{"id":"https://openalex.org/W3137166149","doi":null,"title":"Softmax with Regularization: Better Value Estimation in Multi-Agent Reinforcement Learning.","display_name":"Softmax with Regularization: Better Value Estimation in Multi-Agent Reinforcement Learning.","publication_year":2021,"publication_date":"2021-03-22","ids":{"openalex":"https://openalex.org/W3137166149","mag":"3137166149"},"language":"en","primary_location":{"id":"mag:3137166149","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2103.11883","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"arXiv (Cornell University)","raw_type":null},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":[],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/abs/2103.11883","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5090888372","display_name":"Ling Pan","orcid":"https://orcid.org/0000-0002-4413-281X","observed_orcids":["https://orcid.org/0000-0002-4413-281X"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ling Pan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5003400295","display_name":"Tabish Rashid","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tabish Rashid","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Bei Peng","orcid":"https://orcid.org/0000-0003-2486-6776","observed_orcids":["https://orcid.org/0000-0003-2486-6776"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Bei Peng","raw_affiliation_strings":[],"raw_orcid":"https://orcid.org/0000-0003-2486-6776","affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5082905458","display_name":"Longbo Huang","orcid":"https://orcid.org/0000-0002-7341-447X","observed_orcids":["https://orcid.org/0000-0002-7341-447X"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Longbo Huang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5056879203","display_name":"Shimon Whiteson","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shimon Whiteson","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":3,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10462","display_name":"Reinforcement Learning in Robotics","score":1.0,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T10462","display_name":"Reinforcement Learning in Robotics","score":1.0,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T11574","display_name":"Artificial Intelligence in Games","score":9.999999747378752e-05,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T12101","display_name":"Advanced Bandit Algorithms Research","score":0.0,"subfield":{"id":"https://openalex.org/subfields/1803","display_name":"Management Science and Operations Research"},"field":{"id":"https://openalex.org/fields/18","display_name":"Decision Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/overestimation-bias","display_name":"overestimation bias","score":0.7689999938011169},{"id":"https://openalex.org/keywords/q-learning","display_name":"Q-learning","score":0.6370000243186951},{"id":"https://openalex.org/keywords/multi-agent-reinforcement-learning","display_name":"multi-agent reinforcement learning","score":0.8569999933242798},{"id":"https://openalex.org/keywords/qmix","display_name":"QMIX","score":0.9459999799728394},{"id":"https://openalex.org/keywords/softmax-approximation","display_name":"softmax approximation","score":0.7549999952316284},{"id":"https://openalex.org/keywords/regularization","display_name":"regularization","score":0.9229999780654907},{"id":"https://openalex.org/keywords/cooperative-multi-agent-tasks","display_name":"cooperative multi-agent tasks","score":0.7009999752044678},{"id":"https://openalex.org/keywords/starcraft-ii-micromanagement","display_name":"StarCraft II micromanagement","score":0.8519999980926514},{"id":"https://openalex.org/keywords/value-estimation","display_name":"value estimation","score":0.5630000233650208}],"concepts":[{"id":"https://openalex.org/C188441871","wikidata":"https://www.wikidata.org/wiki/Q7554146","display_name":"Softmax function","level":3,"score":0.9164424538612366},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.8998445272445679},{"id":"https://openalex.org/C2776135515","wikidata":"https://www.wikidata.org/wiki/Q17143721","display_name":"Regularization (linguistics)","level":2,"score":0.757783830165863},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6429563164710999},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5939347743988037},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5231524705886841},{"id":"https://openalex.org/C67203356","wikidata":"https://www.wikidata.org/wiki/Q1321905","display_name":"Reinforcement","level":2,"score":0.4610629975795746},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.3550748825073242},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.17948195338249207},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.15839090943336487},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.0962648093700409},{"id":"https://openalex.org/C66938386","wikidata":"https://www.wikidata.org/wiki/Q633538","display_name":"Structural engineering","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"mag:3137166149","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2103.11883","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"arXiv (Cornell University)","raw_type":null}],"best_oa_location":{"id":"mag:3137166149","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2103.11883","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"arXiv (Cornell University)","raw_type":null},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":33,"referenced_works":["https://openalex.org/W51508254","https://openalex.org/W2012812921","https://openalex.org/W2150468603","https://openalex.org/W2155968351","https://openalex.org/W2173248099","https://openalex.org/W2292533394","https://openalex.org/W2396783599","https://openalex.org/W2419612459","https://openalex.org/W2553109721","https://openalex.org/W2596758708","https://openalex.org/W2596982695","https://openalex.org/W2623431351","https://openalex.org/W2754879180","https://openalex.org/W2787938642","https://openalex.org/W2804380964","https://openalex.org/W2807741983","https://openalex.org/W2946606218","https://openalex.org/W2951984055","https://openalex.org/W2963762747","https://openalex.org/W2978148161","https://openalex.org/W2981021427","https://openalex.org/W2981141126","https://openalex.org/W2995509794","https://openalex.org/W3004640943","https://openalex.org/W3046288222","https://openalex.org/W3089091950","https://openalex.org/W3093287223","https://openalex.org/W3093348818","https://openalex.org/W3100823194","https://openalex.org/W3100860002","https://openalex.org/W3104860527","https://openalex.org/W3118977819","https://openalex.org/W3163926178"],"related_works":["https://openalex.org/W3168274109","https://openalex.org/W4393241368","https://openalex.org/W7124285129","https://openalex.org/W4388232295","https://openalex.org/W2798511001","https://openalex.org/W3199596680","https://openalex.org/W7011180103","https://openalex.org/W4404570667","https://openalex.org/W7131910208","https://openalex.org/W3200702333"],"abstract_inverted_index":{"Overestimation":[0],"in":[1,12,22,81,96,131],"$Q$-learning":[2,36,150],"is":[3,52],"an":[4],"important":[5],"problem":[6,50],"that":[7,32,68,109,162],"has":[8,17],"been":[9],"extensively":[10],"studied":[11],"single-agent":[13],"reinforcement":[14,41],"learning,":[15],"but":[16],"received":[18],"comparatively":[19],"little":[20],"attention":[21],"the":[23,97,103,140],"multi-agent":[24,40,98,136],"setting.":[25],"In":[26],"this":[27,60],"work,":[28],"we":[29,93,145,160],"empirically":[30],"demonstrate":[31,78,108,139],"QMIX,":[33,119],"a":[34,46,63,75,89,132,166,171],"popular":[35],"algorithm":[37],"for":[38],"cooperative":[39,135],"learning":[42],"(MARL),":[43],"suffers":[44],"from":[45,74],"particularly":[47],"severe":[48,125],"overestimation":[49,105,126],"which":[51,92],"not":[53],"mitigated":[54],"by":[55,61],"existing":[56],"approaches.":[57],"We":[58,84,107],"rectify":[59],"designing":[62],"novel":[64],"regularization-based":[65],"update":[66],"scheme":[67],"penalizes":[69],"large":[70],"joint":[71],"action-values":[72],"deviating":[73],"baseline":[76],"and":[77,127,154],"its":[79,121],"effectiveness":[80],"stabilizing":[82],"learning.":[83],"additionally":[85],"propose":[86],"to":[87,100,118,148],"employ":[88],"softmax":[90],"operator,":[91],"efficiently":[94],"approximate":[95],"setting,":[99],"further":[101],"reduce":[102],"potential":[104],"bias.":[106],"our":[110,143,163],"Softmax":[111],"with":[112],"Regularization":[113],"(SR)":[114],"method,":[115,144],"when":[116],"applied":[117],"accomplishes":[120],"goal":[122],"of":[123,134,142,173],"avoiding":[124],"significantly":[128],"improves":[129],"performance":[130,157,168],"variety":[133],"tasks.":[137,178],"To":[138],"versatility":[141],"apply":[146],"it":[147],"other":[149],"based":[151],"MARL":[152],"algorithms":[153],"achieve":[155],"similar":[156],"gains.":[158],"Finally,":[159],"show":[161],"method":[164],"provides":[165],"consistent":[167],"improvement":[169],"on":[170],"set":[172],"challenging":[174],"StarCraft":[175],"II":[176],"micromanagement":[177]},"counts_by_year":[{"year":2021,"cited_by_count":1},{"year":2020,"cited_by_count":2}],"updated_date":"2026-11-03T10:50:19.383937","created_date":"2021-03-29T00:00:00"}