4th International Conference on Spoken Language Processing

Philadelphia, PA, USA
October 3-6, 1996

Acoustic-phonetic Decoding Based on Elman Predictive Neural Networks

F. Freitag, E. Monte

Universitat Politècnica de Catalunya, Department of Signal Theory and Communications, Barcelona, Spain

In this paper we present a phoneme recognition system based on the Elman predictive neural networks. The recurrent neural networks are used to predict the observation vectors of speech frames. Recognition of phonemes is done using the prediction error as distortion measure in the Viterbi algorithm. The performance of the neural predictive networks is evaluated on both the training database and on a speaker independent test database. The results obtained on the training database are similar to a four state continuous density HMM, results on the test database results are comparable to a three state HMM.

Full Paper

Bibliographic reference.  Freitag, F. / Monte, E. (1996): "Acoustic-phonetic decoding based on elman predictive neural networks", In ICSLP-1996, 522-525.