptScore W4366766578C180747234 @default.
- W4366766578 hasConceptScore W4366766578C33923547 @default.
- W4366766578 hasConceptScore W4366766578C41008148 @default.
- W4366766578 hasConceptScore W4366766578C67203356 @default.
- W4366766578 hasConceptScore W4366766578C7149132 @default.
- W4366766578 hasConceptScore W4366766578C77805123 @default.
- W4366766578 hasConceptScore W4366766578C88576662 @default.
- W4366766578 hasConceptScore W4366766578C97541855 @default.
- W4366766578 hasLocation W43667665781 @default.
- W4366766578 hasOpenAccess W4366766578 @default.
- W4366766578 hasPrimaryLocation W43667665781 @default.
- W4366766578 hasRelatedWork W1172213197 @default.
- W4366766578 hasRelatedWork W1964624709 @default.
- W4366766578 hasRelatedWork W2105011545 @default.
- W4366766578 hasRelatedWork W2989932438 @default.
- W4366766578 hasRelatedWork W3022038857 @default.
- W4366766578 hasRelatedWork W3203000071 @default.
- W4366766578 hasRelatedWork W4229023852 @default.
- W4366766578 hasRelatedWork W4318719223 @default.
- W4366766578 hasRelatedWork W4319083788 @default.
- W4366766578 hasRelatedWork W1979038597 @default.
- W4366766578 isParatext "false" @default.
- W4366766578 isRetracted "false" @default.
- W4366766578 workType "article" @default.