Structurally discriminative graphical models for automatic speech recognition - results from the 2001 Johns Hopkins Summer Workshop

Abstract

In recent years there has been growing interest in discriminative parameter training techniques, resulting from notable improvements in speech recognition performance on tasks ranging in size from digit recognition to Switchboard. Typified by Maximum Mutual Information training, these methods assume a fixed statistical modeling structure, and then optimize only the associated numerical parameters (such as means, variances, and transition matrices). In this paper, we explore the significantly different methodology of discriminative structure learning. Here, the fundamental dependency relationships between random variables in a probabilistic model are learned in a discriminative fashion, and are learned separately from the numerical parameters. Tn order to apply the principles of structural discriminability, we adopt the framework of graphical models, which allows an arbitrary set of variables with arbitrary conditional independence relationships to be modeled at each time frame. We present results using a new graphical modeling toolkit (described in a companion paper) from the recent 2001 Johns Hopkins Summer Workshop. These results indicate that significant gains result from discriminative structural analysis of both conventional MFCC and novel AM-FM features on the Aurora continuous digits task.

DOI: 10.1109/ICASSP.2002.5743662

Extracted Key Phrases

7 Figures and Tables

Cite this paper

@article{Zweig2002StructurallyDG, title={Structurally discriminative graphical models for automatic speech recognition - results from the 2001 Johns Hopkins Summer Workshop}, author={Geoffrey Zweig and Jeff A. Bilmes and Thomas O Richardson and Karim Filali and Karen Livescu and Peng Xu and K. Jackson and Yigal Brandman and Eric D. Sandness and E. Holtz and J. Torres and B. Byrne}, journal={2002 IEEE International Conference on Acoustics, Speech, and Signal Processing}, year={2002}, volume={1}, pages={I-93-I-96} }