A Coarse-to-Fine Approach to Computing the k-Best Viterbi Paths
The Hidden Markov Model (HMM) is a probabilistic model used widely in the fields of Bioinformatics and Speech Recognition. Efficient algorithms for solving the most common problems are well known, yet they all have a running time that is quadratic in the number of hidden states, which can be problematic for models with very large state spaces. The Viterbi algorithm is used to find the maximum likelihood hidden state sequence, and it has earlier been shown that a coarse-to-fine modification can significantly speed up this algorithm on some models. We propose combining work on a k-best version of Viterbi algorithm with the coarse-to-fine framework. This algorithm may be used to approximate the total likelihood of the model, or to evaluate the goodness of the Viterbi path on very large models.
Keywordscoarse-to-fine k-best Viterbi Hidden Markov Models
Unable to display preview. Download preview PDF.
- 2.Charniak, E.: A maximum-entropy-inspired parser. In: Proceedings of the 1st North American Chapter of the Association for Computational Linguistics conference, pp. 132–139. Morgan Kaufmann Inc., San Francisco (2000)Google Scholar
- 3.Charniak, E., Johnson, M.: Coarse-to-fine n -best parsing and MaxEnt discriminative reranking. In: Proceedings of the 43rd Annual Meeting on Association for Computational Linguistics - ACL 2005, June 1, pp. 173–180 (2005)Google Scholar
- 4.Chong, J., Yi, Y., Faria, A., Satish, N., Keutzer, K.: Data-parallel large vocabulary continuous speech recognition on graphics processors. In: Proceedings of the 1st Annual Workshop on Emerging Applications and Many Core Architecture (EAMA), pp. 23–35. sn (2008)Google Scholar
- 6.Du, J., Rozowsky, J., Korbel, J., Zhang, Z., Royce, T., Schultz, M., Snyder, M.: A Supervised Hidden Markov Model Framework for Efficiently Segmenting Tiling Array Data in Transcriptional and ChIP-chip Experiments: Systematically Incorporating Validated Biological Knowledge. Bioinformatics (2008)Google Scholar
- 10.Goodman, J.: Global thresholding and multiple-pass parsing. In: Proceedings of the Second Conference on Empirical Methods in Natural Language Processing, pp. 11–25 (1997)Google Scholar
- 11.Huang, L., Chiang, D.: Better k-best parsing. In: Proc. of IWPT, pp. 53–64 (2005)Google Scholar
- 16.Rabiner, L.: A tutorial on hidden Markov models and selected applications in speech recognition (1990)Google Scholar
- 20.Willett, D., Neukirchen, C., Rigoll, G.: Efficient search with posterior probability estimates in HMM-based speech recognition. In: Proceedings of the 1998 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 1998 (Cat. No.98CH36181), vol. 2, pp. 821–824. IEEE, Los Alamitos (1998)CrossRefGoogle Scholar