Segment-Based Classes for Language Modeling Within the Field of CSR

In this work, we propose and formulate two different approaches for the language model integrated in a Continuous Speech Recognition System. Both of them make use of class-based language models where classes are made up of segments or sequences of words. On the other hand, an interpolated model of a class-based language model and a word-based language model is explored as well. The experiments carried out over a spontaneous dialogue corpus in Spanish, demonstrate that introducing segments of words in a class-based language model a better performance of a Continuous Speech Recognition system can be achieved.

[1]  Eduardo Lleida,et al.  Design and acquisition of a telephone spontaneous speech dialogue corpus in Spanish: DIHANA , 2006, LREC.

[2]  Alexander H. Waibel,et al.  Class phrase models for language modeling , 1996, Proceeding of Fourth International Conference on Spoken Language Processing. ICSLP '96.

[3]  Roger K. Moore.  Computer Speech and Language , 1986 .

[4]  Joan-Andreu Sánchez,et al.  Estimation of stochastic context-free grammars and their use as language models , 2005, Comput. Speech Lang..

[5]  M. Inés Torres,et al.  Category-based Language Models in a Spanish Spoken Dialogue System , 2006, Proces. del Leng. Natural.

[6]  Daniel Marcu,et al.  A Phrase-Based,Joint Probability Model for Statistical Machine Translation , 2002, EMNLP.

[7]  M. Inés Torres,et al.  k-TSS language models in speech recognition systems , 2001, Comput. Speech Lang..

[8]  M. Lennig,et al.  A language model for very large-vocabulary speech recognition , 1992 .

[9]  Frederick Jelinek,et al.  Statistical methods for speech recognition , 1997 .

[10]  Enrique Vidal,et al.  Inference of k-Testable Languages in the Strict Sense and Application to Syntactic Pattern Recognition , 1990, IEEE Trans. Pattern Anal. Mach. Intell..

[11]  Imed Zitouni,et al.  Backoff hierarchical class n-gram language models: effectiveness to model unseen events in speech recognition , 2007, Comput. Speech Lang..

[12]  Thomas Niesler,et al.  A variable-length category-based n-gram language model , 1996, 1996 IEEE International Conference on Acoustics, Speech, and Signal Processing Conference Proceedings.

[13]  Hong-Kwang Jeff Kuo,et al.  Phrase-based language models for speech recognition , 1999, EUROSPEECH.

[14]  Franz Josef Och,et al.  An Efficient Method for Determining Bilingual Word Classes , 1999, EACL.

[15]  Frédéric Bimbot,et al.  Language modeling by variable length sequences: theoretical formulation and evaluation of multigrams , 1995, 1995 International Conference on Acoustics, Speech, and Signal Processing.

[16]  Robert L. Mercer,et al.  Class-Based n-gram Models of Natural Language , 1992, CL.

[17]  Thomas Niesler,et al.  Comparison of part-of-speech and automatically derived category-based language models for speech recognition , 1998, Proceedings of the 1998 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP '98 (Cat. No.98CH36181).

[18]  Isabel Trancoso,et al.  Transducer composition for "on-the-fly" lexicon and language model integration , 2001, IEEE Workshop on Automatic Speech Recognition and Understanding, 2001. ASRU '01..

[19]  Hermann Ney,et al.  Improvements in Phrase-Based Statistical Machine Translation , 2004, NAACL.