【8】Deep neural networks
MostcurrentspeechrecognitionsystemsusehiddenMarkovmodels(HMMs)todealwiththetemporalvariabilityofspeechandGaussianmixturemodels(GMMs)todeterminehowwelleachstateofeachHMMfitsaframeorashortwindowofframesofcoefficientsthatrepresentstheacousticinput.Analternativewaytoevaluatethefitis
下载地址
用户评论
