http://www.cnr.it/ontology/cnr/individuo/prodotto/ID241071

Improvements in Neural-Network Training and Search Techniques for Continuous Digit Recognition (Articolo in rivista)

Type

Label

Improvements in Neural-Network Training and Search Techniques for Continuous Digit Recognition (Articolo in rivista) (literal)

Anno

Alternative label

Hosom J.P., Cole R.A., Cosi P. (1998)
Improvements in Neural-Network Training and Search Techniques for Continuous Digit Recognition
in Australian journal of intelligent information processing systems
(literal)

Http://www.cnr.it/ontology/cnr/pubblicazioni.owl#autori

Pagina inizio

Pagina fine

Http://www.cnr.it/ontology/cnr/pubblicazioni.owl#altreInformazioni

Hosom J.P., Cole R.A., Cosi P. \"Improvements in Neural-Network Training and Search Techniques for Continuous Digit Recognition\" in AJIIPS, Australian Journal of Intelligent Information Processing Systems, Vol. 5, No. 4, Summer 1998 pp. 277-284 http://www.cslu.ogi.edu/people/hosom/pubs/hosomcosicole_AJIIPS-digits_1999.pdf http://www2.pd.istc.cnr.it/Papers/PieroCosi/cp-AJIIPS99.pdf (literal)

Http://www.cnr.it/ontology/cnr/pubblicazioni.owl#url

Http://www.cnr.it/ontology/cnr/pubblicazioni.owl#numeroVolume

Rivista

Http://www.cnr.it/ontology/cnr/pubblicazioni.owl#numeroFascicolo

Note

Http://www.cnr.it/ontology/cnr/pubblicazioni.owl#affiliazioni

Center for Spoken Language Understanding (CSLU-OGI) Oregon Graduate Institute (OGI), P.O. Box 91000, Portland, Oregon 97291 USA Hosom J.P. Center for Spoken Language Understanding (CSLU-Boulder) University of Colorado, Boulder, Campus Box 258, Boulder, Colorado 80309 USA Cole R.A. ISTC CNR, UOS Padova Cosi P. (literal)

Titolo

Improvements in Neural-Network Training and Search Techniques for Continuous Digit Recognition (literal)

Abstract

This paper describes a set of experiments on training and search techniques for development of a neural-network based continuous digits recognizer. When the best techniques from these experiments were combined to train a final recognizer, there was a 56% reduction in word-level error on the continuous digits recognition task. The best system had word accuracy of 97.67% on a test set of the OGI 30K Numbers corpus; this corpus contains naturally-produced continuous digit strings recorded over telephone channels. Experiments investigated the effects of the feature set, the amount of data used for training, the type of context-dependent categories to be recognized, the values for duration limits, and the type of grammar. The experiments indicate that the grammar and duration limits had a greater effect on recognition accuracy than the output categories, cepstral features, or a doubling of the amount of training data. In addition, the forwardbackward method of training neural networks was employed in developing the final network. (literal)

Prodotto di