ELM-极限学习机-黄广斌学术报告-讲稿

《ELM-极限学习机-黄广斌学术报告-讲稿》由会员上传分享，免费在线阅读，更多相关内容在教育资源-天天文库。

1、>>ZhengyouZhang:Okay.Soit'smypleasuretointroduceProfessorGuangBinHuangfromNanyangTechnologicalUniversityofSingapore,NTU.Soheistalkingabouthislonginterestresearchinextremelearningmachine.HegraduatedfromNortheasternUniversityinChinainappliedmathematicsandhasbeenwithNTUforquiteafewyears.Andhe'sass

2、ociateeditorofNeurocomputingandalsoIEEETransactionsonSystems,ManandCybernetics.Andhe'sorganizingourworkshop.Youcanadvertiseyourconferencelater.Soplease.>>GuangBinHuang:Thanks,Dr.Zhang,forinvitingmehere.It'smyhonortogiveatalktointroducetheextremelearningmachine.Thisisactually theideawasinitial

3、lyinitiatedin'9 actually2003.Thatwasthefirsttimewesubmittedpapers.Andthenitbecomerecognizedbymoreandmoreresearchersrecently.SowejusthadaworkshopinAustralialastDecember.SowearegoingtohaveaninternationalsymposiumonextremelearningmachinethatiscomingDecemberinChina.It'sDr. Zhang'shometown,ornearby

4、hishometown.Sowe'rehopingyoucanjoin.Okay.Sowhatisextremelearningmachine?Actually,thisisatechnicaltalkabout talkingaboutthekindoflearningmethod.Suchalearningmethodisdifferentfromthetraditionallearningmethod.Sotuningisnotrequired.SoyouwillwishtoknowextremelearningmachineIthinkit'sbettertogobackt

5、oreviewthetraditionalfeedforwardneuralnetworks,includingsupportvectormachines.Iassumemanypeopleconsidersupportvectormachinestonotbeknowntoneuralnetworks.Butactuallyinmyopinion[inaudible]theyhavethesametarget,samearchitectureinsomeaspects,insomesense.Soafterwereviewfeedforwardneuralnetworks,thenw

6、ecangotothe tointroducewhatiscalledextremelearningmachine.Actually,extremelearningmachineisvery,verysimple.Okay.DuringmytalkIalsowishtogivethecomparisonbetweentheyearendandthissquareSVM.SofinallyIwishtoshowlinkagebetweentheELMandalsothetraditionalSVM,soweknowwhat'sthedifference,what'stherelati

7、onship.Okay.Sothefeedforwardneuralnetwork,Ihaveseveraltypesofframeworks,architectures.Oneofthepopularonesismultilayerfeedforwardneuralnetworks.Butthenintheoryandalsoinapplicationspeoplefoundasinglehiddenlayerisenoughforust