Accessing language specific linguistic information for triphone model generation: feature tables in a speech recognition system

• Autorzy: Kanokphara, S. , Geumann, A. , Carson-Berndsen, J.
• Konferencja: Human Language Technologies as a challenge for Computer Science and Linguistics (2; 21-23.04.2005; Poznań, Poland)
• Języki publikacji: EN

Abstrakty

EN

This paper is concerned with a method for generating phonetic questions used in tree-based state tying for speech recognition. In order to implement a speech recognition system, languagedependent knowledge which goes beyond annotated material is usually required. The approach presented here generates phonetic questions for decision trees based on a feature table that summarizes the articulatory characteristics of each sound. On the one hand, this method allows better language-specific triphone models to be defined given only a feature-table as linguistic input. On the other hand, the feature-table approach facilitates efficient definition of triphone models for other languages since again only a feature table for this language is required. The approach is exemplified with speech recognition systems for English and Thai.

Słowa kluczowe

EN

speech recognition; tree-based state tying; phonetic features

• Czasopismo: Archives of Control Sciences
• Rocznik: 2005
• Tom: Vol. 15, no. 3
• Strony: 321-328
• Opis fizyczny: Bibliogr. 11 poz., tab.

Twórcy

autor

Kanokphara, S.

• School of Computer Science and Informatics, University College, Dublin, Ireland,

autor

Geumann, A.

• School of Computer Science and Informatics, University College, Dublin, Ireland,

autor

Carson-Berndsen, J.

• School of Computer Science and Informatics, University College, Dublin, Ireland,

Bibliografia

• [1] K. Beulen and H. Ney: Automatic Question Generation for Decision Tree Based Slate Tying. Proc. IEEE Int. Conf. on Acoustics, Speech, and Signal Processing. 2 (1998), 805-809.
• [2] J. Carson-Berndsen: Time Map Phonology. Kluwer, Dordrecht, 1998.
• [3] C. Chelba and R. Morton: Mutual Information Phone Clustering for Decision Tree Induction. Proc. Int. Conf. on Spoken Language Processing, (2002).
• [4] F. Diehl and A. Moreno: Acoustic Phonetic Modelling using Local Codebook Features. Proc. Int. Conf. on Spoken Language Processing, (2004).
• [5] J. S. Garofolo, L. F. Lamel. W. M. Fisher, J. G. Fiscus, D. S. Pallett and N. L. Dahlgren: DARPA TIMIT Acoustic-Phonetic Continuous Speech Corpus CDROM, NIST, (1993).
• [6] A. Geumann: Towards a New Level of Annotation Detail of Multilingual Speech Corpora. Proc. Int. Conf. on Spoken Language Pnxessing. (2004), 1096-1099.
• [7] S. Kanokphara and J. Carson-Berndsen: Automatic Question Generation for KMM State Tying using a Feature Table. Proc. Australian Int. Conf. on Speech Science & Technology. (2004).
• [8] S. Kanokphara and J. Carson-Berndsen: Feature-Tablc-Based Automatic Question Generation for Tree-Based State Tying: A Practical Implementation. Proc. of the IHth Int. Conf. on Industrial and Engineering Applications of Artificial Intelligence and Expert Systems. Ban. Italy, (2005).
• [9] L. Netsch and A. Bernard: Automatic and language independent triphone training using phonetic tables. Proc. IEEE Int. Conf. on Acoustics, Speech, and Signal Processing. Montreal, Canada, (2004).
• [10] J. J. Odell: The Use of Context in Large Vocabulary Speech Recognition. Ph.D. diss., Cambridge University, Cambridge, 1995.
• [11] R. Singh. B. Raj and R. M. Stern: Automatic Clustering and Generation of Contextual Questions for Tied States in Hidden Markov Models. Proc. Int. Conf. on Spoken Language Processing, 1 (1999), 117-120.