Recurrent Neural Networks_Xiaolin Hu, P. Balasubramaniam文档文件分享.pdfVIP

RecurrentNeuralNetworks

RecurrentNeuralNetworks

Editedby

XiaolinHuandP.Balasubramaniam

I-Tech

IV

PublishedbyIn-Teh

In-TehisCroatianbranchofI-TechEducationandPublishingKG,Vienna,Austria.

Abstractingandnon-profituseofthematerialispermittedwithcredittothesource.Statementsand

opinionsexpressedinthechaptersaretheseoftheindividualcontributorsandnotnecessarilythoseof

theeditorsorpublisher.Noresponsibilityisacceptedfortheaccuracyofinformationcontainedinthe

publishedarticles.Publisherassumesnoresponsibilityliabilityforanydamageorinjurytopersonsor

propertyarisingoutoftheuseofanymaterials,instructions,methodsorideascontainedinside.After

thisworkhasbeenpublishedbytheIn-Teh,authorshavetherighttorepublishit,inwholeorpart,in

anypublicationofwhichtheyareanauthororeditor,andthemakeotherpersonaluseofthework.

©2008In-teh

Additionalcopiescanbeobtainedfrom:

publication@

FirstpublishedSeptember2008

PrintedinCroatia

AcataloguerecordforthisbookisavailablefromtheUniversityLibraryRijekaunderno.111225073

RecurrentNeuralNetworks,EditedbyXiaolinHuandP.Balasubramaniam

p.cm.

ISBN978-953-7619-08-4

1.RecurrentNeuralNetworks,XiaolinHuandP.Balasubramaniam

Preface

Theresearchofneuralnetworkshasexperiencedseveralupsanddownsinthe20th

century.ThelastresurgenceisbelievedtobeinitiatedbyseveralseminalworksofHopfield

andTankinthe1980s,andthisupsurgehaspersistedforthreedecades.TheHopfield

neuralnetworks,eitherdiscr