Learning strategies in speech recognition

Gavăt, Inge; Dumitru, Octavian; Iancu, Claudia; Costache, Gabriel

doi:10.1109/elmar.2005.193686

Cited by 5 publications

(2 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the learning phase the structures are trained and the weights of the networks are established in two steps: (a) the determination of the winner neurons; (b) the adaptation of the weights for the winner neurons and for the neurons existing in a certain neighborhood. In this step important are: (a) the neighborhood dimension r(t) decreasing during the learning; (b) the learning rate η(t) following in our experiments one of the laws [5]:…”

Section: Kohonen Mapsmentioning

confidence: 99%

See 1 more Smart Citation

NN and Hybrid Strategies for Speech Recognition in Romanian Language

Dumitru¹,

Gavăt²

2008

Proceedings of the 4th International Workshop on Artificial Neural Networks and Intelligent Information Processing

View full text Add to dashboard Cite

In this paper we present results obtained with learning structures more "human likely" than the very effective and widely used hidden Markov model. Good results were obtained with simple artificial neural networks like the multilayer perceptron or the Kohonen maps. Hybrid structures have proven also their efficiency, the neuro-statistical hybrid applied enhancing the digit recognition rate of the initial HMM. Also fuzzy variants of the MLP and HMM gave good results in the tested tasks of vowel recognition.

show abstract

Section: Kohonen Mapsmentioning

confidence: 99%

“…The results obtained for hybrid strategies are compared with other learning strategies (hidden Markov models, Support Vector Machine) and the performance being appreciated by their recognition rate and by their generalization capacity [5]. The word recognition rate (WRR) is reported in Table 4.…”

Section: Digit Recognitionmentioning

confidence: 99%