"Therearemanywaysofbeingsmartt-查字典问答网

来自任红菊的问题

　　"Therearemanywaysofbeingsmartthataren'tsmartlikeus．"ThesearethewordsofPatrickWinston，aleadingvoiceinthefieldofartificialintelligence．Althoughhisideaissimple，itssignificancehasbeenlostonmostpeople

　　Fromthe1950stothe1980s．itwasgenerallythoughtthatthebestwaytobuildsystemscapableofperformingtaskstothelevelofhumanexpertsorhigherwastocopythewaythatexpertsworked．Buttherewasaproblem：humanexpertsoftenstruggledtoexplainhowtheyperformedmanytasks．Chess-playingwasagoodexample．Whenaskedtoexplainhowtheyplayedchesswell，somegrandmastersappealedto"intuition"（真爱），othersto"experience"．Manysaidtheydidnotreallyknowatall．Howcouldresearchersbuildachess-playingsystemtobeatagrandmasterifthebestplayersthemselvescouldnotexplainhow theyweresogood？

　　Aturningpointcamein1997．DeepBlue，IBM'ssupercomputer，beatGarryKasparov，theihenworldchesschampion，inasix-gamematch．Whatwasmostremarkablewashowthesystemdidit．DeepBluedidnotshareMrKasparov's"intuition"or"experience"．Itwonbymeansofabsoluteprocessingpowerandmassivedatastorageability．

　　Considerthelegalworld．DanielMartinKatz，alawprofessor，hasdesignedasystemtopredictthevotingbehaviouroftheUSSupremeCourt．Itcanperformaswellasmostspecialists，butitdoesnotmirrorthejudgementofahumanbeing．InsteaditdrawsondatathatcapturesixdecadesofCourtbehaviour．

　　Weseesimiliardevelopmentsinotherpartsoflife．Surely，allthesepointtoafuturethatisverydifferentfromtheonemostexpertsarepredicting．Itisoftensaidthattherearemanytasksthatwillalwaysrequirehumanbeingstoperformbecausemachinescannot"think"，"reason"or"feel"likehumanbeings．Butthisistofailtograspthattomorrow'ssystemswillhandlemanytasksnotbycopyingusbutbyworkinginentirelydifferent，unhumanways．Thesetoftasksreservedparticularlyforhumanbeingsislikelytobemuchsmallerthanmanyexpect．

　　46．Theexampleofchess-playingmainlyshowsthat___．

1回答

2020-05-17 00:52

我要回答

请先登录

冯国峰

　　46．C细节理解题．根据句子humanexpertsoftenstruggledtoexplainhowtheyperformedmanytasks．Howcouldresearchersbuildachess-playingsystemtobeatagrandmasterifthebestplayersthemselvescouldnotexplainhow theyweresogood？人类专家经常努力解释他们是如何完成许多任务的．如何建立一个国际象棋博弈系统研究打败大师即使最好的球员本身不能解释他们是如何这么好，所以答案选C．

　　47．D细节理解题．根据句子DItwonbymeansofabsoluteprocessingpowerandmassivedatastorageability．它通过绝对的处理能力和海量的数据存储能力．所以答案选D．

　　48．B考虑到法律世界，作者写了法律教授DanielMartinKatz设计了一个系统来预测美国最高法院的投票行为，可见本段是作者给出的另一个例子，所以答案选B．

　　49．C本文第一段提出观点Therearemanywaysofbeingsmartthataren'tsmartlikeus，二三四段通过举例说明有很多的聪明的方法．最后一段Weseesimiliardevelopmentsinotherpartsoflife延伸到生活的其他领域，所以答案为C、

　　50．B推理判断题．根据句子Butthisistofailtograspthattomorrow'ssystemswillhandlemanytasksnotbycopyingusbutbyworkinginentirelydifferent，unhumanways．但这是无法把握明天的系统会处理很多任务不是通过复制我们的工作方式，而是在完全不同的，非人的工作方式中，可以推断将来有更多的任务是由系统处理的．所以答案选B．．

2020-05-17 00:56:27

最新问答

推荐文章

猜你喜欢

附近的人在看

推荐阅读

拓展阅读