Too many cooks: Bayesian inference for coordinating multi-agent collaboration

Collaboration requires agents to coordinate their behavior on the fly, sometimes cooperating to solve a single task together and other times dividing it up into sub-tasks to work on in parallel. Underlying the human ability to collaborate is theory-of-mind, the ability to infer the hidden mental sta...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Wang, Rose E, Wu, Sarah A, Evans, James A, Tenenbaum, Joshua B, Parkes, David C, Kleiman-Weiner, Max
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator Wang, Rose E
Wu, Sarah A
Evans, James A
Tenenbaum, Joshua B
Parkes, David C
Kleiman-Weiner, Max
description Collaboration requires agents to coordinate their behavior on the fly, sometimes cooperating to solve a single task together and other times dividing it up into sub-tasks to work on in parallel. Underlying the human ability to collaborate is theory-of-mind, the ability to infer the hidden mental states that drive others to act. Here, we develop Bayesian Delegation, a decentralized multi-agent learning mechanism with these abilities. Bayesian Delegation enables agents to rapidly infer the hidden intentions of others by inverse planning. We test Bayesian Delegation in a suite of multi-agent Markov decision processes inspired by cooking problems. On these tasks, agents with Bayesian Delegation coordinate both their high-level plans (e.g. what sub-task they should work on) and their low-level actions (e.g. avoiding getting in each other's way). In a self-play evaluation, Bayesian Delegation outperforms alternative algorithms. Bayesian Delegation is also a capable ad-hoc collaborator and successfully coordinates with other agent types even in the absence of prior experience. Finally, in a behavioral experiment, we show that Bayesian Delegation makes inferences similar to human observers about the intent of others. Together, these results demonstrate the power of Bayesian Delegation for decentralized multi-agent collaboration.
doi_str_mv 10.48550/arxiv.2003.11778
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2003_11778</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2003_11778</sourcerecordid><originalsourceid>FETCH-LOGICAL-a678-ef6712d9bd93699b5231f5f6a9a6484dcf2b1123c15cce5fc933407e876f8d283</originalsourceid><addsrcrecordid>eNotz7tOwzAYhmEvDKhwAUz1DST4EJ_YSsVJatUle_THh8oisSsnIHL3tIXpGx7pk16EHiipGy0EeYTyE79rRgivKVVK36J9mzMeIS3Y5vw5PeFnWPwUIeGYgi8-WY9DLhctLiaYYzri8WuYYwVHn-YzDAP0uZwlpzt0E2CY_P3_rlD7-tJu36vd4e1ju9lVIJWufJCKMmd6Z7g0pheM0yCCBAOy0Y2zgfWUMm6psNaLYA3nDVFeKxm0Y5qv0Prv9trTnUocoSzdpau7dvFfj7BI9A</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Too many cooks: Bayesian inference for coordinating multi-agent collaboration</title><source>arXiv.org</source><creator>Wang, Rose E ; Wu, Sarah A ; Evans, James A ; Tenenbaum, Joshua B ; Parkes, David C ; Kleiman-Weiner, Max</creator><creatorcontrib>Wang, Rose E ; Wu, Sarah A ; Evans, James A ; Tenenbaum, Joshua B ; Parkes, David C ; Kleiman-Weiner, Max</creatorcontrib><description>Collaboration requires agents to coordinate their behavior on the fly, sometimes cooperating to solve a single task together and other times dividing it up into sub-tasks to work on in parallel. Underlying the human ability to collaborate is theory-of-mind, the ability to infer the hidden mental states that drive others to act. Here, we develop Bayesian Delegation, a decentralized multi-agent learning mechanism with these abilities. Bayesian Delegation enables agents to rapidly infer the hidden intentions of others by inverse planning. We test Bayesian Delegation in a suite of multi-agent Markov decision processes inspired by cooking problems. On these tasks, agents with Bayesian Delegation coordinate both their high-level plans (e.g. what sub-task they should work on) and their low-level actions (e.g. avoiding getting in each other's way). In a self-play evaluation, Bayesian Delegation outperforms alternative algorithms. Bayesian Delegation is also a capable ad-hoc collaborator and successfully coordinates with other agent types even in the absence of prior experience. Finally, in a behavioral experiment, we show that Bayesian Delegation makes inferences similar to human observers about the intent of others. Together, these results demonstrate the power of Bayesian Delegation for decentralized multi-agent collaboration.</description><identifier>DOI: 10.48550/arxiv.2003.11778</identifier><language>eng</language><subject>Computer Science - Artificial Intelligence ; Computer Science - Learning ; Computer Science - Multiagent Systems</subject><creationdate>2020-03</creationdate><rights>http://arxiv.org/licenses/nonexclusive-distrib/1.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,776,881</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2003.11778$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2003.11778$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Wang, Rose E</creatorcontrib><creatorcontrib>Wu, Sarah A</creatorcontrib><creatorcontrib>Evans, James A</creatorcontrib><creatorcontrib>Tenenbaum, Joshua B</creatorcontrib><creatorcontrib>Parkes, David C</creatorcontrib><creatorcontrib>Kleiman-Weiner, Max</creatorcontrib><title>Too many cooks: Bayesian inference for coordinating multi-agent collaboration</title><description>Collaboration requires agents to coordinate their behavior on the fly, sometimes cooperating to solve a single task together and other times dividing it up into sub-tasks to work on in parallel. Underlying the human ability to collaborate is theory-of-mind, the ability to infer the hidden mental states that drive others to act. Here, we develop Bayesian Delegation, a decentralized multi-agent learning mechanism with these abilities. Bayesian Delegation enables agents to rapidly infer the hidden intentions of others by inverse planning. We test Bayesian Delegation in a suite of multi-agent Markov decision processes inspired by cooking problems. On these tasks, agents with Bayesian Delegation coordinate both their high-level plans (e.g. what sub-task they should work on) and their low-level actions (e.g. avoiding getting in each other's way). In a self-play evaluation, Bayesian Delegation outperforms alternative algorithms. Bayesian Delegation is also a capable ad-hoc collaborator and successfully coordinates with other agent types even in the absence of prior experience. Finally, in a behavioral experiment, we show that Bayesian Delegation makes inferences similar to human observers about the intent of others. Together, these results demonstrate the power of Bayesian Delegation for decentralized multi-agent collaboration.</description><subject>Computer Science - Artificial Intelligence</subject><subject>Computer Science - Learning</subject><subject>Computer Science - Multiagent Systems</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2020</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNotz7tOwzAYhmEvDKhwAUz1DST4EJ_YSsVJatUle_THh8oisSsnIHL3tIXpGx7pk16EHiipGy0EeYTyE79rRgivKVVK36J9mzMeIS3Y5vw5PeFnWPwUIeGYgi8-WY9DLhctLiaYYzri8WuYYwVHn-YzDAP0uZwlpzt0E2CY_P3_rlD7-tJu36vd4e1ju9lVIJWufJCKMmd6Z7g0pheM0yCCBAOy0Y2zgfWUMm6psNaLYA3nDVFeKxm0Y5qv0Prv9trTnUocoSzdpau7dvFfj7BI9A</recordid><startdate>20200326</startdate><enddate>20200326</enddate><creator>Wang, Rose E</creator><creator>Wu, Sarah A</creator><creator>Evans, James A</creator><creator>Tenenbaum, Joshua B</creator><creator>Parkes, David C</creator><creator>Kleiman-Weiner, Max</creator><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20200326</creationdate><title>Too many cooks: Bayesian inference for coordinating multi-agent collaboration</title><author>Wang, Rose E ; Wu, Sarah A ; Evans, James A ; Tenenbaum, Joshua B ; Parkes, David C ; Kleiman-Weiner, Max</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a678-ef6712d9bd93699b5231f5f6a9a6484dcf2b1123c15cce5fc933407e876f8d283</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2020</creationdate><topic>Computer Science - Artificial Intelligence</topic><topic>Computer Science - Learning</topic><topic>Computer Science - Multiagent Systems</topic><toplevel>online_resources</toplevel><creatorcontrib>Wang, Rose E</creatorcontrib><creatorcontrib>Wu, Sarah A</creatorcontrib><creatorcontrib>Evans, James A</creatorcontrib><creatorcontrib>Tenenbaum, Joshua B</creatorcontrib><creatorcontrib>Parkes, David C</creatorcontrib><creatorcontrib>Kleiman-Weiner, Max</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Wang, Rose E</au><au>Wu, Sarah A</au><au>Evans, James A</au><au>Tenenbaum, Joshua B</au><au>Parkes, David C</au><au>Kleiman-Weiner, Max</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Too many cooks: Bayesian inference for coordinating multi-agent collaboration</atitle><date>2020-03-26</date><risdate>2020</risdate><abstract>Collaboration requires agents to coordinate their behavior on the fly, sometimes cooperating to solve a single task together and other times dividing it up into sub-tasks to work on in parallel. Underlying the human ability to collaborate is theory-of-mind, the ability to infer the hidden mental states that drive others to act. Here, we develop Bayesian Delegation, a decentralized multi-agent learning mechanism with these abilities. Bayesian Delegation enables agents to rapidly infer the hidden intentions of others by inverse planning. We test Bayesian Delegation in a suite of multi-agent Markov decision processes inspired by cooking problems. On these tasks, agents with Bayesian Delegation coordinate both their high-level plans (e.g. what sub-task they should work on) and their low-level actions (e.g. avoiding getting in each other's way). In a self-play evaluation, Bayesian Delegation outperforms alternative algorithms. Bayesian Delegation is also a capable ad-hoc collaborator and successfully coordinates with other agent types even in the absence of prior experience. Finally, in a behavioral experiment, we show that Bayesian Delegation makes inferences similar to human observers about the intent of others. Together, these results demonstrate the power of Bayesian Delegation for decentralized multi-agent collaboration.</abstract><doi>10.48550/arxiv.2003.11778</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.2003.11778
ispartof
issn
language eng
recordid cdi_arxiv_primary_2003_11778
source arXiv.org
subjects Computer Science - Artificial Intelligence
Computer Science - Learning
Computer Science - Multiagent Systems
title Too many cooks: Bayesian inference for coordinating multi-agent collaboration
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-30T23%3A43%3A36IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Too%20many%20cooks:%20Bayesian%20inference%20for%20coordinating%20multi-agent%20collaboration&rft.au=Wang,%20Rose%20E&rft.date=2020-03-26&rft_id=info:doi/10.48550/arxiv.2003.11778&rft_dat=%3Carxiv_GOX%3E2003_11778%3C/arxiv_GOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true