搜索
ArtificialIntelligenceCommunications1202Group7ArtificialIntelligenceSearchenginesLaborAppliancesSciencecontent•Definition•Meanings•Historyandgreatmen•Developinganddifficulties•Knowledge•Currentsituation•ApplicationsandAchievements•FutureDefinition定义ArtificialIntelligence(AI)istheintelligenceofmachinesandthebranchofcomputersciencethataimstocreateit.DefinitioninAItextbook:”thestudyanddesignofintelligentagents”人工智能（AI）是机器智能与计算机科学，旨在创建它的分支。在AI教科书的定义：“学习的智能代理的设计。”•人工智能是计算机科学的前沿，充满机遇和挑战。•“AstudentinphysicsmightreasonablyfeelthatallthegoodideashavealreadybeentakenbyGalileo,Newton,Einstein,andtherest,andthatittakesmanyyearsofstudybeforeonecancontributenewideas,AI,ontheotherhand,stillhasopeningsforafull-timeEinstein.”•_______《ArtificialIntelligence:AmodernApproach》•（一位在物理学领域的学生会理所当然的认为所有的好点子已经被伽利略，牛顿，爱因斯坦和其他人，它需要许多年前的研究能做出贡献的新思路，另一方面，AI，仍然作为一个全职的爱因斯坦。）“Scienceofmakingmachinesdothingsthatwouldrequireintelligenceifdonebyman.”让机器做的事情，需要智慧，如果做了人的科学。Minsky“makecomputermoreusefulandtounderstandprinciplesthatmakeintelligencepossible.”使电脑更加有用和明白原则，使智力成为可能。WinstonTheexcitingnewefforttomakecomputersthinks…machinewithminds,inthefullandliteralsense”(Haugeland1985)“Theartofcreatingmachinesthatperformfunctionsthatrequireintelligencewhenperformedbypeople”(Kurzweil,1990)“Thestudyofmentalfacultiesthroughtheuseofcomputationalmodels”(Charniaketal.1985)Afieldofstudythatseekstoexplainandemulateintelligentbehaviorintermsofcomputationalprocesses”(Schalkol,1990)SystemsthatthinklikehumansSystemsthatthinkrationallySystemsthatactlikehumansSystemsthatactrationallyViewsofAIfallintofourcategories:WhatisAI?1956BirthEnthusiasm（狂热）1966DifficultiesNeuralNetworksReborn1986ComputationalIntelligence1969IJCAIwasheld1970AIwaslaunchedKnowledgeEngineering1977Present•国际人工智能联合会议TheHistoryofAISecondaryTranslation-MachineTranslationthespiritiswillingbutthefleshisweakthevodkaisgoodbutthemeatisrottenCombinatorialexplosionproblem(组合爆炸)Thefactthataprogramcanfindasolutioninprincipledoesnotmeansthattheprogramcontainsanyofthemechanismsneededtofinditinpractice.Perceptronlimitations(感知机局限性)Atwo-inputperceptroncannotbetrainedtorecognizewhenitstwoinputsaredifferent.DifficultiesMcCarthy麦卡锡Minsky明斯基Shannon香农TheAmericanscholarMcCarthywhoisconsideredthefatherofartificialintelligenceoforganizedaSociety,manyexpertsandscholarsgatheredtogetherforadiscussionofuptotwomonths.InvitingthemtoparticipateintheSummerDartmouthArtificialIntelligenceResearchSociety.Sincethen,thisfieldisnamedartificialintelligenceMcCarthy麦卡锡Inmachinelearning,in1957Rosenblatt(罗森布拉特)successfullydevelopedPerceptron(感知机);在1957年机器产生时代，罗森布拉特成功的研制出感知机。achievementIntheoremproving,in1958WangHaointheIBM-704machinewitha3-5minuteprovedall220theorem“PrincipiaMathematica”intherelevantpropositionalcalculus;In1965,Robinsonmadeduetotheresolutionprinciple(归结原理);定理证明，1958年王皓用IBM-704机用3-5分钟的证明所有220定理“数学原理”中的相关命题演算;1965年，罗宾逊作出了应有的决议原则。Inpatternrecognition(模式识别),1959Selfridge(塞尔福里奇)launchedapatternrecognitionprogram;1965Robertspreparedadistinguishedbuildingblockconstructionprocedures(可分辨积木构造程序);1959年塞尔弗里奇用模式识别推出了模式识别程序;1965年罗伯茨准备了杰出的建筑模块结构的程序。Intermsofproblemsolving(问题求解):1960Newell(纽厄尔)compiledauniversalproblemsolver(GPS),cansolve11differenttypesofproblems;Inexpertsystems(专家系统):theresultsofthe1968Feigenbaum(费根鲍姆)developedDENDRALexpertsystemandputintouse;achievement在专家系统：1968的结果费根鲍姆开发DENDRAL“专家系统”并投入使用；在解决问题方面：1960纽厄尔编译一个普遍的问题解决者(GPS),可以解决11个不同类型的问题;May1997,theU.S.IBM‘s“DeepBlue”supercomputerbeattheworldchesschampionGarryKasparov(卡斯帕罗夫).Thegreatachievementsofartificialintelligence1997五月，美国IBM的“深蓝”超级计算机击败了国际象棋世界冠军卡斯帕罗夫（卡斯帕罗夫）。ComputationalIntelligencethriving(90s-now)ArtificialNeuralNetwork人工神经网络GeneticAlgorithm遗传算法FuzzyReasoning模糊推理ParticleSwarmOptimization粒子群优化ArtificialImmuneSystem人工免疫系统Artificialintelligence(人工智能)Turingtest(图灵测试)patternrecognition(模式识别)expertsystems(专家系统)NeuralNetwork(神经网络)TheTuringtest(1936)Alanturing（1912-1954）TheTuringtestisatestofamachine'sabilitytoexhibitintelligentbehaviororindistinguishablefromthatofahuman.Ifthejudgercannotreliablytellthemachinefromthehuman,themachineissaidtohavepassedthetest.图灵测试是一种测试机器的能力,表现出人类的智能行为或难以区分的。如果测试的人不能明确地区分出人类和机器,这台机器就通过了测试。Applications•Themainareas•Generalmachineintelligence,conversationlbehavior,data-mining,•一般机器智能,对话行为,数据挖掘•Driverlesscars,robotsoccerandgames•无人驾驶汽车,机器人足球和游戏RobotsoccerConversationalbehavior•Acomprehensivesurveypaperprovidesathoroughreviewofintelligentsystemsinprocessengineering.TheprocesscontrolresearchinthisareahaslargelybeenconcernedwiththreeAImethods:knowledge-basedsystems,neuralnerworksandfuzzylogic.•一个广泛的调查提供了一个全面的关于智能技术在工程进程中的一系列问题的复查，针对此领域的这项进程控制调查研究主要集中在三个方面：知识库系统、神经网络和模糊逻辑。AImethods:knowledge-basedsystems,NeuralnetworksandFuzzylogicprovides人工智能主要集中在三个方面：知识库系统、神经网络和模糊逻辑•Knowledge-basedsystems(KBS),alsoreferredtoasexpertsystems,useasetof“rules”toperformlogicalinferencesaboutthestateofaprocessoperationorsomeotheractivityofinterest.IndustrialapplicationsofKBSsystemshavelargelybeenconcernedwithsupervisory,whichhaveincludedthefollowingproblems:comp