Recurrent neural prediction models for speech recognition

Na, Kyungmin; Ryu, Jekwan; Chang, Dong-Il; Chae, Soo-Ik; Ann, Souguil

doi:10.21437/eurospeech.1995-421

Cited by 4 publications

(1 citation statement)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It was shown that speech recognition results comparable to the results using HMMs can be obtained by approaches, in which neural networks are used to estimate posterior probabilities of phonemes [1]. Another approach is to use neural networks as predictors of observation vectors of speech frames as shown in [2], [3], and [4]. When working as predictors, neural networks map past observation vectors into a predicted observation vector, and the prediction error is used in a Viterbi decoding.…”

Section: Introductionmentioning

confidence: 99%

Acoustic-phonetic decoding based on elman predictive neural networks

Freitag¹,

Monte²

1996

4th International Conference on Spoken Language Processing (ICSLP 1996)

View full text Add to dashboard Cite

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.

show abstract