 |
|
What people are saying - Write a reviewWe haven't found any reviews in the usual places. Related books
Contents
Other editions - View all
Common terms and phrasesachieved acoustic models acoustic score adaptation data algorithm amounts of adaptation Artificial Intelligence ASR systems Australia ICS98 Automatic Speech Recognition baseline dictionary British English canonical pronunciation cepstral command corpus computed Confidence Measures considered database decision tree decoding described dynamic weight English phoneme error rate estimated Eurospeech97 Gaussians German German speakers Greece Eur97 Hidden Markov Models HMM models ICSLP98 Italian speakers Linear Regression Maximum A Posteriori Maximum Likelihood means misrecognised MLLR adaptation monophone non-native pronunciation non-native speakers non-native speech normalised number of frames OOV words parameters performance phoneme recogniser probability problem Proceedings pronunciation dictionary pronunciation modelling pronunciation rules pronunciation variants recognition rates recognition result regression class SAMPA semi-supervised adaptation sentences source language speaking rate speech data speech signal spoken spontaneous speech standard MLLR static weight target language test set training data transcriptions triphones updated utterance vocabulary vowel reduction WERs word error rate word sequences Popular passagesPage 8 - are beyond the scope of this book and will not be considered Page 41 - conducts adaptation on the phoneme level. Although yielding a phoneme specific and thus finer adaptation than MLLR, the disadvantage of MAP is that it only updates the parameters of those phonemes that were observed in the adaptation data. As a consequence it needs a lot of adaptation data to reliably re-estimate the parameters for all phonemes used in the system, which takes Page 24 - that is, no transitions are allowed to states whose indices are lower than the current state, Page 41 - MLLR adaptation parameters are pooled and updated with transformation matrices as will be described in the following section, the MAP approach Page 24 - on the state transition coefficients to make sure that large changes in state indices do not occur; hence a constraint Page 114 - N. Cremelie and J.-P. Martens. In Search of Better Pronunciation Models for Speech Recognition. Speech Communication, Page vii - at this point I would like to take the opportunity to thank Page 23 - increases (or stays the same), ie, the states proceed from left to right, Page 116 - Automatic Generation of Multiple Pronunciations Based on Neural Networks and Language Statistics. In Page 116 - and M. Picheny. New Adaptation Techniques for Large Vocabulary Continuous Speech Recognition. In References to this bookFrom other books
From Google ScholarPronunciation Variations of Spanish-Accented English Spoken by ...Hong You, Abeer Alwan, Abe Kazemzadeh, Shrikanth Narayanan - 2005 - Ninth European Conference on Speech Communication and Technology Scalable Localization with Mobility Prediction for Underwater ...Zhong Zhou, Jun-Hong Cui, Amvrossios Bagtzoglou A French Non-Native Corpus for Automatic Speech RecognitionTien-Ping Tan, Laurent Besacier Acoustic Model Interpolation For Non-native Speech RecognitionTien-Ping Tan, Laurent Besacier References from web pagesRobust adaptation to non-native accents in automatic speech ... Lecture Notes in Computer Science, 2002(2560 ) A French Non-Native Corpus for Automatic Speech Recognition Publications Silke Goronzy ROBUST ADAPTATION TO NON - NATIVE ACCENTS IN AUTOMATIC SPEECH ... ROBUST ADAPTATION TO NON - NATIVE ACCENTS IN AUTOMATIC SPEECH ... Livros - ROBUST ADAPTATION TO NON - NATIVE ACCENTS IN AUTOMATIC ... Segnala "Robust Adaptation to Non-Native Accents in Automatic ... Robust Adaptation to Non-Native Accents in Automatic Speech ... book robust adaptation to non-native accents in automatic speech ... Bibliographic information
|