Third European Conference on Speech Communication and Technology

Berlin, Germany
September 22-25, 1993


A Neural Network Based, Speaker Independent, Large Vocabulary, Continuous Speech Recognition System: The WERNICKE Project

Tony Robinson (2), L. Almeida (3), Jean-Marc Boite (1), Hervé Bourlard (1), Frank Fallside (2), Michael M. Hochberg (2), D. Kershaw (2), P. Kohn (4), Y. Konig (4), Nelson Morgan (4), J. P. Neto (3), Steve Renals (2), Marco Saerens (3), C. Wooters (4)

(1) Lernout and Hauspie SpeechProducts, Belgium
(2) Cambridge University Engineering Department, UK
(3) Instituto de Engenharia de Sistemas e Computadores (INESC), Portugal
(4) International Computer Science Institute (ICSI), USA

This paper describes the research underway for the ESPRIT WERNICKE project. The project brings together a number of different groups from Europe and the US and focuses on extending the state-of-the-art for hybrid hidden Markov model/connectionist approaches to large vocabulary, continuous speech recognition. This paper describes the specific goals of the research and presents the work performed to date. Results are reported for the resource management talker-independent recognition task. The paper concludes with a discussion of the projected future work.

Keywords: Recognition, Neural Nets, HMM

Full Paper

Bibliographic reference.  Robinson, Tony / Almeida, L. / Boite, Jean-Marc / Bourlard, Hervé / Fallside, Frank / Hochberg, Michael M. / Kershaw, D. / Kohn, P. / Konig, Y. / Morgan, Nelson / Neto, J. P. / Renals, Steve / Saerens, Marco / Wooters, C. (1993): "A neural network based, speaker independent, large vocabulary, continuous speech recognition system: the WERNICKE project", In EUROSPEECH'93, 1941-1944.