Matches in SemOpenAlex for { <https://semopenalex.org/work/W2895921264> ?p ?o ?g. }
Showing items 1 to 87 of
87
with 100 items per page.
- W2895921264 abstract "We propose a unified mechanism for achieving coordination and communication in Multi-Agent Reinforcement Learning (MARL), through rewarding agents for having causal influence over other agents' actions. Causal influence is assessed using counterfactual reasoning. At each timestep, an agent simulates alternate actions that it could have taken, and computes their effect on the behavior of other agents. Actions that lead to bigger changes in other agents' behavior are considered influential and are rewarded. We show that this is equivalent to rewarding agents for having high mutual information between their actions. Empirical results demonstrate that influence leads to enhanced coordination and communication in challenging social dilemma environments, dramatically increasing the learning curves of the deep RL agents, and leading to more meaningful learned communication protocols. The influence rewards for all agents can be computed in a decentralized way by enabling agents to learn a model of other agents using deep neural networks. In contrast, key previous works on emergent communication in the MARL setting were unable to learn diverse policies in a decentralized manner and had to resort to centralized training. Consequently, the influence reward opens up a window of new opportunities for research in this area." @default.
- W2895921264 created "2018-10-26" @default.
- W2895921264 creator A5006947993 @default.
- W2895921264 creator A5035060247 @default.
- W2895921264 creator A5041145688 @default.
- W2895921264 creator A5042659302 @default.
- W2895921264 creator A5046953322 @default.
- W2895921264 creator A5054808675 @default.
- W2895921264 creator A5082304130 @default.
- W2895921264 creator A5089226694 @default.
- W2895921264 date "2018-09-27" @default.
- W2895921264 modified "2023-10-18" @default.
- W2895921264 title "Intrinsic Social Motivation via Causal Influence in Multi-Agent RL" @default.
- W2895921264 hasPublicationYear "2018" @default.
- W2895921264 type Work @default.
- W2895921264 sameAs 2895921264 @default.
- W2895921264 citedByCount "29" @default.
- W2895921264 countsByYear W28959212642018 @default.
- W2895921264 countsByYear W28959212642019 @default.
- W2895921264 countsByYear W28959212642020 @default.
- W2895921264 countsByYear W28959212642021 @default.
- W2895921264 crossrefType "posted-content" @default.
- W2895921264 hasAuthorship W2895921264A5006947993 @default.
- W2895921264 hasAuthorship W2895921264A5035060247 @default.
- W2895921264 hasAuthorship W2895921264A5041145688 @default.
- W2895921264 hasAuthorship W2895921264A5042659302 @default.
- W2895921264 hasAuthorship W2895921264A5046953322 @default.
- W2895921264 hasAuthorship W2895921264A5054808675 @default.
- W2895921264 hasAuthorship W2895921264A5082304130 @default.
- W2895921264 hasAuthorship W2895921264A5089226694 @default.
- W2895921264 hasConcept C108650721 @default.
- W2895921264 hasConcept C111472728 @default.
- W2895921264 hasConcept C138885662 @default.
- W2895921264 hasConcept C154945302 @default.
- W2895921264 hasConcept C15744967 @default.
- W2895921264 hasConcept C187206662 @default.
- W2895921264 hasConcept C26517878 @default.
- W2895921264 hasConcept C2778496695 @default.
- W2895921264 hasConcept C38652104 @default.
- W2895921264 hasConcept C41008148 @default.
- W2895921264 hasConcept C56739046 @default.
- W2895921264 hasConcept C77805123 @default.
- W2895921264 hasConcept C79416737 @default.
- W2895921264 hasConcept C89611455 @default.
- W2895921264 hasConcept C97541855 @default.
- W2895921264 hasConceptScore W2895921264C108650721 @default.
- W2895921264 hasConceptScore W2895921264C111472728 @default.
- W2895921264 hasConceptScore W2895921264C138885662 @default.
- W2895921264 hasConceptScore W2895921264C154945302 @default.
- W2895921264 hasConceptScore W2895921264C15744967 @default.
- W2895921264 hasConceptScore W2895921264C187206662 @default.
- W2895921264 hasConceptScore W2895921264C26517878 @default.
- W2895921264 hasConceptScore W2895921264C2778496695 @default.
- W2895921264 hasConceptScore W2895921264C38652104 @default.
- W2895921264 hasConceptScore W2895921264C41008148 @default.
- W2895921264 hasConceptScore W2895921264C56739046 @default.
- W2895921264 hasConceptScore W2895921264C77805123 @default.
- W2895921264 hasConceptScore W2895921264C79416737 @default.
- W2895921264 hasConceptScore W2895921264C89611455 @default.
- W2895921264 hasConceptScore W2895921264C97541855 @default.
- W2895921264 hasLocation W28959212641 @default.
- W2895921264 hasOpenAccess W2895921264 @default.
- W2895921264 hasPrimaryLocation W28959212641 @default.
- W2895921264 hasRelatedWork W1542941925 @default.
- W2895921264 hasRelatedWork W1641379095 @default.
- W2895921264 hasRelatedWork W2064675550 @default.
- W2895921264 hasRelatedWork W2119717200 @default.
- W2895921264 hasRelatedWork W2145339207 @default.
- W2895921264 hasRelatedWork W2736601468 @default.
- W2895921264 hasRelatedWork W2758442112 @default.
- W2895921264 hasRelatedWork W2786533521 @default.
- W2895921264 hasRelatedWork W2891661335 @default.
- W2895921264 hasRelatedWork W2952370515 @default.
- W2895921264 hasRelatedWork W2962938168 @default.
- W2895921264 hasRelatedWork W2963000099 @default.
- W2895921264 hasRelatedWork W2963276097 @default.
- W2895921264 hasRelatedWork W2963627051 @default.
- W2895921264 hasRelatedWork W2964043796 @default.
- W2895921264 hasRelatedWork W2964121744 @default.
- W2895921264 hasRelatedWork W2964345382 @default.
- W2895921264 hasRelatedWork W2973029245 @default.
- W2895921264 hasRelatedWork W2979363950 @default.
- W2895921264 hasRelatedWork W3126457215 @default.
- W2895921264 isParatext "false" @default.
- W2895921264 isRetracted "false" @default.
- W2895921264 magId "2895921264" @default.
- W2895921264 workType "article" @default.