ASurveyofArtificialIntelligence11302010033_马骁烊IntroductionArtificialIntelligence,abbreviationforAIisaresearchanddevelopmentforsimulation,extensionandexpansionofhumanintelligencetheories,techniquesandapplicationsofanewtechnicalsciences.Artificialintelligenceisabranchofcomputersciencethatattemptstounderstandtheessenceofintelligence,andcanproduceanewkindofhumanintelligenceinasimilarmannertorespondtointelligentmachines,thefieldofresearchincludingrobotics,speechrecognition,imagerecognition,naturallanguageprocessingandexpertsystems.In1956theterm“AI”wasfirstraisedatDartmouthInstitute.Sincethen,researchershavedevelopedanumberoftheoriesandprinciplesofartificialintelligence.Ingeneral,amajorgoalofartificialintelligenceresearchistomakeanumberofmachinestocompletecomplextask.ThedevelopmentstagesofAITheriseanddownfallofAIin1950s:Aftertheconceptwasfirstproposed,alotofsignificantresultshavebeenappeared,suchasmachinetheoremproving,checkersprogram,commonproblemsolver,LISPlanguage.Howeverduetolimiteddigestionreasoningability,aswellasthefailureofmachinetranslation,AIhadfallenintothevalley.Feature:emphasisonproblemsolvingapproach,ignoringtheimportanceofknowledge.Thelate1960sand1970s,theexpertsystemhasenabledtheemergenceofnewupsurgeAIresearch.DENDRALchemicalmassspectrometrysystems,MYCINdiseasediagnosisandtherapysystems,Hearsay-IIspeechunderstandingsystems.TheexpertsystemleadstheAItopractice.And,TheInternationalJointConferencesonArtificialIntelligencewasestablishedin1969.1980s,withthesuccessoffifthgenerationcomputer,AIhasbeengreatdevelopment.Japan'sbegana"fifth-generationcomputerplan,"namely"KnowledgeinformationprocessingcomputersystemKIPS",itspurposeistomakethespeedofreasoningreachnumericalcomputation.Althoughthisplanseventuallyfailed,butitmadethetopicofAIhotagain.Theneuralnetworkwasrapidlydevelopmentinthelate1980s.USAhelditsfirstinternationalconferenceonneuralnetworkin1987,andthisannouncedthebirthofthisnewdiscipline.Sincethen,theinvestmentintheneuralnetworkgraduallyincreased