Continuous Speech Recognition Predictive Systems

Mellouk, Abdelhamid; Gallinari, Patrick

doi:10.1007/978-1-4471-2063-6_92

Cited by 2 publications

(3 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We have proposed [6,7] Table 1. Performances (% correct) on phone recognition for the six phone classes in the NPS and the DNPS.…”

Section: Implementationsmentioning

confidence: 99%

“…It has been found much more efficient , both for training time and accuracy, to perform discrimination between the correct model and selected nearest compe titors at each time instead of all competitors [6].…”

Section: Probabilistic Interpretation Of Frame Discriminative Trainingmentioning

confidence: 99%

“…Simple discriminative training at the frame level[6] raised the performances of NPS up-to the state of the art (74.9% of phonetic accuracy). This work and others proved the feasibility of the predictive NN approach for CSR.…”

mentioning

confidence: 97%

See 2 more Smart Citations

Global discrimination for neural predictive systems based on N-best algorithm

Mellouk

Gallinari²

1995 International Conference on Acoustics, Speech, and Signal Processing

Self Cite

View full text Add to dashboard Cite

We d .esc . ribe a general formalism for training neural pred1ctIve systems. We then introduce discrimination at the frame level and show how it relates to maximum mutual information training. Last, we propose an approach for performing discri�natiol1 in predictive systems at the sequence level , 1t makes use of N-Best sequence selection. Performances: for acoustic-phonetic decoding reach 77.4% phone accuracy on 1988 version of TIMIT.

show abstract

“…We have proposed [6,7] Table 1. Performances (% correct) on phone recognition for the six phone classes in the NPS and the DNPS.…”

Section: Implementationsmentioning

confidence: 99%

Section: Probabilistic Interpretation Of Frame Discriminative Trainingmentioning

confidence: 99%