人工智能英語(yǔ)教程參考試卷

參考試卷、寫(xiě)出以下單詞的中文意思（每小題0.5分，共10分）1accuracy11customize2actuator12definition3adjust13defuzzification4agent14deployment5algorithm15effector6analogy16entity7attribute17extract8backtrack18feedback9blockchain19finite10cluster20framework二、根據(jù)給出的中文意思，寫(xiě)出英文單詞（每小題0.5分，共10分）1V.收集，搜集11n.神經(jīng)元；神經(jīng)細(xì)胞2adj.嵌入的，內(nèi)置的12n.節(jié)點(diǎn)3n.指示器；指標(biāo)13V.運(yùn)轉(zhuǎn)；操作4n.基礎(chǔ)設(shè)施，基礎(chǔ)架構(gòu)14n.模式5V.合并；集成15V.察覺(jué)，發(fā)覺(jué)6n.解釋器，解釋程序16n.前提7n.迭代；循環(huán)17adj.程序的；過(guò)程的8n.庫(kù)18n.回歸9n.元數(shù)據(jù)19adj.健壯的，強(qiáng)健的；結(jié)實(shí)的10V.監(jiān)視；控制；監(jiān)測(cè)20V.篩選三、根據(jù)給出的短語(yǔ)，寫(xiě)出中文意思（每小題I分，共10分）dataobjectcybersecuritysmartmanufacturingclusteredsystemdatavisualizationopensourceanalyzetextcloudcomputingcomputationpowerobjectrecognition四、根據(jù)給出的中文意思，寫(xiě)出英文短語(yǔ)（每小題1分，共10分）數(shù)據(jù)結(jié)構(gòu) 決策樹(shù) 演繹推理 4貪婪最佳優(yōu)先搜索隱臧模式，隱含模式知識(shí)挖掘 邏輯推理預(yù)測(cè)性維護(hù)搜索引擎 文本挖掘技術(shù)五、寫(xiě)出以下縮略語(yǔ)的完整形式和中文意思（每小題1分，共10分）縮略語(yǔ) 完整形式 中文意思 ANNARBFSCVDFSESIAKNNNLPVR六、閱讀短文，回答問(wèn)題(每小題2分，共10分)ArtificialNeuralNetwork(ANN)Anartificialneuralnetwork(ANN)isthepieceofacomputingsystemdesignedtosimulatethewaythehumanbrainanalyzesandprocessesinformation.Itisthefoundationofartificialintelligence(AI)andsolvesproblemsthatwouldproveimpossibleordifficultbyhumanorstatisticalstandards.ANNshaveself-learningcapabilitiesthatenablethemtoproducebetterresultsasmoredatabecomesavailable.Artificialneuralnetworksarebuiltlikethehumanbrain,withneuronnodesinterconnectedlikeaweb.Thehumanbrainhashundredsofbillionsofcellscalledneurons.Eachneuronismadeupofacellbodythatisresponsibleforprocessinginformationbycarryinginformationtowards(inputs)andaway(outputs)fromthebrain.AnANNhashundredsorthousandsofartificialneuronscalledprocessingunits,whichareinterconnectedbynodes.Theseprocessingunitsaremadeupofinputandoutputunits.Theinputunitsreceivevariousformsandstructuresofinformationbasedonaninternalweightingsystem,andtheneuralnetworkattemptstolearnabouttheinformationpresentedtoproduceoneoutputreport.Justlikehumansneedrulesandguidelinestocomeupwitharesultoroutput,ANNsalsouseasetoflearningrulescalledbackpropagation,anabbreviationforbackwardpropagationoferror,toperfecttheiroutputresults.AnANNinitiallygoesthroughatrainingphasewhereitlearnstorecognizepatternsindata,whethervisually,aurally,ortextually.Duringthissupervisedphase,thenetworkcomparesitsactualoutputproducedwithwhatitwasmeanttoproduce—thedesiredoutput.Thedifferencebetweenbothoutcomesisadjustedusingbackpropagation.Thismeansthatthenetworkworksbackward,goingfromtheoutputunittotheinputunitstoadjusttheweightofitsconnectionsbetweentheunitsuntilthedifferencebetweentheactualanddesiredoutcomeproducesthelowestpossibleerror.Aneuralnetworkmaycontainthefollowing3layers:Inputlayer-Theactivityoftheinputunitsrepresentstherawinformationthatcanfeedintothenetwork.Hiddenlayer-Todeterminetheactivityofeachhiddenunit.Theactivitiesoftheinputunitsandtheweightsontheconnectionsbetweentheinputandthehiddenunits.Theremaybeoneormorehiddenlayers.Outputlayer-Thebehavioroftheoutputunitsdependsontheactivityofthehiddenunitsandtheweightsbetweenthehiddenandoutputunits.Whatisanartificialneuralnetwork(ANN)?Whatiseachneuronmadeupof?Whadotheinputunitsdo?WhatdoesanANNinitiallygothrough?Howmanylayersmayaneuralnetworkcontain?Whatarethey?七、將下列詞填入適當(dāng)?shù)奈恢?每詞只用一次)。(每小題10分，共20分)填空題1供選擇的答案:transactionsinformationtechniquesfraudnodesunstructuredsubsetsharedautomatedexplosionDeepLearningWhatIsDeepLearning?Deeplearningisanartificialintelligence(AI)functionthatimitatestheworkingsofthehumanbraininprocessingdataandcreatingpatternsforuseindecisionmaking.Deeplearningisa 1 ofmachinelearninginartificialintelligencethathasnetworkscapableoflearningunsupervisedfromdatathatis 2 orunlabeled.Alsoknownasdeepneurallearningordeepneuralnetwork.HowDoesDeepLearningWork?Deeplearninghasevolvedhand-in-handwiththedigitalera,whichhasbroughtaboutan 3 ofdatainallformsandfromeveryregionoftheworld.Thisdata,knownsimplyasbigdata,isdrawnfromsourceslikesocialmedia,internetsearchengines,e-commerceplatforms,andonlinecinemas,amongothers.Thisenormousamountofdataisreadilyaccessibleandcanbe 4 throughfintechapplicationslikecloudcomputing.However,thedata,whichnormallyisunstructured,issovastthatitcouldtakedecadesforhumanstocomprehenditandextractrelevant 5 .CompaniesrealizetheincrediblepotentialthatcanresultfromunravelingthiswealthofinformationandareincreasinglyadaptingtoAIsystemsfor 6 support.DeepLearningvs.MachineLearningOneofthemostcommonAI 7 usedforprocessingbigdataismachinelearning,aself-adaptivealgorithmthatgetsincreasinglybetteranalysisandpatternswithexperienceorwithnewlyaddeddata.Ifadigitalpaymentscompanywantedtodetecttheoccurrenceorpotential 8 initssystem,itcouldemploymachinelearningtoolsforthispurpose.Thecomputationalalgorithmbuiltintoacomputermodelwillprocessall 9 happeningonthedigitalplatform,findpatternsinthedataset,andpointoutanyanomalydetectedbythepattern.Deeplearningutilizesahierarchicallevelofartificialneuralnetworkstocarryouttheprocessofmachinelearning.Theartificialneuralnetworksarebuiltlikethehumanbrain,withneuron 10 connectedtogetherlikeaweb.Whiletraditionalprogramsbuildanalysiswithdatainalinearway,thehierarchicalfunctionofdeeplearningsystemsenablesmachinestoprocessdatawithanonlinearapproach.填空題2供選擇的答案:storedresolutionmatchlookunlockdatabasephotographeyesreturn,identifyingFaceRecognitionFacerecognitionsystemsusecomputeralgorithmstopickoutspecific,distinctivedetailsaboutaperson'sface.Thesedetails,suchasdistancebetweenthe 1 orshapeofthechin,arethenconvertedintoamathematicalrepresentationandcomparedtodataonotherfacescollectedinafacerecognitiondatabase.Thedataaboutaparticularfaceisoftencalledafacetemplateandisdistinctfroma 2 becauseit'sdesignedtoonlyincludecertaindetailsthatcanbeusedtodistinguishonefacefromanother.Somefacerecognitionsystems,insteadofpositively 3 anunknownperson,aredesignedtocalculateaprobabilitymatchscorebetweentheunknownpersonandspecificfacetemplates 4 inthedatabase.Thesesystemswillofferupseveralpotentialmatches,rankedinorderoflikelihoodofcorrectidentification,insteadofjustreturningasingleresult.Facerecognitionsystemsvaryintheirabilitytoidentifypeopleunderchallengingconditionssuchaspoorlighting,lowqualityimage 5 ,andsuboptimalangleofview(suchasinaphotographtakenfromabovelookingdownonanunknownperson).Whenitcomestoerrors,therearetwokeyconceptstounderstand:A"falsenegative^^iswhenthefacerecognitionsystemfailsto 6 matchaperson'sfacetoanimagethatis,infact,containedinadatabase.Inotherwords,thesystemwillerroneously 7 zeroresultsinresponsetoaquery.A"falsepositive^^iswhenthefacerecognitionsystemdoesmatchaperson'sfacetoanimageina 8 ,butthatmatchisactuallyincorrect.Thisiswhenapoliceofficersubmitsanimageof"Joe,"butthesystemerroneouslytellstheofficerthatthephotoisof"Jack.”Whenresearchingafacerecognitionsystem,itisimportantto 9 closelyatthet€falsepositive“rateandthe“falsenegative''rate,sincethereisalmostalwaysatrade-off.Forexample,ifyouareusingfacerecognitionto 10 yourphone,itisbetterifthesystemfailstoidentifyyouafewtimes(falsenegative)thanitisforthesystemtomisidentifyotherpeopleasyouandletsthosepeopleunlockyourphone(falsepositive).Iftheresultofamisidentificationisthataninnocentpersongoestojail(likeamisidentificationinamugshotdatabase),thenthesystemshouldbedesignedtohaveasfewfalsepositivesaspossible.六、將下面兩篇短文翻譯成中文(每小題10分，共20分)短文1DifferencesbetweenStrongAIandWeakAI1.MeaningStrongAIisatheoreticalformofartificialintelligencewhichsupportstheviewthatmachinescanreallydevelophumanintelligenceandconsciousnessinthesamewaythatahumaninconscious.StrongAIreferstoahypotheticalmachinethatexhibitshumancognitiveabilities.WeakAI(alsoknownasnarrowAI),ontheotherhand,isaformofartificialintelligencethatreferstotheuseofadvancedalgorithmstoaccomplishspecificproblemsolvingorreasoningtasksthatdonotencompassthefullrangeofhumancognitiveabilities.2.FunctionalityFunctionsarelimitedinweakAIascomparedtostrongAI.WeakAIdoesnotachieveself-awarenessordemonstrateawiderangeofhumancognitiveabilitiesthatahumanmayhave.WeakAIreferstosystemsthatareprogrammedtoaccomplishawiderangeproblemsbutoperatewithinapre-determinedorpre-definedrangeoffunctions.StrongAI,ontheotherhand,referstomachinesthatexhibithumanintelligence.Theideaistodevelopartificialintelligencetothepointwherehumaninteractwithmachinesthatareconscious,intelligentanddrivenbyemotionsandself-awareness.3.GoalThegoalofweakAlistocreateatechnologythatallowsallowsmachinesandcomputerstotoaccomplishspecificproblemsolvingorreasoningtasksatasignificantlyquickerpacethanahumancan.Butitdoesnotnecessarilyincorporateanyrealworldknowledgeabouttheworldoftheproblemthatisbeingsolved.ThegoalofstrongAIistodevelopartificialintelligencetothepointwhereitcanbeconsideredtruehumanintelligence.StrongAIisatypeofwhichdoesnotexistyetinitstrueform.短文2PatternRecognitionPatternRecognitionisdefinedastheprocessofidentifyingthetrends(globalorlocal)inthegivenpattern.Apatterncanbedefinedasanythingthatfollowsatrendandexhibitssomekindofregularity.Therecognitionofpatternscanbedonephysically,mathematicallyorbytheuseofalgorithms.Whenwetalkaboutpatternrecognitioninmachinelearning,itindicatestheuseofpowerfulalgorithmsforidentifyingtheregularitiesinthegivendata.Patternrecognitioniswidelyusedinthenewagetechnicaldomainslikecomputervision,speechrecognition,facerecognition,etc.Therearetwotypesofpatternrecognitionalgorithmsinmachinelearning.SupervisedAlgorithmsThepatternrecognitioninasupervisedapproachiscalledclassification.Thesealgorithmsuseatwo-stagemethodologyforidentifyingthepatterns.Thefirststageisthedevelopment/constructionofthemodelandthesecondstageinvolvesthepredictionfbrneworunseenobjects.Thekeyfeaturesinvolvingthisconceptarelistedbelow.Classifythegivendataintotwosets—trainingsetandtestingset.TrainthemodelusingasuitablemachinelearningalgorithmsuchasSVM(SupportVectorMachines),decisiontrees,randomforest,etc.Themodelistrainedonthetrainingsetandtestedonthetestingset.Theperformanceofthemodelisevaluatedbasedoncorrectpredictionsmade.UnsupervisedAlgorithmsIncontrasttothesupervisedalgorithmsforpatternmakeuseoftrainingandtestingsets,thesealgorithmsuseagroupbyapproach.Theyobservethepatternsinthedataandgroupthembasedonthesimilarityintheirfeaturessuchasdimensiontomakeaprediction.Let'ssaythatwehaveabasketofdifferentkindsoffruitssuchasapples,oranges,pears,andcherries.Weassumethatwedonotknowthenamesofthefruits.Wekeepthedataasunlabeled.Now,supposeweencounterasituationwheresomeonecomesandtellsustoidentifyanewfruitthatwasaddedtothebasket.Insuchacasewemakeuseofaconceptcalledclustering.Clusteringcombinesorgroupsitemshavingthesamefeatures.Nopreviousknowledgeisavailableforidentifyinganewitem.Theyusemachinelearningalgorithmslikehierarchicalandk-mansclustering.Basedonthefeaturesorpropertiesofthenewobject,itisassignedtoagrouptomakeaprediction.

參考試卷答案\寫(xiě)出以下單詞的中文意思（每小題0.5分，共10分）accuracyn.精確（性），準(zhǔn)確（性）11customizeVt.定制，定做；用戶化2actuatorn.執(zhí)行器12definitionn.定義3adjustV.調(diào)整，調(diào)節(jié)；適應(yīng)；校準(zhǔn)13defuzzificationn.逆模糊化，去模糊化4agentn.實(shí)體：代理14deploymentn.部署5algorithmn.算法15effectorn.效應(yīng)器6analogyn.類推16entityn.實(shí)體7attributen.屬性；性質(zhì)；特征17extractV.提取，提煉8backtrackvi.回溯18feedbackn.反饋9blockchainn.區(qū)塊鏈19finiteadj.有限的：限定的1020n.構(gòu)架；框架；（體系的）clusterv.聚集n.團(tuán)，群，簇framework結(jié)構(gòu)二、根據(jù)給出的中文意思，寫(xiě)出英文單詞（每小題0.5分，共10分）1V.收集，搜集gather11n.神經(jīng)元；神經(jīng)細(xì)胞neuron2adj.嵌入的，內(nèi)置的inbuilt12n.節(jié)點(diǎn)node3n.指示器；指標(biāo)indicator13V.運(yùn)轉(zhuǎn)；操作operate4n.基礎(chǔ)設(shè)施，基礎(chǔ)架構(gòu)infrastructure14n.模式pattern5v.合并：集成integrate15V.察覺(jué)，發(fā)覺(jué)perceive6n.解釋器，解釋程序interpreter16n.前提premise7n.迭代；循環(huán)iteration17adj.程序的：過(guò)程的procedural8n.庫(kù)library18n.回歸regression919adi.健壯的，強(qiáng)健的；n.元數(shù)據(jù)metadatarobust結(jié)實(shí)的10V.監(jiān)視；控制；監(jiān)測(cè)monitor20V.篩選screen三、根據(jù)給出的短語(yǔ)，寫(xiě)出中文意思（每小題1分，共10分）1dataobject數(shù)據(jù)對(duì)象2cybersecurity網(wǎng)絡(luò)安全3smartmanufacturing智能制造4clusteredsystem集群系統(tǒng)5datavisualization數(shù)據(jù)可視化6opensource開(kāi)源7analyzetext分析文本8cloudcomputing云計(jì)算9computationpower計(jì)算能力10objectrecognition物體識(shí)別四、根據(jù)給出的中文意思，寫(xiě)出英文短語(yǔ)（每小題1分，共10分）

1數(shù)據(jù)結(jié)構(gòu)datastructure2決策樹(shù)decisiontree3演繹推理deductivereasoning4貪婪最佳優(yōu)先搜索greedybest-firstsearch5隱藏模式，隱含模式hiddenpattern6知識(shí)挖掘knowledgemining7邏輯推理logicalreasoning8預(yù)測(cè)性維護(hù)predictivemaintenance9搜索引擎searchengine10文本挖掘技術(shù)textminingtechnique五、寫(xiě)出以下縮略語(yǔ)的完整形式和中文意思（每小題1分，共10分）縮略語(yǔ)完整形式中文意思1ANNArtificialNeuralNetwork人工神經(jīng)網(wǎng)絡(luò)2ARAugmentedReality增強(qiáng)現(xiàn)實(shí)3BFSBreadth-FirstSearch寬度優(yōu)先搜索4CVComputerVision計(jì)算機(jī)視覺(jué)5DFSDepth-FirstSearch深度優(yōu)先搜索6ESExpertSystem專家系統(tǒng)7IAIntelligentAgent智能體8KNNK-NearestNeighborK最近鄰算法9NLPNaturalLanguageProcessing自然語(yǔ)言處理10VRVirtualReality虛擬現(xiàn)實(shí)六、閱讀短文，回答問(wèn)題（每小題2分，共10分）Anartificialneuralnetwork(ANN)isthepieceofacomputingsystemdesignedtosimulatethewaythehumanbrainanalyzesandprocessesinformation.Itisthefoundationofartificialintelligence(AI)andsolvesproblemsthatwouldproveimpossibleordifficultbyhumanorstatisticalstandards.Eachneuronismadeupofacellbodythatisresponsibleforprocessinginformationbycarryinginformationtowards(inputs)andaway(outputs)fromthebrain.Theinputunitsreceivevariousformsandstructuresofinformationbasedonaninternalweightingsystem.A