Abstract: This paper summarizes recent activities at LIMSI in multilingual speech recognition and its applications. While the main goal of speech recognition is to provide a transcription of the speech signal as a sequence of words, the same basic technolo
RecentActivitiesinSpokenLanguageProcessingatLIMSI
Jean-LucGauvainandLoriLamel
SpokenLanguageProcessingGroup,LIMSI-CNRS,91403Orsay,FRANCE
gauvain,lamel@limsi.frdoc.guandang.net
Abstract:ThispapersummarizesrecentactivitiesatLIMSIinmultilingualspeechrecognitionanditsapplications.Whilethemaingoalofspeechrecognitionistoprovideatranscriptionofthespeechsignalasasequenceofwords,thesamebasictechnologyservesasthe rststepinotherapplicationareas,suchasinautomaticsystemsforinformationaccessandforautomaticindexationofaudiovisualdata.
SPEECHRECOGNITION
Speechrecognitionisprincipallyconcernedwiththeproblemoftranscribingthespeechsignalasasequenceofwords.TheLIMSIsystem,incommonwithmostoftoday’sstate-of-the-artsystems(4),makesuseofstatisticalmodelsofspeechgeneration.Fromthispointofview,messagegenerationisrepresentedbyalanguagemodelwhichprovidesanestimateoftheprobabilityofanygivenwordstring,andtheencodingofthemessageintheacousticsignalisrepresentedbyaprobabilitydensityfunction(HMM).Thespeechdecodingproblemthenconsistsofmaximizingtheaposterioriprobabilityofthewordstringgiventheobservedacousticsignal.
TheLIMSIwordrecognizer(makesuseofcontinuousdensityHMMsforacousticmodelingofcontextualphoneunitsandn-gramstatisticsestimatedontrainingtexts(newspapertextsand/orspeechtranscriptions)forlanguagemodeling.AcousticmodelingusesPLP-likecepstralfeaturesderivedfromaMelfrequencyspectrum,withsentence-basedcepstralnormalization.Foreachtasktherecognitionvocabularyisselectedtomaximizelexicalcoverage,thusminimizingerrorsduetounknownwords.Lexicalpronunciationsarephonemically-based,withlanguage-speci csymbols(7).Speci cunitsareusedtomodelnon-speecheffectssuchassilence, llerwords,andbreathnoises.Fortranscriptiontasks,wordrecognitioniscarriedoutinmultipledecodingpasses,wheretheinformationbetweenlevelsistransmittedviawordgraphs(5).Inthe rstpassawordgraphisgeneratedusingasmallbigramlanguagemodel,whichisthenusedtoconstrainthesearchspaceinfurtherdecodingpasseswithmoreaccurateacousticandlanguagemodels.Unsupervisedacousticmodeladaptationmayoptionallybeperformedtoimproveperformance.TheapplicabilityofspeechrecognitiontechniquesfordifferentlanguagesisofparticularimportanceinEurope.AtLIMSI,speaker-independent,largevocabulary,continuousspeechrecognitionsystemsfordifferentEuropeanlan-guages(French,GermanandBritishEnglish)andforAmericanEnglishhavebeendeveloped(9).ForAmericanEnglishworderrorratesofabout8%canbeachievedforunrestrictednewspapertextsfromunknownspeaker.WorderrorsonasimilartaskinFrenchareabout11%(1).Thewidelyusedn-gramlanguagemodelsarereasonablysuc-cessfulfordictationinEnglish,butarelessef cientformorehighlyin ectedlanguagessuchasFrenchandGerman.Applicationsofthistechnologyaremainlyinprofessionaldomains,wherethesystemshandleaspokenencodingofwrittenlanguage.
TRANSCRIPTIONOFRADIOANDTELEVISIONBROADCASTS
Recentadvancesintheunderlyingspeechrecognitiontechnologyhaveledtoactivityusingreal-world(or“found”)speechdataincontrasttospeechproducedwiththeaimofbeingrecognizedbyamachine.Oneparticularlyexcitingareaofresearchisthetranscriptionoftelevisionandradiobroadcastsforwhichautomaticprocessingcaneasetheworkloadrequiredforindexationandretrievalpurposes(2).Thistaskischallengingfromthetechnicalview,astheshowscontaindataofvariousacousticandlinguisticnature,withabruptorgradualtransitionsbetweensegments.Thesignaltypicallyconsistsofpreparedspeechandspontaneousspeech,whichmaybeofstudioqualityorhavebeentransmittedoveratelephoneorothernoisychannel(ie.,corruptedbyadditivenoiseandnonlineardistorsions),aswellasspeechovermusicandpuremusicsegments.Acousticmodelstrainedoncleanspeechareclearlyinadequatetoprocesssuchinhomogeneousdata.Thespeechisproducedbyawidevarietyofspeakers:newsanchorsandtalkshowhosts,reportersinremotelocations,interviewswithpoliticiansandcommonpeople,unknownspeakers,newdialects,non-nativespeakers,etc.
Inadditiontoproblemsassociatedwithspeechtranscription,thecontinuousstreamofdataneedstobepartitionedintohomogeneoussegmentsinordertoachieveacceptableperformance(6).Fromrecentresearchitappearsthatcurrentwordrecognitionperformanceisnotcriticallydependentuponthepartitioningaccuracyandthatanyreason-ableapproachthatseparatesspeakerturnsandmajoracousticboundariesissuf cient.Inordertoaddressvariability
NEW DEVELOPMENTS IN THE INRS CONTINUOUS SPEECH RECOGNITION SYSTEM New techniques are developed for the second pass search in our large vocabulary continuous ...
Continuous Speech Recognition in Radiology Reporting Vinod Aggarwal, M.D., Venkatesh Rudrapatna, M.B.,B.S., Sridhar Rajappan, M.B.,B.S., Bharat ...
Continuous speech recognition in the WAXHOLM dialogue system Nikko Strm This paper presents the status of the continuous speech recognition engine of the ...
AUDIO-VISUAL LARGE VOCABULARY CONTINUOUS SPEECH RECOGNITION IN THE BROADCAST DOMAIN S. Basu, C. Neti, N. Rajput, A. Senior, L. Subramaniam, A. ...
Large vocabulary, continuous speech recognition is only at the beginning of development for Vietnamese. The equivalent task is considered challenging even in ...
sentence (PD) set, which are parts of the text prompt for speech recording in Large Vocabulary Continuous Speech Recognition (LVCSR) corpus for Thai ...
Generation and Minimization of Word Graphs in Continuous Speech Recognition The ability of an automatic speech recognition system (ASR) to output multiple ...
Eigenvalues Driven Gaussian Selection in continuous speech recognition 隐藏>> Appl Intell (2010) 33: 107–116 DOI 10.1007/s10489-008-0152-9 Eigenvalues...
The Effectiveness of Oral Expression through the use of Continuous Speech Recognition Technology in Supporting the Written Composition of Postsecondary Students...
cial Neural Network (HMM/ANN) Large Vocabulary Continuous Speech Recognition (LVCSR) system [8]. Experiments were carried out for a number of baseform ...
热门文档