基于TANDEM的区分性训练在语音评测中的应用研究-信号与信息处理专业论文.docxVIP

摘要关键词：语音评测系统语音检错语音评分区分性训练最小音素错误TANDEM 摘要 关键词：语音评测系统语音检错语音评分区分性训练最小音素错误TANDEM 多层感知器 Ⅱ AB AB STRACT In recent years，the speech assessment and evaluation systems such as computer assisted language learning system are more and more applied in the oral exams and language learning activities．These systems can not only help teachers give scores of oral tests much more objectively and efficiently but also help students evaluate their pronunciation proficiency immediately and accurately．Now most of speech assessment and evaluation systems USe MFCC features and maximum likelihood estimation(MLE)to establish the statiStical models．This popular MLE based statistical method also has some disadvantages．The most prominent one is that the discriminability of the MLE statistical models is limited．Another one is when the training data is not large enough，MLE method is unlikely to reach an optimization solution．To solve these problems，this thesis proposes discriminative training criterions and TANDEM features which intend to improve the performance of the current speech evaluation system． The whole thesis is organized as follows： Chapter 1 gives a brief summary on the development and background of speech evaluation system．Then We explain the basic principle and system structure for speech scoring system and speech error detection system respectively．Finally,we give introduction to some concept of speech recognition technology as the foundation of speech evaluation．such as acoustic features，acoustic model，language model and SO On． Chapter 2 gives an overview on Bayesian decision theory firstly．To overcome the Weakness of MLE，we bring discriminative training methods for hidden Markov models into speech evaluation system．Four typical discriminative training criterions and some updating methods of acoustic model parameters are introduced，then，they are deftned in a unified framework．After,we analyze the relationship between the target of speech evaluation task and the objection function