




已閱讀5頁,還剩16頁未讀, 繼續(xù)免費(fèi)閱讀
版權(quán)說明:本文檔由用戶提供并上傳,收益歸屬內(nèi)容提供方,若內(nèi)容存在侵權(quán),請(qǐng)進(jìn)行舉報(bào)或認(rèn)領(lǐng)
文檔簡(jiǎn)介
1Machine-LearningResearchFourCurrentDirectionsThomasG.DietterichMachine-learningresearchhasbeenmakinggreatprogressinmanydirections.Thisarticlesummarizesfourofthesedirectionsanddiscussessomecurrentopenproblems.Thefourdirectionsare(1)theimprovementofclassificationaccuracybylearningensemblesofclassifiers,(2)methodsforscalingupsupervisedlearningalgorithms,(3)reinforcementlearning,and(4)thelearningofcomplexstochasticmodels.Thelastfiveyearshaveseenanexplosioninmachine-learningresearch.Thisexplosionhasmanycauses:First,separateresearchcommunitiesinsymbolicmachinelearning,computationlearningtheory,neuralnetworks,statistics,andpatternrecognitionhavediscoveredoneanotherandbeguntoworktogether.Second,machine-learningtechniquesarebeingappliedtonewkindsofproblem,includingknowledgediscoveryindatabases,languageprocessing,robotcontrol,andcombinatorialoptimization,aswellastomoretraditionalproblemssuchasspeechrecognition,facerecognition,handwritingrecognition,medicaldataanalysis,andgameplaying.Inthisarticle,Iselectedfourtopicswithinmachinelearningwheretherehasbeenalotofrecentactivity.ThepurposeofthearticleistodescribetheresultsintheseareastoabroaderAIaudienceandtosketchsomeoftheopenresearchproblems.Thetopicareasare(1)ensemblesofclassifiers,(2)methodsforscalingupsupervisedlearningalgorithms,(3)reinforcementlearning,and(4)thelearningofcomplexstochasticmodels.Thereadershouldbecautionedthatthisarticleisnotacomprehensivereviewofeachofthesetopics.Rather,mygoalistoprovidearepresentativesampleoftheresearchineachofthesefourareas.Ineachoftheareas,therearemanyotherpapersthatdescriberelevantwork.IapologizetothoseauthorswhoseworkIwasunabletoincludeinthearticle.EnsemblesofClassifiersThefirsttopicconcernsmethodsforimprovingaccuracyinsupervisedlearning.Ibeginbyintroducingsomenotation.Insupervisedlearning,alearningprogramisgiventrainingexamplesoftheform(x1,y1),(xm,ym)forsomeunknownfunctiony=f(x).Thexivaluesaretypicallyvectorsoftheformwhosecomponentsarediscreteorrealvalued,suchasheight,weight,color,andage.ThesearealsocalledthefeatureofXi,IusethenotationXijto.referto2thejthfeatureofXi.Insomesituations,Idroptheisubscriptwhenitisimpliedbythecontext.Theyvaluesaretypicallydrawnfromadiscretesetofclasses1,kinthecaseofclassificationorfromthereallineinthecaseofregression.Inthisarticle,Ifocusprimarilyonclassification.Thetrainingexamplesmightbecorruptedbysomerandomnoise.GivenasetSoftrainingexamples,alearningalgorithmoutputsaclassifier.Theclassifierisahypothesisaboutthetruefunctionf.Givennewxvalues,itpredictsthecorrespondingyvalues.Idenoteclassifiersbyh1,,hi.Anensembleofclassifierisasetofclassifierswhoseindividualdecisionsarecombinedinsomeway(typicallybyweightedorunweightedvoting)toclassifynewexamples.Oneofthemostactiveareasofresearchinsupervisedlearninghasbeenthestudyofmethodsforconstructinggoodensemblesofclassifiers.Themaindiscoveryisthatensemblesareoftenmuchmoreaccuratethantheindividualclassifiersthatmakethemup.Anensemblecanbeemoreaccuratethanitscomponentclassifiersonlyiftheindividualclassifiersdisagreewithoneanother(HansenandSalamon1990).Toseewhy,imaginethatwehaveanensembleofthreeclassifiers:h1,h2,h3,andconsideranewcasex.Ifthethreeclassifiersareidentical,thenwhenh1(x)iswrong,h2(x)andh3(x)arealsowrong.However,iftheerrorsmadebytheclassifiersareuncorrelated,thenwhenh1(x)iswrong,h2(x)andh3(x)mightbecorrect,sothatamajorityvotecorrectlyclassifiesx.Moreprecisely,iftheerrorratesofLhypotheseshiareallequaltopL/2andiftheerrorsareindependent,thentheprobabilitythatbinomialdistributionwheremorethanL/2hypothesesarewrong.Figure1showsthisareaforasimulatedensembleof21hypotheses,eachhavinganerrorrateof0.3.Theareaunderthecurvefor11ormorehypothesesbeingsimultaneouslywrongis0.026,whichismuchlessthantheerrorrateoftheindividualhypotheses.Ofcourse,iftheindividualhypothesesmakeuncorrelatederrorsatratesexceeding0.5,thentheerrorrateofthevotedensembleincreasesasaresultofthevoting.Hence,thekeytosuccessfulensemblemethodsistoconstructindividualclassifierswitherrorratesbelow0.5whoseerrorsareatleastsomewhatuncorrelated.MethodsforConstructingEnsemblesManymethodsforconstructingensembleshavebeendeveloped.Somemethodsaregeneral,andtheycanbeappliedtoanylearningalgorithm.Othermethodsarespecifictoparticularalgorithms.Ibeginbyreviewingthegeneraltechniques.SubsamplingtheTrainingExamplesThefirstmethodmanipulatesthetrainingexamplestogeneratemultiple3hypotheses.Thelearningalgorithmisrunseveraltimes,eachtimewithadifferentsubsetofthetrainingexamples.Thistechniqueworksespeciallywellforunstablelearningalgorithms-algorithmswhoseoutputclassifierundergoesmajorchangesinresponsetosmallchangesinthetrainingdata.Decisiontree,neuralnetwork,andrule-learningalgorithmsareallunstable.Linear-regression,nearest-neighbor,andlinear-thresholdalgorithmsaregenerallystable.Themoststraightforwardwayofmanipulatingthetrainingsetiscalledbagging.Oneachrun,baggingpresentsthelearningalgorithmwithatrainingsetthatconsistofasampleofmtrainingexamplesdrawnrandomlywithreplacementfromtheoriginaltrainingsetofmitems.Suchatrainingsetiscalledabootstrapreplicateoftheoriginaltrainingset,andthetechniqueiscalledbootstrapaggregation(Breiman1996a).Eachbootstrapreplicatecontains,ontheaverage,63.2percentoftheoriginalset,withseveraltrainingexamplesappearingmultipletimes.Anothertraining-setsamplingmethodistoconstructthetrainingsetsbyleavingoutdisjointsubsets.Then,10overlappingtrainingsetscanbedividedrandomlyinto10disjointsubsets.Then,10overlappingtrainingsetscanbeconstructedbydroppingoutadifferentisusedtoconstructtrainingsetsfortenfoldcross-validation;so,ensemblesconstructedinthiswayaresometimescalledcross-validatedcommittees(Parmanto,Munro,andDoyle1996).ThethirdmethodformanipulatingthetrainingsetisillustratedbytheADABOOSTalgorithm,developedbyFreundandSchapire(1996,1995)andshowninfigure2.Likebagging,ADABOOSTmanipulatesthetrainingexamplestogeneratemultiplehypotheses.ADABOOSTmaintainsaprobabilitydistributionpi(x)overthetrainingexamples.Ineachiterationi,itdrawsatrainingsetofsizembysamplingwithreplacementaccordingtotheprobabilitydistributionpi(x).Thelearningalgorithmisthenappliedtoproduceaclassifierhi.Theerrorrateiofthisclassifieronthetrainingexamples(weightedaccordingtopi(x)iscomputedandusedtoadjusttheprobabilitydistributiononthetrainingexamples.(Infigure2,notethattheprobabilitydistributionisobtainedbynormalizingasetofweightswi(i)overthetrainingexamples.)Theeffectofthechangeinweightsistoplacemoreweightonexamplesthatweremisclassifiedbyhiandlessweightonexamplesthatwerecorrectlyclassified.Insubsequentiterations,therefore,ADABOOSTconstructsprogressivelymoredifficultlearningproblems.Thefinalclassifier,hiisconstructsbyaweightedvoteoftheindividualclassifiers.Eachclassifierisweightedaccordingtoitsaccuracyforthedistributionpithatitwastrainedon.Inline4oftheADABOOSTalgorithm(figure2),thebaselearningalgorithmLearniscalledwiththeprobabilitydistributionpi.IfthelearningalgorithmLearncanusethisprobabilitydistributiondirectly,4thenthisproceduregenerallygivesbetterresults.Forexample,Quinlan(1996)developedaversionofthedecisiontree-learningprogramc4.5thatworkswithaweightedtrainingsample.Hisexperimentsshowedthatitworkedextremelywell.Onecanalsoimagineversionsofbackpropagationthatscaledthecomputedoutputerrorfortrainingexample(Xi,Yi)bytheweightpi(i).Errorsforimportanttrainingexampleswouldcauselargergradient-descentstepsthanerrorsforunimportant(low-weight)examples.However,ifthealgorithmcannotusetheprobabilitydistributionpidirectly,thenatrainingsamplecanbeconstructedbydrawingarandomsamplewithreplacementinproportiontotheprobabilitiespi.ThisproceduremakesADABOOSTmorestochastic,butexperimentshaveshownthatitisstilleffective.Figure3comparestheperformanceofc4.5toc4.5withADABOOST.M1(usingrandomsampling).Onepointisplottedforeachof27testdomainstakenfromtheIrvinerepositoryofmachine-learningdatabases(MerzandMurphy1996).Wecanseethatmostpointslieabovetheliney=x,whichindicatesthattheerrorrateofADABOOSTislessthantheerrorrateofc4.5.Figure4comparestheperformanceofbagging(withc4.5)toc4.5alone.Again,weseethatbaggingproducessizablereductionsintheerrorrateofc4.5formanyproblems.Finally,figure5comparesbaggingwithboosting(bothusingc4.5astheunderlyingalgorithm).Theresultsshowthatthetwotechniquesarecomparable,althoughboostingappearstostillhaveanadvantageoverbagging.Wecanseethatmostpointslieabovetheliney=x,whichindicatesthattheerrorrateofADABOOSTislessthantheerrorrateofc4.5.Figure4comparestheperformanceofbagging(withc4.5)toc4.5alone.Again,weseethatbaggingproducessizabler
溫馨提示
- 1. 本站所有資源如無特殊說明,都需要本地電腦安裝OFFICE2007和PDF閱讀器。圖紙軟件為CAD,CAXA,PROE,UG,SolidWorks等.壓縮文件請(qǐng)下載最新的WinRAR軟件解壓。
- 2. 本站的文檔不包含任何第三方提供的附件圖紙等,如果需要附件,請(qǐng)聯(lián)系上傳者。文件的所有權(quán)益歸上傳用戶所有。
- 3. 本站RAR壓縮包中若帶圖紙,網(wǎng)頁內(nèi)容里面會(huì)有圖紙預(yù)覽,若沒有圖紙預(yù)覽就沒有圖紙。
- 4. 未經(jīng)權(quán)益所有人同意不得將文件中的內(nèi)容挪作商業(yè)或盈利用途。
- 5. 人人文庫網(wǎng)僅提供信息存儲(chǔ)空間,僅對(duì)用戶上傳內(nèi)容的表現(xiàn)方式做保護(hù)處理,對(duì)用戶上傳分享的文檔內(nèi)容本身不做任何修改或編輯,并不能對(duì)任何下載內(nèi)容負(fù)責(zé)。
- 6. 下載文件中如有侵權(quán)或不適當(dāng)內(nèi)容,請(qǐng)與我們聯(lián)系,我們立即糾正。
- 7. 本站不保證下載資源的準(zhǔn)確性、安全性和完整性, 同時(shí)也不承擔(dān)用戶因使用這些下載資源對(duì)自己和他人造成任何形式的傷害或損失。
最新文檔
- 攝影器材品牌區(qū)域代理權(quán)合同
- 材料疲勞損傷累積分析模型合同
- 邊疆地區(qū)古代神話與服飾關(guān)系考古合同
- 保險(xiǎn)業(yè)務(wù)流程重組合同
- 主題餐廳廚房承包及特色氛圍營(yíng)造合同
- 拆遷工程后期維護(hù)承包合同
- 測(cè)量工作總結(jié)200字
- 光伏電站安全工作總結(jié)及計(jì)劃
- 美術(shù)6分鐘技能展示課件
- 防火安全重于泰山
- 游戲策劃師招聘筆試題與參考答案2025年
- 安全、環(huán)境、職業(yè)健康安全目標(biāo)、指標(biāo)及管理方案
- 課件:《中華民族共同體概論》第一講 中華民族共同體基礎(chǔ)理論
- 2024年檔案知識(shí)競(jìng)賽考試題庫300題(含答案)
- 中國(guó)人民抗日戰(zhàn)爭(zhēng)勝利紀(jì)念日紀(jì)念暨世界反法西斯戰(zhàn)爭(zhēng)勝利課件
- 殯葬禮儀策劃方案
- 行政效能提升路徑研究
- (完整版)無菌醫(yī)療器械耗材生產(chǎn)企業(yè)體系文件-質(zhì)量手冊(cè)模板
- JBT 3300-2024 平衡重式叉車 整機(jī)試驗(yàn)方法(正式版)
- 鉆井及井下作業(yè)井噴事故典型案例
評(píng)論
0/150
提交評(píng)論