Parallel distributed processes for speaker independent acoustic- phonetic decoding
Résumé
We aim at examining to what extent a knowledge-based model can recognise segmental structure without feedback from semantic information and without stochastic modelling. The system pro-posed is inspired by some features of human cognitive processing: the speech signal activates parallel distributed processes of decoding. A supervisor module takes the final decision after the access to a dictionary and a top-down verification.
Origine | Fichiers produits par l'(les) auteur(s) |
---|