Tied Mixture Continuous Parameter Modeling for Speech Recognition

Abstract

The acoustic-modeling problem in automatic speech recognition is examined with the specific goal of unifying discrete and continuous parameter approaches. To model a sequence of information-bearing acoustic feature vectors which has been extracted from the speech waveform via some appropriate front-end signal processing, a speech recognizer basically faces two alternatives: a) assign a multivariate probability distribution directly to the stream of vectors, or b) use a time-synchronous labeling acoustic processor to perform vector quantization on this stream, and assign a multinomial probability distribution to the output of the vector quantizer. With few exceptions, these two methods have traditionally been given separate treatment. Here we consider a class of very general hidden Markov models which can accommodate feature vector sequences lying either in a discrete or in a continuous space; the new class allows one to represent the prototypes in an assumption limited, yet convenient way, as tied mixtures of simple multivariate densities. Speech recognition experiments, reported for two (5000- and 20 000-word vocabularly) office correspondence tasks, demonstrate some of the benefits associated with this technique. © 1990 IEEE