《The Truth about AI 2of3《人工智能的真相（2023）（2）》完整中英文对照剧本.docx》由会员分享，可在线阅读，更多相关《The Truth about AI 2of3《人工智能的真相（2023）（2）》完整中英文对照剧本.docx（63页珍藏版）》请在优知文库上搜索。
1、当今世界人工智能无处不在Intoday,sworld,AIiseverywhere.我今天有哪些安排WhatappointmentsdoIhavetoday?今天有英国皇家科学院圣诞讲座TodayistheRoyalInstitutionChristmas1.ectures.请播放一首提神醒脑的歌♥Playmeawakemeupmorningmix.*你的生活又过去了一天*Anotherdayspassinginyourlife.*它突然出现在Itspoppingupinplaces许多我们几年前无法想象的地方wecouldneverhaveimaginedafewyearsag
2、o.*我正走在*monmyway.*麦克今晚的讲座你准备好了吗Mike,youallsetfortonight?我喜欢你的背景Iloveyourbackground!是啊我正在感受圣诞氛围Yeah,mjustgettingintheChristmasmood.电脑我今晚需要穿外套吗Computer,willIneedacoattonight?是的请携带外套Yes,takeacoat.我们对日常生活中的人工智能了解多少HowawareareweoftheAIinoureverydaylives?*我们正要登上.*We,reonarideto.*在哪里能找到它Wherecanwefindit?*我
3、身无分文.*Aintgotnomoney.*它能做些什么Whatdoesitdo?*你无法忽视.*Youcantignoreit.*这就是今晚讲座中我们将探讨的Thatswhatwe,llbeaskingintonightslectures.皇家科学院科学讲座演播厅假象真♥相♥人工智能大揭秘2023年英国皇家科学院圣诞讲座第二讲我的人工智能生活大家好我是麦克伍德里奇Hello,mMikeWooldridge,麦克伍德里奇教授欢迎回到本年度第二场andwelcomebacktothesecondofthisyears英国皇家科学院圣诞讲座Christmaslectur
4、esfromtheRoyalInstitution,本讲座由CGI集团赞助播出aneventsupportedbyCGI.在上场讲座中我们了解到Inthelastlecture,wesawhowtoday,s当今人工智能是受人脑后发发展而来的artificialintelligencewasinspiredbythehumanbrain.本次讲座我们将深入探讨Inthislecture,we,regoingtoinvestigate人工智能如何塑造我们今天的生活howAIisshapingourlivestoday,我们将看到游戏在推动人工智能发展方面andweregoingtoseehowg
5、ameshaveplayed起到了怎样重要的作用anincrediblyimportantroleinadvancingAI.我们过会儿将再回到We,regoingtocomebacktothevideothatwejustsaw刚才所看的影片alittlebitlateroninthelecture,仔细聊聊该影片中andwe,regoingtodiscussexactly人工智能扮演的角色whereAIfeaturesinthatvideo.但首先我们来探讨一下人工智能Butfirst,weregoingtostartbyexploring在我们最流行的娱乐形式电子游戏中howAIfeat
6、uresinoneofourmostpopularforms起到的重要作用ofentertainment-videogames.有请跑车浪漫旅GT赛车顶尖选手马丁格雷迪PleasewelcometopGTdriverMartinGrady,和索尼人工智能公♥司♥的考什克苏布拉马尼安andfromSonyAI,KaushikSubramanian.欢迎你马丁谢谢Welcome,Martin.-Thankyou.请坐Takeaseat.别客气.andmakeyourselfcomfortable.马丁你玩赛车游戏多久了Now,Martin,howlonghaveyoub
7、eenracing?我玩赛车游戏二十五年了vebeenracingfor25years.其中十二年是职业选干Andvebeenapro-racingdriverfor12ofthose.好吧我们今天想请你OK,sowhatweregoingtogetyoutodotoday和人工智能一较高下isweregoingtogetyoutoplayagainstAI.我们将安排你和索尼人工智能公♥司♥WeregoingtogetyoutoplayagainstaracingAI为跑车浪漫旅开♥发♥的thatwasdevelopedbySonyAI
8、赛车人工智能进行比赛forthegameofGranTurismo.顺便问一下观众中Bytheway,whointheaudiencehas谁家有这些东西吗gotoneofthesethingsathome?谁家有吗Anybody?少数几个人家有这些东西Acoupleofpeoplehavegotoneofthose.马丁请开始吧So,Martin,yougetyourselfgoing.告诉我们考什克Kaushik,tellus,你构建该人工智能是想做什么whatdidyouwanttodowhenyoubuiltthisAI?好吧我们向自己提出的第一个挑战是Right,thefirstch
9、allengewhatweaskedourselvesis,我们能开♥发♥出一款能在比赛中击败canwedevelopanAIthatcanracetobeattheworld,sbestGT赛车世界冠军的人工智能吗GranTurismoracersinacompetitiverace?人工智能可以学会赛车所需的AndcantheAIlearntherequiredskills高阶技能吗toperformatthatlevel?事实证明我们能Itturnsoutwecan.索菲赛车游戏中人工智能的名字考什克该游戏中索菲在哪里Kaushik,whereexactlyi
10、sSophyappearinginthis,theAI?马丁正和其他十九辆人工智能赛车比赛So,Martinsracedwith19otherAIcars,它们都由GT索菲控制andthey*reallcontrolledbyGranTurismoSophyhere.从这里就能看出来Andthewayyoucanseethatis,在左手边ifyoulookontheleft-handside,你可以看到所有玩家车辆的名字youcanseeallthenamesoftheplayercars.这里有个心形图标Andtheresalogowhichlookslikeaheart,所有带这个图标的
11、andallofthoselogos都是GT索菲控制的车aretheGranTurismoSophycars.所有这些车都知道自己是哪辆Andallofthosecars,theyknowwhatcartheyredriving,它们知道彼此theyreawareofeachother当然也能意识到马丁的存在andtheyreofcourseawareofMartinaswell.好吧但它们并不想联手对付马丁对吗OK,buttheyrenotalltryingtoganguponMartin,arethey?它们都是独&hearts位♥的吗Theyreallindependent
12、?它们都想让自己获得第一名Theyrealltryingtogettofirstplaceindependently.索菲是如何变得这么出色的呢SohowdidSophy,theAI,getgoodenoughinthisrace?人工智能GT索菲ThisA1.GranTurismoSophy,经过了一种称作“强化学习”的方法训练istrainedusinganapproachcalledreinforcementlearning,也就是人工智能通过试错来学习anapproachwheretheAIlearnsbytrialanderror.一开始索菲不太了解驾驶So,atthestart,i
13、treallydoesntknowmuchaboutdriving,更别提比赛规则了muchlesstherulesofracing.我们这里有训练视频OK,Ithinkwevegotthevideoshowinghereofthetraining.好吧在槐顷里Right,andwhenyouwatchthis,你们能看到赛车相撞youregoingtoseethecarscolliding,车辆完全失控youregoingtoseethecarscompletelylosingcontrol,冲出赛道goingoffthetrack.视频里是刚训练几小时后的结果Andthisisjustwi
14、thinafewhoursoftraining.就是这个样子Thisiswhatitlookslike.过了一段时间通过强化学习Andafterawhile,withreinforcementlearning,它尝试了许多不同的事ittriesmanydifferentthings最终学会正确的操作顺序andthenitlearns,finally,therightsequence能够进行比赛了ofactionstobeabletorace,就像马丁现在正看到的那样likethewayMartinslookingatrightnow.显然马丁撞车了Orcrashing,apparently.好
15、吧我们稍后回来OK,now,weregoingtocomebacklateroninthelecture继续r解更多“强化学习的相关内容andweregoingtolearnalotmoreaboutreinforcementlearning.考什克但为什么像这样的电子游戏But,Kaushik,whyisavideogamelikethissodifficult很难用强化学习的方法forreinforcementlearning,来让机器进行学习andjustformachinelearninggenerally?赛车游戏有许多变量RacinghasalotOfvariability.情况瞬息万变Thereareeverchangingsituations.当你在那些情形下处于失控边缘时Whenyouredrivingattheedgeofcontrolinthosesituations,很难进
免费区 