人工智能与专家系统 外文文献译文和原文.docx
- 文档编号:9409167
- 上传时间:2023-02-04
- 格式:DOCX
- 页数:17
- 大小:45.77KB
人工智能与专家系统 外文文献译文和原文.docx
《人工智能与专家系统 外文文献译文和原文.docx》由会员分享,可在线阅读,更多相关《人工智能与专家系统 外文文献译文和原文.docx(17页珍藏版)》请在冰豆网上搜索。
人工智能与专家系统外文文献译文和原文
毕业设计(论文)外文翻译
(全日制本科生)
课题名称基于正则表达式的汉语特殊句式
元素抽取与处理
翻译论文名称人工智能与专家系统
专业、班级
学生
学号
指导教师
二o一3年4月
ARTIFICIALINTELLIGENCEANDEXPERTSYSTEM
1.HistoryofAl
TheseedofAlweresownonlytwoyearsafterGeneralElectricinstalledthefirstconputerforbusinessuse.Theyearwas1956,andthetermartificialintelligence(Al)wascoinedbyjohnMcCarthyasthethemeofaconferenceheldatDartmouthCollege.Thatsameyear,thefirstAlconputerprogram,calledLogicTheoristwasannounced.LogicTheorist'slimitedabilitytothereason(provingcalculustheorems)encourageresearcherstodevelopanotherprogramcalledtheGeneralProblemSolver(GPS),whichwasintendedtosolveproblemsofallkinds.Thetaskturnedouttobemorethentheearlypioneerscouldhandle.
Alresearchcontinued,butittookbackseattothelessambitiousconputerapplicationssuchasMISandDSS.Overtime,however,persistentresearchcontinuedtopushbackthefrontiersofusingthecomputerfortasksthatnormallyrequirehumanintelligence.
2.AreasofAl
Aliscurrentlybeingappliedinbusinessintheformofknowledgesystems,whichusehumanknowledgetosolveproblems.Themostpopulartypeofknowledge-basedsystemistheexpertsystem.AnexpertsystemisacomputerprogramthatatterrptstorepresenttheknowledgeofhumanexpertintheformofheuristicsisderivedfromthesameGreekrootasthewordeureka,whichmeans"todiscover".Aheuristicis,therefore,aruleofgoodguessing.
HeuristicsdonotguaranteeresultsasabsolutelyasdoconventionalalgorithmsthatareincorporatedintoDSSs,buttheyofferresultsthatarespecificenoughmostofthetimetobeuseful.Theheuristicsallowtheexpertsystemtofunctioninamannerconsistentwithahumane冲ert,advisingtheuseronhowtosolveaproblemSincetheexpertsystemfunctionsasaconsultant,theactofusingitiscalledaconsultation—theuserconsultstheexpertsystemforadvice.
Inadditiontoexpertsystem,Alincludesworkinthefollowingareas:
neuralnetworks,perceptivesystems,learning,robotics,Alhardware,andnaturallanguageprocessing.Theseareasareillustratedthewaythatoneareacanbenefittheothers.
3.TheAppealofExpertSystem
Theconceptofe冲ertsystemisbasedontheassunptionthatane冲ert'sknowledgecanbecapturedinconputerstorageandthenappliedbyotherswhentheneedarises.
AnexpertsystemoffersuniquecapabilitiesasadecisionssupportsystemFirst,anexpertsystemoffertheopportunitytomakedecisionsthatexceedthemanager5scapabilities.Forexample,anewinvestmentofficerforabankcanuseanexpertsystemdesignedbyaleadingfinancialexpertand,indoingso,incorporatetheexpert'sknowledgeintohisorreachingaparticularsolution.Veryoften,theexplanationofhowasolutionwasreachedismorevaluablethanthesolutionitself
4.AnExpertSystemModel
Themodelofanexpertsystemconsistsoffourmainparts.Theknowledgebasehousestheaccumulatedknowledgeoftheparticularproblemtobesolved.Theinferenceengineprovidesthereasoningabilitythatinterpretsthecontentsoftheknowledgebase.Theexpertandtheknowledgeengineerusethedevelopmentenginetocreatetheexpertsystem
LTheUserinterface
Theuserinterfaceenablesthemanagertoenterinstructionsandinformationintotheexpertsystemandtoreceiveinformationfromit.Theinstructionsspecifytheparametersthatguidetheexpertsystemthroughitsreasoningprocessing.Theinformationisintheformofvaluesassignedtocertainvariables.
(1)ExpertSystemInputs
Themostpopularinterfeceformattodayisthegraphicaluserinterface,whichfeaturesaWindowslook.Somesystemsemployacustominterfecetailoredtotheproblembeingsolved.Forexample,thescreenmightdisplayadrawingofamechicalassembly.
(2)ExpertSystemoutputs
Expertsystemaredesignedtorecommendsolutions.Thesesolutionsaresupplementedbyexplanations.Therearetwotypesofexplanation:
Expertsystemaredesignedtorecommendsolutions.Thesesolutionsaresupplementedbye邳lanationswhiletheexpertsystemperformsitsreasoning.PerhapstheexpertsystemwillpronptthemanagertoentersomeinformatioiLThemanageraskswhytheinformationisneeded.Theexpertsystemprovidesane冲lanation
ExplanationoftheproblemsolutionAftertheexpertsystemprovidesaproblemsolution,themanagercanaskforanexplanationofhowitwasreached.Theexpertsystemwilldisplayeachofthereasoningstepsleadingtothesolution
Althoughtheinnerworkingoftheexpertsystemcanbecorrplex,theuserinterfeceisuser-friendlyAmanageraccustomedtointeractingwithaconputershouldhavenodifficultyinusinganexpertsystem
2.TheKnowledgebase
Theknowledgebasecontainsbothfactsthatdescribetheproblemareaandknowledgerepresentationtechniquesthatdescribehowthefectsfittogetherinalogicalmanager.Thetermproblemdomainisusedtodescribetheproblemarea.
(1)Rules
Apopularknowledgerepresentationtechniqueistheuseofrulesspecifieswhattodoinagivensituationtechniqueistheuseofrules.Arulespecifieswhattodoinagivensituationandconsistsoftwoparts:
aconditionthatmayormaynotbetrueandanactiontobetakenwhentheconditionistrue.Anexanpleofaruleis:
IFECONOMIC.INDEX>1.20ANDSEASONAL.INDEX>1.30
THENSALES.OUTLOOK=”EXCELLENT”
Alloftherulescontainedinanexpertsystemarecalledtheruleset.Therulesetcanvaryfromadozenofrules.Adozenofrulesforasimpleexpertsystem,and500,1,000,or10,000rulesforaco叫lexone.
(2)NetworkofRules
Therulesofarolesetarenotphysicallylinked,buttheirlogicalrelationshipscanbeillustratedwithahierarchicaldiagram.Therulesatthebottomofthehierarchyprovideevidencefortherulesontheupperlevels.Theevidenceenablestherulesontheupperlevelstoproduceconclusions.
Thetoplevelmightconsistofasingleconclusion,indicatingthattheproblemhasonlyasinglesolutionThetermgoalvariableisusedtodescribethesolution,whichcanbeaconputedvalue,anactiontobetaken,orsomeotherrecommendationForexample,ifanexpertsystemistoadvisetop-levelmanagementonwhethertoenteranewmarketarea,avalueofYesorNotwouldbeassignedtothesingle-goalvariableMARKETDECISION.
Itisalsopossibleforthetoplevelofthehierarchytoincludemultipleconclusions,indictingthepossibilityofmorethanonesolutionAnexampleisanexpertsystemthatmakesrecommendationsconcerningthebeststrategytofollowinreactingtoincreasedcompetitiveactivityThesystemmightselectfromamongpossiblestrategiesofimprovingthequalityofthefirm'sproducts,investingmoreinadvertising,orloweringprices.
3.TheInferenceEngine
Theinferenceengineistheportionoftheexpertsystemthatperformsreasoningbyusingthecontentsoftheknowledgebaseinaparticularsequence.
Duringtheconsultation,theinferenceengineexaminestherulesoftheknowledgebaseoneatatime,andwhenarule'sconditionistrue,thespecifiedactionistakenInexpertsystemsterminologytheruleis"fired"whentheactionistaken
Twomainmethodshavebeendevisedfortheinferenceenginetouseinexaminingtherules:
forwardreasoningandreversereasoning.
(1)Forwardreasoning
Inforwardreasoning,alsocalledforwardchaining,therulesareexaminedoneafteranotherinacertainorder.Theordermightbethesequenceinwhichtheruleswereenteredintotheruleset,oritmightbesomeothersequencespecifiedbytheuser.Aseachruleisexamined,theexpertsystemattenptstoevaluatewhethertheconditionstrueorfelse.
RULEEVALUSTION.Whentheconditionistrue,theruleisfiredandthenextruleisexamined.Whentheconditionisfalse,theruleisnotfiredthenextruleisexamined.Itispossiblethatarulecannotbeevaluatedastrueorfelse.Perhapstheconditionincludesoneormorevariableswithunknownvalues.Inthatcase,theruleconditionisunknownWhenaroleconditionisunknown,theruleisnotfiredandthenextruleisexamined.
THEITERAIIVEREASONINGPROCESS.TheprocessofexaminingoneruleaftertheothercontinuesuntilaconpletepasshasbeenmadethroughtheentirerulesetMorethanonepassusuallyisnecessarytoassignavaluetothegoalvariable.PerhapstheinformationneededtoevaluateoneruleisproducedbyanotherrulethatisexaminedsubsequentlyForexanple,aftertheeleventhruleisfired,thefifthrulecanbeevaluatedonthenextpass.Thepassescontinueaslongasitispossibletofirerules.Whennomorerulescanbefired,thereasoningprocessceases.
(2)ReverseReasoning
Inreversereasoning,alsocalledbackwardchaining,theinferenceengineselectsaruleandregardsitasaproblemtobesolved.Usingtherulesetasshowninfigure20-1.Rule12istheproblem,sinceitassignsavaluetothegoalvariableP.TheinferenceengineatterrptstoevaluateRule12butrecognizesthatRule10orRule11mustbeevaluatedfirstRule10and11becomesubproblemsofRule12.Theinferenceenginethenselectsoneofthesubproblemstoevaluate,andtheselectedsubproblembecomesthenewproblem
THEFIRSTLOGCALPATHISPURSUED.WewillassumethatRule10becomestheproblemTheinferenceenginethendeterminesthatRule7and8mustbeevaluatedbeforeRule10canbeevaluated.Rules7and8becomethesubproblemsinthismanner;searchingforarulethatcanbeevaluated.
THENEXTLOGICALPATHISPURSUED.Whenthee冲ertsystemattenptstoevaluateRule11,Rule9becomestheproblem;itcanbeevaluatedusingtheoutcomesofRules4and5.BecausebothRules4and5aretrue,Rule9canbeevaluatedastruewithout
- 配套讲稿:
如PPT文件的首页显示word图标,表示该PPT已包含配套word讲稿。双击word图标可打开word文档。
- 特殊限制:
部分文档作品中含有的国旗、国徽等图片,仅作为作品整体效果示例展示,禁止商用。设计者仅对作品中独创性部分享有著作权。
- 关 键 词:
- 人工智能与专家系统 外文文献译文和原文 人工智能 专家系统 外文 文献 译文 原文