Improved hidden Markov model for speech recognition and POS tagging

Abstract

In order to overcome defects of the classical hidden Markov model (HMM), Markov family model (MFM), a new statistical model was proposed. Markov family model was applied to speech recognition and natural language processing. The speaker independently continuous speech recognition experiments and the part-of-speech tagging experiments show that Markov family model has higher performance than hidden Markov model. The precision is enhanced from 94.642% to 96.214% in the part-of-speech tagging experiments, and the work rate is reduced by 11.9% in the speech recognition experiments with respect to HMM baseline system.

Keywords

hidden Markov model / Markov family model / speech recognition / part-of-speech tagging

References

[1] RabinerL., JuangB. H.Fundamentals of speech recognition [M], 1993, New Jersey, USA, Prentice Hall: 210-235

[2] LiuS., LiS., ZhaoT.-jun.. Directly smooth interpolation algorithm in head-driven parsing [J]. Journal of Software, 2009, 20(11): 2915-2924

[3] AviranS., SiegelP. H., WolfJ. K.. Optimal parsing trees for run-length coding of biased data [J]. IEEE Transaction on Information Theory, 2008, 54(2): 841-849

[4] ZhouD.-y., HeY.-lan.. Discriminative training of the hidden vectors state model for semantic parsing [J]. IEEE Transaction on Knowledge and Data Engineering, 2009, 21(1): 66-77

[5] LiZ.-h., CheW.-x., LiuT.. Beam-search based high-order dependency parser [J]. Journal of Chinese Information Processing, 2010, 24(1): 37-41

[6] GOLDBERG Y, ELHADAD M. SplitSVM: Fast space-efficient non-heuristic, polynomial kernel computation for NLP applications [C]// Proceedings of the 46th Annual Meeting of the Association for Computational Linguistics on Human Language Technologies (ALT). Columbus USA, 2008: 237–240.

[7] NIVRE J, McDONALD R. Integrating graph-based and transition-based dependency parsers [C]// Proceedings of the 46th Annual Meeting of the Association for Computational Linguistics on Human Language Technologies (ALT). Columbus USA, 2008: 950–958.

[8] CHE Wan-xiang, LI Zheng-hua, HU Yu-xuan, LI Yong-qiang, QIN Bing, LIU Ting, LI Sheng. A Cascaded syntactic and semantic dependency parsing system [C]// CoNLL 2008: Proceedings of the 12th Conference on Computational Natural Language Learning, 2008: 238–242.

[9] WANG Hong-ling, ZHOU Guo-dong, ZHU Qiao-ming, QIAN Pei-de. Exploring various features in semantic role labeling [C]// Proceedings of 2008 International Conference on Advanced Language Processing and Web Information Technology (ALPIT2008). 2008: 23–25.

[10] VilaresJ., AlonsoM. A., VilaresM.. Extraction of complex index terms in non-English IR: A shallow parsing based approach [J]. Information Processing and Management, 2008, 44: 1517-1537

[11] ZhengY., ChengX.-c., ChenK.. Filtering noise in Web pages based on parsing tree [J]. Journal of China Universities of Posts and Telecommunications, 2008, 15(9): 46-50

[12] TURISH B. Part-of-speech tagging with finite-state morphology [C]// Proceedings of the International conference on Collocations and Idioms: linguistic, Computational, and Psycholinguistic perspective. Berlin German, 2003: 18–20.

[13] BRANTS T. A statistical Part-of-Speech tagger [C]// Proceeding of the Sixth Applied Natural Language Processing Conference (ANLP-2000). Seattle USA, 2000: 224–231.

[14] GIMENEZ J, MARQUEZ L. Fast and accurate part-of-speech tagging: The SVM approach revisited [C]// Proceedings of the 4th International Conference on Recent Advances in National Language Processing (4th RANLP). Bulgaria, 2003: 153–163.

[15] GongY.. Stochastic trajectory modeling and sentence searching for continuous speech recognition [J]. IEEE Transactions on Speech Audio Processing, 1997, 5(1): 33-44

[16] YanB.-f., ZhuX.-y., ZhangZ.-j., et al.. Robust speech recognition based on neighborhood space [J]. Journal of Software, 2007, 18(11): 878-883

[17] LiuY.-h., LiuQ., RenQ.. Speech recognition based on fuzzy clustering neural network [J]. Chinese Journal of Computers, 2006, 29(10): 1894-1900

[18] TangY., LiuW.-j., XuB.. Mandarin digit string recognition based on segment model using posterior probability decoding [J]. Chinese Journal of Computers, 2006, 29(4): 635-641.

[19] Hui, DuL.-min.. Combination of acoustic models trained from different unit sets for chinese continuous speech recognition [J]. Journal of Electronics and Information Technology, 2006, 28(11): 2045-2049

[20] YanL., LiuG., GuoJ.. A study on robustness of large vocabulary chinese continuous speech recognition system based on wavelet analysis [J]. Journal of Chinese Information Processing, 2006, 20(2): 60-65

[21] WANG W J, CHEN S H. The study of prosodic modeling for mandarin speech [C]// Proceedings of the International Computer Symposium (ICS). Hualien, China, 2002: 1777–1784.

[22] CHANG E, ZHOU J L, DI S, HUANG C, LEE K F. Large vocabulary mandarin speech recognition with different approaches in modeling tones [C]// Proceedings of the 6th International Conference on Spoken Language Processing (ICSLP 200). San Jose, USA, 2000(II): 983–986.

[23] WangZ.-y., XiaoX.. Duration distribution based HMM speech recognition models [J]. Chinese Journal of Electronics, 2004, 32(1): 46-49

[24] MANNING C D, SCHUTZE H. Foundations of statistical natural language processing [M]. London: The MIT Press, 1999: 121–147.

[25] RabinerL. R.. A tutorial on hidden Markov models and selected applications in speech recognition [J]. Proceedings of the IEEE, 1989, 77(2): 257-285

[26] HON Hsiao-wuen, WANG Kuan-san. Unified frame and segment based models for automatic speech recognition [C]// Proceedings of the IEEE International Conference on Acoustic, Speech and Signal Processing (ICASSP 2000). Turkey, 2000: 1017–1020.