Combining Neural Networks and Hidden Markov Models for Continuous Speech Recognition

Citation

Cohen, M., Cohen, I., Rumelhart, D., Morgan, N., Franco, H., Abrash, V., & Konig, Y. (1992). Combining neural networks and hidden markov models for continuous speech recognition. In In ICSLP-92, 915—918.

Abstract

We present a speaker-independent, continuous-speech recognition system based on a hybrid multilayer perceptron (MLP)/hidden Markov model (HMM). The system combines the advantages of both approaches by using MLPs to estimate the state-dependent observation probabilities of an HMM. New MLP architectures and training procedures are presented that allow the modeling of multiple distributions for phonetic classes and context-dependent phonetic classes. Comparisons with a pure HMM system illustrate advantages of the hybrid approach both in recognition accuracy and in number of parameters required.

↓ Download

Read more from SRI

  • The US Capitol Dome

    Quantum on Capitol Hill

    The SRI-managed Quantum Economic Development Consortium convened quantum innovators and members of Congress to explore the future of quantum technology.

  • Rays of light

    Building the photonic circuits of the future

    SRI’s work on DARPA’s HAPPI program seeks to measurably advance the capability of circuits that transmit information using light rather than electrons.

  • Turning AI into a problem-solving teammate

    To chart the future of human-machine teaming, SRI’s COLLEAGUE project is building an AI-based system designed to act as a true collaborative partner.