Summary: Proceedings of Speech Research Symposium XII, June, 1992.
s
Multiple­State Context­Dependent Phonetic Modeling with MLP
Michael Cohen*, Horacio Franco*, Nelson Morgan**,
*
David Rumelhart***, and Victor Abrash*
Speech Research Program, SRI International, Menlo Park, CA 94025
4
** Intl. Computer Science Inst., 1947 Center Street, Suite 600, Berkeley, CA 9470
*** Stanford University, Dept. of Psychology, Stanford, CA 94305
E
Abstract
arlier hybrid multilayer perceptron (MLP)/hidden Markov model (HMM) continuous speech recognition sys­
r
g
tems have not modeled context­dependent phonetic effects, sequences of distributions for phonetic models, o
ender­based speech consistencies. In this paper we present a new MLP architecture and training procedure for
t
"
modeling context­dependent phonetic classes with a sequence of distributions. A new training procedure tha

Source: Abrash, Victor - Speech Technology & Research Laboratory, SRI International

Collections: Computer Technologies and Information Sciences