资源描述：
《人工神经网络 毕业论文外文翻译》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。
1、外文资料ArtificialNeuralNetworksArtificialNeuralNetworks-BasicFeaturesComposedofalargenumberofprocessingunitsconnectedbyanonlinear,adaptiveinformationprocessingsystem.Itisthebasisformodernneuroscienceresearchfindingspresented,tryingtosimulatealargeneuralnetworkprocessing,memory,
2、informationprocessingwayofinformation.Artificialneuralnetworkhasfourbasiccharacteristics:(1)non-linearnon-linearrelationshipisthegeneralcharacteristicsofthenaturalworld.Thewisdomofthebrainisanonlinearphenomenon.Artificialneuralactivationorinhibitionintwodifferentstates,thisb
3、ehaviormathematicallyexpressedasalinearrelationship.Thresholdneuronshaveanetworkwithbetterperformance,canimprovefaulttoleranceandstoragecapacity.(2)non-limitationofaneuralnetworkisusuallymoreextensiveneuronalconnectionsmade.Theoverallbehaviorofasystemdependsnotonlyonthechara
4、cteristicsofsingleneurons,andmayprimarilybyinteractionbetweenunits,connectedbythedecision.Byalargenumberofconnectionsbetweenthecellsofnon-simulatedbrainlimitations.Associativememorylimitationsofatypicalexampleofnon.(3)characterizationofartificialneuralnetworkisadaptive,self-
5、organizing,self-learningability.Neuralnetworkscannotonlydealwiththechangesofinformation,butalsoprocessinformationthesametime,nonlineardynamicsystemitselfisalsochanging.Iterativeprocessisfrequentlyusedindescribingtheevolutionofdynamicalsystems.(4)Non-convexityofthedirectionof
6、theevolutionofasystem,undercertainconditions,willdependonaparticularstatefunction.Suchasenergyfunction,anditsextremevaluecorrespondingtothestateofthesystemmorestable.Non-convexityofthisfunctionismorethanoneextremum,thissystemhasmultiplestableequilibrium,whichwillcausethesyst
7、emtotheevolutionofdiversity.Artificialneuralnetwork,neuralprocessingunitcanbeexpressedindifferentobjects,suchasfeatures,letters,concepts,orsomeinterestingabstractpatterns.Thetypeofnetworkprocessingunitisdividedintothreecategories:
inputunits,outputunitsandhiddenunits.Inputun
8、itreceivingthesignalanddataoutsideworld;outputunitforprocessingtheresultsto
