Application of EalphaNets to Feature Recognition of Articulation Manner in Knowledge-Based Automatic Speech Recognition

Filippo Sorbello, Antonio Gentile, Giorgio Vassallo, Sabato Marco Siniscalchi, Sabato M. Siniscalchi, Mark A. Clements, Jinyu Li, Giovanni Pilato

Risultato della ricerca: Chapter

1 Citazione (Scopus)

Abstract

Speech recognition has become common in many application domains. Incorporating acoustic-phonetic knowledge into Automatic Speech Recognition (ASR) systems design has been proven a viable approach to rise ASR accuracy. Manner of articulation attributes such as vowel, stop, fricative, approximant, nasal, and silence are examples of such knowledge. Neural networks have already been used successfully as detectors for manner of articulation attributes starting from representations of speech signal frames. In this paper, a set of six detectors for the above mentioned attributes is designed based on the E-αNet model of neural networks. This model was chosen for its capability to learn hidden activation functions that results in better generalization properties. Experimental set-up and results are presented that show an average 3.5% improvement over a baseline neural network implementation
Lingua originaleEnglish
Titolo della pubblicazione ospiteLecture Notes in Computer Science
Pagine140-146
Numero di pagine7
Volume3931
Stato di pubblicazionePublished - 2006

Serie di pubblicazioni

NomeLECTURE NOTES IN COMPUTER SCIENCE

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

Cita questo

Sorbello, F., Gentile, A., Vassallo, G., Siniscalchi, S. M., Siniscalchi, S. M., Clements, M. A., ... Pilato, G. (2006). Application of EalphaNets to Feature Recognition of Articulation Manner in Knowledge-Based Automatic Speech Recognition. In Lecture Notes in Computer Science (Vol. 3931, pagg. 140-146). (LECTURE NOTES IN COMPUTER SCIENCE).

Application of EalphaNets to Feature Recognition of Articulation Manner in Knowledge-Based Automatic Speech Recognition. / Sorbello, Filippo; Gentile, Antonio; Vassallo, Giorgio; Siniscalchi, Sabato Marco; Siniscalchi, Sabato M.; Clements, Mark A.; Li, Jinyu; Pilato, Giovanni.

Lecture Notes in Computer Science. Vol. 3931 2006. pag. 140-146 (LECTURE NOTES IN COMPUTER SCIENCE).

Risultato della ricerca: Chapter

Sorbello, F, Gentile, A, Vassallo, G, Siniscalchi, SM, Siniscalchi, SM, Clements, MA, Li, J & Pilato, G 2006, Application of EalphaNets to Feature Recognition of Articulation Manner in Knowledge-Based Automatic Speech Recognition. in Lecture Notes in Computer Science. vol. 3931, LECTURE NOTES IN COMPUTER SCIENCE, pagg. 140-146.
Sorbello F, Gentile A, Vassallo G, Siniscalchi SM, Siniscalchi SM, Clements MA e altri. Application of EalphaNets to Feature Recognition of Articulation Manner in Knowledge-Based Automatic Speech Recognition. In Lecture Notes in Computer Science. Vol. 3931. 2006. pag. 140-146. (LECTURE NOTES IN COMPUTER SCIENCE).
Sorbello, Filippo ; Gentile, Antonio ; Vassallo, Giorgio ; Siniscalchi, Sabato Marco ; Siniscalchi, Sabato M. ; Clements, Mark A. ; Li, Jinyu ; Pilato, Giovanni. / Application of EalphaNets to Feature Recognition of Articulation Manner in Knowledge-Based Automatic Speech Recognition. Lecture Notes in Computer Science. Vol. 3931 2006. pagg. 140-146 (LECTURE NOTES IN COMPUTER SCIENCE).
@inbook{67ce42e927c64d3ebf00f99c10eb37a0,
title = "Application of EalphaNets to Feature Recognition of Articulation Manner in Knowledge-Based Automatic Speech Recognition",
abstract = "Speech recognition has become common in many application domains. Incorporating acoustic-phonetic knowledge into Automatic Speech Recognition (ASR) systems design has been proven a viable approach to rise ASR accuracy. Manner of articulation attributes such as vowel, stop, fricative, approximant, nasal, and silence are examples of such knowledge. Neural networks have already been used successfully as detectors for manner of articulation attributes starting from representations of speech signal frames. In this paper, a set of six detectors for the above mentioned attributes is designed based on the E-αNet model of neural networks. This model was chosen for its capability to learn hidden activation functions that results in better generalization properties. Experimental set-up and results are presented that show an average 3.5{\%} improvement over a baseline neural network implementation",
keywords = "Robots ; Anthropomorphic robots ; cognitive architecture",
author = "Filippo Sorbello and Antonio Gentile and Giorgio Vassallo and Siniscalchi, {Sabato Marco} and Siniscalchi, {Sabato M.} and Clements, {Mark A.} and Jinyu Li and Giovanni Pilato",
year = "2006",
language = "English",
isbn = "3540331832",
volume = "3931",
series = "LECTURE NOTES IN COMPUTER SCIENCE",
pages = "140--146",
booktitle = "Lecture Notes in Computer Science",

}

TY - CHAP

T1 - Application of EalphaNets to Feature Recognition of Articulation Manner in Knowledge-Based Automatic Speech Recognition

AU - Sorbello, Filippo

AU - Gentile, Antonio

AU - Vassallo, Giorgio

AU - Siniscalchi, Sabato Marco

AU - Siniscalchi, Sabato M.

AU - Clements, Mark A.

AU - Li, Jinyu

AU - Pilato, Giovanni

PY - 2006

Y1 - 2006

N2 - Speech recognition has become common in many application domains. Incorporating acoustic-phonetic knowledge into Automatic Speech Recognition (ASR) systems design has been proven a viable approach to rise ASR accuracy. Manner of articulation attributes such as vowel, stop, fricative, approximant, nasal, and silence are examples of such knowledge. Neural networks have already been used successfully as detectors for manner of articulation attributes starting from representations of speech signal frames. In this paper, a set of six detectors for the above mentioned attributes is designed based on the E-αNet model of neural networks. This model was chosen for its capability to learn hidden activation functions that results in better generalization properties. Experimental set-up and results are presented that show an average 3.5% improvement over a baseline neural network implementation

AB - Speech recognition has become common in many application domains. Incorporating acoustic-phonetic knowledge into Automatic Speech Recognition (ASR) systems design has been proven a viable approach to rise ASR accuracy. Manner of articulation attributes such as vowel, stop, fricative, approximant, nasal, and silence are examples of such knowledge. Neural networks have already been used successfully as detectors for manner of articulation attributes starting from representations of speech signal frames. In this paper, a set of six detectors for the above mentioned attributes is designed based on the E-αNet model of neural networks. This model was chosen for its capability to learn hidden activation functions that results in better generalization properties. Experimental set-up and results are presented that show an average 3.5% improvement over a baseline neural network implementation

KW - Robots ; Anthropomorphic robots ; cognitive architecture

UR - http://hdl.handle.net/10447/59647

M3 - Chapter

SN - 3540331832

VL - 3931

T3 - LECTURE NOTES IN COMPUTER SCIENCE

SP - 140

EP - 146

BT - Lecture Notes in Computer Science

ER -