Too many cooks: Coordinating multi-agent collaboration through inverse planning
© 2020 International Foundation for Autonomous Agents and Multiagent Systems (IFAAMAS). All rights reserved. Humans collaborate in dynamic and flexible ways. Collaboration requires agents to coordinate their behavior on the fly, sometimes jointly solving a single task together and other times dividi...
Main Authors: | , , , , , |
---|---|
Other Authors: | |
Format: | Article |
Language: | English |
Published: |
2021
|
Online Access: | https://hdl.handle.net/1721.1/138369 |
_version_ | 1811083559848378368 |
---|---|
author | Wang, RE Wu, SA Evans, JA Tenenbaum, JB Parkes, DC Kleiman-Weiner, M |
author2 | Massachusetts Institute of Technology. Department of Brain and Cognitive Sciences |
author_facet | Massachusetts Institute of Technology. Department of Brain and Cognitive Sciences Wang, RE Wu, SA Evans, JA Tenenbaum, JB Parkes, DC Kleiman-Weiner, M |
author_sort | Wang, RE |
collection | MIT |
description | © 2020 International Foundation for Autonomous Agents and Multiagent Systems (IFAAMAS). All rights reserved. Humans collaborate in dynamic and flexible ways. Collaboration requires agents to coordinate their behavior on the fly, sometimes jointly solving a single task together and other times dividing it up into sub-tasks to work on in parallel. We develop Bayesian Delegation, a learning mechanism for decentralized multi-agent coordination that enables agents to rapidly infer the sub-tasks that other agents are working on by inverse planning. These inferences enable agents to determine, in the absence of communication, whether to plan jointly with others or work on complementary sub-tasks. We test this model in a suite of decentralized multi-agent environments inspired by cooking problems. To succeed, agents must coordinate both their high-level plans (sub-task) and their low-level actions (avoiding collisions). Including joint sub-tasks in the prior of Bayesian delegation enables agents to carry out sub-tasks that neither agent can finish independently. The full system outperforms lesioned systems that are missing one or more of these capabilities. |
first_indexed | 2024-09-23T12:35:00Z |
format | Article |
id | mit-1721.1/138369 |
institution | Massachusetts Institute of Technology |
language | English |
last_indexed | 2024-09-23T12:35:00Z |
publishDate | 2021 |
record_format | dspace |
spelling | mit-1721.1/1383692023-02-03T19:41:52Z Too many cooks: Coordinating multi-agent collaboration through inverse planning Wang, RE Wu, SA Evans, JA Tenenbaum, JB Parkes, DC Kleiman-Weiner, M Massachusetts Institute of Technology. Department of Brain and Cognitive Sciences Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory Center for Brains, Minds, and Machines © 2020 International Foundation for Autonomous Agents and Multiagent Systems (IFAAMAS). All rights reserved. Humans collaborate in dynamic and flexible ways. Collaboration requires agents to coordinate their behavior on the fly, sometimes jointly solving a single task together and other times dividing it up into sub-tasks to work on in parallel. We develop Bayesian Delegation, a learning mechanism for decentralized multi-agent coordination that enables agents to rapidly infer the sub-tasks that other agents are working on by inverse planning. These inferences enable agents to determine, in the absence of communication, whether to plan jointly with others or work on complementary sub-tasks. We test this model in a suite of decentralized multi-agent environments inspired by cooking problems. To succeed, agents must coordinate both their high-level plans (sub-task) and their low-level actions (avoiding collisions). Including joint sub-tasks in the prior of Bayesian delegation enables agents to carry out sub-tasks that neither agent can finish independently. The full system outperforms lesioned systems that are missing one or more of these capabilities. 2021-12-07T20:04:14Z 2021-12-07T20:04:14Z 2020-01-01 2021-12-07T19:57:09Z Article http://purl.org/eprint/type/ConferencePaper https://hdl.handle.net/1721.1/138369 Wang, RE, Wu, SA, Evans, JA, Tenenbaum, JB, Parkes, DC et al. 2020. "Too many cooks: Coordinating multi-agent collaboration through inverse planning." Proceedings of the International Joint Conference on Autonomous Agents and Multiagent Systems, AAMAS, 2020-May. en https://cognitivesciencesociety.org/cogsci20/papers/0157/index.html Proceedings of the International Joint Conference on Autonomous Agents and Multiagent Systems, AAMAS Creative Commons Attribution 4.0 International license https://creativecommons.org/licenses/by/4.0/ application/pdf Cognitive Science Society |
spellingShingle | Wang, RE Wu, SA Evans, JA Tenenbaum, JB Parkes, DC Kleiman-Weiner, M Too many cooks: Coordinating multi-agent collaboration through inverse planning |
title | Too many cooks: Coordinating multi-agent collaboration through inverse planning |
title_full | Too many cooks: Coordinating multi-agent collaboration through inverse planning |
title_fullStr | Too many cooks: Coordinating multi-agent collaboration through inverse planning |
title_full_unstemmed | Too many cooks: Coordinating multi-agent collaboration through inverse planning |
title_short | Too many cooks: Coordinating multi-agent collaboration through inverse planning |
title_sort | too many cooks coordinating multi agent collaboration through inverse planning |
url | https://hdl.handle.net/1721.1/138369 |
work_keys_str_mv | AT wangre toomanycookscoordinatingmultiagentcollaborationthroughinverseplanning AT wusa toomanycookscoordinatingmultiagentcollaborationthroughinverseplanning AT evansja toomanycookscoordinatingmultiagentcollaborationthroughinverseplanning AT tenenbaumjb toomanycookscoordinatingmultiagentcollaborationthroughinverseplanning AT parkesdc toomanycookscoordinatingmultiagentcollaborationthroughinverseplanning AT kleimanweinerm toomanycookscoordinatingmultiagentcollaborationthroughinverseplanning |