上传者: lull0815
|
上传时间:2024/4/15 1:20:12
|
文件大小:3.84MB
|
文件类型:pdf
MasteringthegameofGowithouthumanknowledge英文高清完整.pdf版下载
讲述alphazero的原文,发表在nature。
Along-standinggoalofartificialintelligenceisanalgorithmthatlearns,tabularasa,superhumanproficiencyinchallengingdomains.Recently,AlphaGobecamethefirstprogramtodefeataworldchampioninthegameofGo.ThetreesearchinAlphaGoevaluatedpositionsandselectedmovesusingdeepneuralnetworks.Theseneuralnetworksweretrainedbysupervisedlearningfromhumanexpertmoves,andbyreinforcementlearningfromself-play.Hereweintroduceanalgorithmbasedsolelyonreinforcementlearning,withouthumandata,guidanceordomainknowledgebeyondgamerules.AlphaGobecomesitsownteacher:aneuralnetworkistrainedtopredictAlphaGo’sownmoveselectionsandalsothewinnerofAlphaGo’sgames.Thisneuralnetworkimprovesthestrengthofthetreesearch,resultinginhigherqualitymoveselectionandstrongerself-playinthenextiteration.Startingtabularasa,ournewprogramAlphaGoZeroachievedsuperhumanperformance,winning100–0againstthepreviouslypublished,champion-defeatingAlphaGo.
本软件ID:10029951