Structured Speech Modeling

Li Deng, Dong Yu, and Alex Acero

Abstract

Abstract—Modeling dynamic structure of speech is a novel paradigm in speech recognition research within the generative modeling framework, and it offers a potential to overcome limitations of the current hidden Markov modeling approach. Analogous to structured language models where syntactic structure is exploited to represent long-distance relationships among words [5], the structured speech model described in this paper makes use of the dynamic structure in the hidden vocal tract resonance space to characterize long-span contextual influence among phonetic units. A general overview is provided first on hierarchically classified types of dynamic speech models in the literature. A detailed account is then given for a specific model type called the hidden trajectory model, and we describe detailed steps of model construction and the parameter estimation algorithms. We show how the use of resonance target parameters and their temporal filtering enables joint modeling of long-span coarticulation and phonetic reduction effects. Experiments on phonetic recognition evaluation demonstrate superior recognizer performance over a modern hidden Markov model-based system. Error analysis shows that the greatest performance gain occurs within the sonorant speech class.

Index Terms—Hidden dynamics, hidden trajectory, long span modeling, maximum-likelihood, nonlinear prediction, parameter learning, structured modeling, vocal tract resonance.

Details

Publication typeArticle
Published inIEEE Trans. on Audio, Speech and Language Processing
Pages1492-1504
Volume14
Number5
PublisherInstitute of Electrical and Electronics Engineers, Inc.
> Publications > Structured Speech Modeling