Article

Slovak language model from internet text data

Authors:

Ján Staš,

Daniel Hládek,

Matúš Pleva,

Jozef JuhárAuthors Info & Claims

Proceedings of the Third COST 2102 international training school conference on Toward autonomous, adaptive, and context-aware multimodal interfaces: theoretical and practical issues

Pages 340 - 346

Published: 15 March 2010 Publication History

Publisher Site

Abstract

Automatic speech recognition system is one of the parts of the multimodal dialogue system. It is necessary to create correct vocabulary and to generate suitable language model for this purpose. The main aim of this article is to describe a process of building statistical models of the Slovak language with large vocabulary trained on the text data gathered mainly from Internet sources. Several smoothing techniques for different sizes of vocabulary have been used in order to obtain an optimal model of the Slovak language. We have also employed pruning technique based on relative entropy for size reduction of a language model to find the maximum threshold of pruning with minimum degradation in recognition accuracy. Tests were performed by the decoder based on the HTK Toolkit.

References

[1]

Chollet, G., Esposito, A., Gentes, A., Horain, P., Karam, W., Li, Z., Pelachaud, C., Perrot, P., Petrovska-Delacrétaz, D., Zhou, D., Zouari, L.: Multimodal Human Machine Interactions in Virtual and Augmented Reality. In: Esposito, A., Hussain, A., Marinaro, M., Martone, R. (eds.) COST Action 2102. LNCS(LNAI), vol. 5398, pp. 1-23. Springer, Heidelberg (2009).

Digital Library

Google Scholar

[2]

Jurafsky, D., Martin, J.H.: Speech and Language Processing: An Introduction to Natural Language Processing, Computational Linguistics, and Speech Recognition, 2nd edn., p. 998. Prentice Hall, Englewood Cliffs (2009) ISBN-13 978-0-13-504196-3.

Digital Library

Google Scholar

[3]

Chen, S.F., Goodman, J.: An Empirical Study of Smoothing Techniques for Language Modeling. Technical Report TR-10-98, p. 63 (1998).

Google Scholar

[4]

Stolcke, A.: Entropy-based Pruning of Backoff Language Models. In: Proc. DARPA Broadcast News Transcription and Understanding Workshop, pp. 270-274 (1998).

Google Scholar

[5]

Mirilovič, M., Juhár, J., Cižmár, A.: Large Vocabulary Continuous Speech Recognition in Slovak. In: Proc. of International Conference on Applied Electrical Engineering and Informatics, Athens, Greece, pp. 73-77 (2008) ISBN 978-80-553-0066-5.

Google Scholar

[6]

Stolcke, A.: SRILM - An Extensible Language Modeling Toolkit. In: Proc. of the 7th International Conference on Spoken Language Processing, Denver, Colorado, pp. 901-904 (2002).

Google Scholar

[7]

Cowan, I.A., Moore, D., Dines, J., Gatiza-Perez, D., Flynn, M., Wellner, P., Bourlard, H.: On the Use of Information Retrieval Measures for Speech Recognition Evaluation. In: IDIAP-RR-73, Martigny, Switzerland, p. 15 (2005).

Google Scholar

[8]

Young, S., Odell, J., Ollason, D., Valtchev, V., Woodland, P., Evermann, G., Hain, T., Kershaw, D., Moore, G.: The HTK Book (v3.4). Cambridge University, Cambridge (2009).

Google Scholar

[9]

Rusko, M., Trnka, M., Daržagín, S.: MobilDat-SK - A Mobile Telephone Extension to the SpeechDat-E SK Telephone Speech Database in Slovak. In: Proc. of the 11th International Conference Speech and Computer, SPECOM 2006, pp. 485-488 (2006).

Google Scholar

[10]

Mirilovič, M., Juhár, J., Čižmár, A.: Comparison of Grapheme and Phoneme Based Acoustic Modeling in LVCSR Task in Slovak. In: Proc. of the 7th International Conference on Spoken Language Processing, Denver, Colorado, pp. 901-904 (2002).

Google Scholar

Recommendations

Topic-Dependent Language Model with Voting on Noun History

Language models (LMs) are an important field of study in automatic speech recognition (ASR) systems. LM helps acoustic models find the corresponding word sequence of a given speech signal. Without it, ASR systems would not understand the language and it ...
Toward enhanced Arabic speech recognition using part of speech tagging

One major source of suboptimal performance in automatic continuous speech recognition systems is misrecognition of small words. In general, errors resulting from small words are much more than errors resulting from long words. Therefore, compounding ...
A Language Model for Intelligent Speech Recognition of Power Dispatching
ACM TURC '21: Proceedings of the ACM Turing Award Celebration Conference - China

The accuracy of power dispatching speech recognition system is related to the effect of language model. In order to improve the accuracy of power dispatching speech recognition, this paper proposes a class label language model based on double ...

Comments

Please enable JavaScript to view thecomments powered by Disqus.

Information & Contributors

Information

Published In

Proceedings of the Third COST 2102 international training school conference on Toward autonomous, adaptive, and context-aware multimodal interfaces: theoretical and practical issues

March 2010

472 pages

ISBN:9783642181832

Editors:
Anna Esposito
Second University of Naples and IIASS, International Institute for Advanced Scientific Studies, Vietri sul Mare, SA, Italy
,
Antonietta M. Esposito
Istituto Nazionale di Geofisica e Vulcanologia, Osservatorio Vesuviano, Napoli, Italy
,
Raffaele Martone
Second University of Naples, Department of Engineering and Informatics, Aversa, CE, Italy
,
Vincent C. Müller
Anatolia College, ACT, Department of Humanities and Social Sciences, Pylaia, Greece
,
Gaetano Scarpetta
University of Salerno and IIASS, International Institute for Advanced Scientific Studies, Baronissi, SA, Italy

Publisher

Springer-Verlag

Berlin, Heidelberg

Publication History

Published: 15 March 2010

Author Tags

Qualifiers

Article

Recommendations

Topic-Dependent Language Model with Voting on Noun History

Toward enhanced Arabic speech recognition using part of speech tagging

A Language Model for Intelligent Speech Recognition of Power Dispatching