计算机科学导论第三版答案Ch-18

1CHAPTER18ArtificialIntelligence(SolutionstoOdd-NumberedReviewQuestionsandProblems)ReviewQuestionsQ18-1.Aninterrogatorasksasetofquestionsthatareforwardedtoacomputerandahumanbeing.Theinterrogatorreceivestwosetsofresponses:onefromthecomputerandonefromthehuman.Aftercarefulexaminationofthetwosets,iftheinterrogatorcannotdefinitelytellwhichsethascomefromthecom-puter,thecomputerhaspassedtheintelligenttest.Someexpertsthinkthatthisisanaccuratedefinitionofanintelligentsystem;somethinkthatthetestisnotnecessarilythedefinitionofanintelligentsystem.Q18-3.LISPisaprogramminglanguagethatmanipulateslists.LISPtreatsdata,aswellasaprogram,asalist,.ThismeansaLISPprogramcanchangeitself.Thisfeaturematcheswiththeideaofanintelligentagentthatcanlearnfromtheenvironmentandimprovesitsbehavior.PROLOGisalanguagethatcanbuildadatabaseoffactsandaknowledgebaseofrules.AprograminPRO-LOGcanuselogicalreasoningtoanswerquestionsthatcanbeinferredfromtheknowledgebase.Q18-5.Propositionallogicisalanguagemadeofasetofsentencesthatcanbeusedtodologicalreasoningabouttheworld.Inpropositionallogic,asymbolthatrep-resentsasentenceisatomic;itcannotbebrokentofindsomeinformationaboutitscomponents.Todoso,weneedpredicatelogic,thelogicthatdefinestherelationbetweenthepartsinaproposition.Q18-7.Aruled-basedsystemrepresentsknowledgeusingasetofrulesthatcanbeusedtodeducesomenewfactsfromalready-knownfacts.Thesemanticnet-workisagraphicalrepresentationofentitiesandtheirrelationships.Q18-9.Thefivestagesofimageprocessingareedgedetection,segmentation,findingdepth,findingorientation,andobjectrecognition.Q18-11.Neuralnetworkstrytosimulatethelearningprocessofthehumanbrainusinganetworksofartificialneurons.2ProblemsP18-1.ThesemanticnetworkisshowninFigureP18-1.P18-3.a.¬Rb.¬Sc.(¬R)∨(¬S)d.R∧Se.S→(¬R)f.R→(¬S)g.S↔(¬R)or(¬R)↔Sh.¬[R→(¬S)]P18-5.a.∃x[Fl(x)∧Wh(x)]b.∃x[Fl(x)∧¬Re(x)]c.¬∀x[Fl(x)→Re(x)]d.∃x{Fl(x)∧[Re(x)∨Wh(x)]}e.¬∃x[Fl(x)∧Gr(x)]f.¬∃x[Fl(x)∧Gr(x)]g.∃x[Fl(x)∧¬Wh(x)]FigureP18-1SemanticnetworktoshowrelationshipbetweenprofessionalsProfesionalCareerAccountantMedicalDoctorSubclassSubclassSubclassInternistSubclassSubclassSubclassIsInstanceEngineerDr.PascalFrenchFamilyPractitionerGynocologist3P18-7.a.∀x[Expensive(x)]b.∀x[Cheap(x)]c.∀x[Cheap(x)→Buys(Bob,x)]d.∀x[Expensive(x)→Sells(John,x)]e.¬∀x[Expensive(x)]f.¬∀x[Cheap(x)]g.∃x{[Cheap(x)]→¬Expensive(x)}P18-9.Thetruthtableisshownbelow.Theargument{P→Q,P}|−Qisvalid.P18-11.Thetruthtableisshownbelow.Theargument{P∧Q,P}|−Qisvalid:P18-13.Thedesignoftheneuralnetwork,withweightsw1=w2=0.5andthethresh-oldofT=0.5,isshowninFigureP18-13.PQP→ΟPQFFTFFFTTFTTFFTFTTTTTOKPremisePremiseConclusionPQP∧ΟPQFFFFFFTFFTTFFTFTTTTTOKPremisePremiseConclusionFigureP18-13SimulationofanORgatewithneuralnetworkOutputInputsy(0or1)x1.w2w1=w2=0.5x2.w44Thetruthtableforthisneuralnetworkisshowninbelow.ItisthesameasthetruthtableforanORgate.P18-15.TheheuristicsearchtreeforsolvingthepuzzleisshowninFigureP18-15.InputsS=x1.w1+x2.w2CompareSwithTOutput000ST0010.5S=T1100.5S=T1111ST1FigureP18-15TheinitialandgoalstateofapuzzleInitialStateGoalState5432331024676413856724138567241385672413856721348567213485672134856721348567213485672134856721348567213485672134856725P18-17.FigureP18-17showsthedepth-firstsearchforthetree(relatedtoProblemP18-16).P18-19.FigureP18-19showsthebreadth-firstsearchforthemaze(ProblemP18-18).FigureP18-17Depth-firstsearchforthetreeFigureP18-19Breadth-firstsearchforthemazeAFBEDGKJIHCLStartFinishABCDEFGIHLJMKNOPQ