《人工智能(复习版)》由会员分享，可在线阅读，更多相关《人工智能(复习版)（13页珍藏版）》请在金锄头文库上搜索。
1、杰的小灵儿整理，版权问题，请勿删除，谢谢!人工智能复习后Ml杰的小灵儿整理一、术语解释1、人工智能 artificial intelligence4、产生式规贝U production rules6、后项链接 backward chaining2、知识knowledge3、专家系统expertsystem5、前向链接forwardchaining7、冲突conflict8、推理弓I擎inferenceengine9、排中律lawoftheexcludedmiddle10、贝叶斯推理Bayesianreasoning11、确定因子 certainty factors14、语言变量 linguis
2、tic variable17、模糊推理 fuzzy inference19、质心技术 centroid technique21、单态模式singleton12、模糊集fuzzyset13、模糊逻辑fuzzylogic15、模糊限制语hedge16模糊规则fuzzyrule18、模糊关联记忆fuzzyassociativememory20、隶属函数membershipfunction22、人工神经网络artificialneuralnetworks23、神经元neuron24、权重weight25、符号激活函数signactivationfunction26、感知器perceptron27、线性
3、分割函数linearseparablefunction28、超平面hyperplane29、向后传送back-propagation30、硬限幅函数hard-limitfunction31、阶跃函数stepfunction32、多层神经网络multilayerneuralnetworks33、隐含层hiddenlayer34、自反馈self-feedback35、循环神经网络recurrentneuralnetwork36、双向相关记忆bidirectionalassociativememory37、监督学习supervisedlearning38、无监督学习unsupervisedlearn
4、ing39、竞争学习competitivelearning40、自组织神经网络self-organizingneuralnetworks41、存储模式对storepatternpairs42、基本记忆fundamentalmemory43、自组织特征映射self-organizingfeaturemap44、进化计算evolutionarycomputation45、遗传算法geneticalgorithms46、染色体 chromosome49、遗传编程 genetic programming51、交叉操作 crossover operator53、模式定理 schema theorem47
5、、基因gene48、进化策略evolutionstrategy50、突变操作mutationoperator52、性能图performancegraph54、秩rank55、链表处理语言linkedlistprocessinglanguage56、知识工程knowledgeengineering57、数据挖掘datamining58、原型prototype59、测试用例testcase60、系统维护systemmaintenance二、简述第1章1、图灵测试内容。（网上总结的）Amanwithouttheknowledgeoftheconditionisthroughaspecialway,w
6、ithamachinetoanswerquestions.Ifinaquitelongtime,hecouldnottellthedifferencebetweentheobjectofcommunicatewithhimispeopleormachines,thenthismachinecanbeconsideredtothinking.ThisisthefamousTuringtest.2、智能的定义。（P2）Intelligenceistheirabilitytounderstandandlearnthings.Intelligenceistheabilitytothinkandunde
7、rstandinsteadofdoingthingsbyinstinctorautomatically.3、人工智能的定义。(P18加点第二段)Artificialintelligenceisasciencethatmakemachinesdothingsthatwouldrequireintelligenceifdonebyhumans.第2章4、什么是产生式规则及其组成？(P26、P262.2)(1) ThesestatementsrepresentedintheIF-THENformarecalledroductionrules.(2) Anyruleconsistsoftwoparts
8、:theIFpart,calledtheantecedent(premiseorcondition)andtheTHENpartcalledtheconsequent(conclusionoraction).ExplanationFacilitiesUserInterface专家系统的基本结构5、描述前向链接和后向链接推理技术。(1) Forwardchainingisthedata-drivenreasoning.Thereasoningstartsfromtheknowndataandproceedsforwardwiththatdata.Eachtimeonlythetopmostrul
9、eisexecuted.Whenfired,theruleaddsanewfactinthedatabase.Anyrulecanbeexecutedonlyonce.Thematch-firecyclestopswhennofurtherrulescanbefired(P37)NlatchFireKuowicdge BaseMatchFireX相片超色f YCZL攫- ACxle 1MatchFreKnowledge Ease上MatchFgDatabas?Knowicdge Dascr a? p C上上一L 思 5fjVCvck 3Cck2(P38)(2) Backwardchaining
10、isthegoal-drivenreasoning.Inbackwardchaining,anexpertsystemhasthegoal(hypotheticalsolution)andtheinferenceengineattemptstofindtheevidencetoproveit.First,theknowledgebaseissearchedtofindrulesthatmighthavethedesiredsolution.SuchrulesmusthavethegoalintheirTHEN(action)parts.IfsucharuleisfoundanditsIF(co
11、ndition)partmatchesdatainthedatabase,thentheruleisfiredandthegoalisproved.(P38)Pass 6Pass 4Knovdedge BasePass 5Knocdge BaseDatabaseRpjqqq-3/f克FDfZ X&S&YL&MfNM加幽用咫Khawledge 的眈Pass 3航通GmJ: YPass 1IY&D ZL&MtN匕民Mf NL&MtNGoaJ; Z5讪Ga/；rDatabaseKnovletige BaseKnoMedge BaseKnovdidge BaseF&DN 丫&去汇yA-XAXDatab
12、ase，&方&万f FA-XSj-GoqJ; K(P39)第三章6、贝叶斯推理、确定因子技术的定义区别及适用情况。(1)pH Ep(e|h：p(H)-p(e|h) p(H_pEj_h1 p(H(P62)s belief. The maximum(3) Certaintyfactorisanumbertomeasuretheexpertvalueofthecertaintyfactoris,say,+1.0(definitelytrue)andtheminimum-1.0(definitelyfalse).Apositionvaluerepresentedadegreeofbeliefandan
13、egativeadegreeofdisbelief.(P743.6中间)(4) Certaintyfactorsareusediftheprobabilitiesarenotknownorcannotbeeasilyobtains.（P84最后一段）7、模糊集、模糊推理、模糊逻辑、隶属函数的定义。Deg1eoAfuzzysetisasetwithfuospsetsdaries.（P91下面）Mem2erFhuzzyinferencecanbedefinedasaprocessofmappingfromagiveninputto1.0,anoutput,usingthetheoryoffuzzy
14、sets.（P1064.6）0.8（3）FuzzyhorgicwasintrodAverdgeyJanLukasiewiczTalthe1920sFuzzylogicisase06fmathematicalprinciplesforknowledgerepresentationbasedondegreesofmembershipratherthanonthecrispmembershipofclassicalbinarylogic.Unliketwo-ValuedBooleanlogic,fuzzylogicismulti-valued.（P125力口点第二、三段）02（4）Membershi
15、pfunctionuA（x）equalsthedegreetowhichxisanelementofsetA.（P92英文的第三段中间）8。画图表示模糊集。工50模糊集极示：tal70nan=（0/1180,0.5/185,19090）B图:200P93）210以二4（工）二1 一以6A的模糊操作）DegreeofMembershipHeight, cmFuzzy Sets9、模糊集的操作（补，交，并，包含）。（P98-P100）一、Complement （补）tallmennottallmen、Containment（包含）tallmenverytallmen、Intersection（交）以=min8,H霆Wl二%（乃ch4刈tallmenAaveragemen四、Union（并）=max%（x）,/（x）=以uis（x）tallmenUaveragemenContainmentx
