Earlier Version

Inference algorithms for pattern-based CRFs on sequence data

R. Takhanov, V. Kolmogorov, in:, ICML’13 Proceedings of the 30th International Conference on International, International Machine Learning Society, 2013, pp. 145–153.

Conference Paper | Published | English
Department
Series Title
JMLR
Abstract
We consider Conditional Random Fields (CRFs) with pattern-based potentials defined on a chain. In this model the energy of a string (labeling) x1...xn is the sum of terms over intervals [i,j] where each term is non-zero only if the substring xi...xj equals a prespecified pattern α. Such CRFs can be naturally applied to many sequence tagging problems. We present efficient algorithms for the three standard inference tasks in a CRF, namely computing (i) the partition function, (ii) marginals, and (iii) computing the MAP. Their complexities are respectively O(nL), O(nLℓmax) and O(nLmin{|D|,log(ℓmax+1)}) where L is the combined length of input patterns, ℓmax is the maximum length of a pattern, and D is the input alphabet. This improves on the previous algorithms of (Ye et al., 2009) whose complexities are respectively O(nL|D|), O(n|Γ|L2ℓ2max) and O(nL|D|), where |Γ| is the number of input patterns. In addition, we give an efficient algorithm for sampling. Finally, we consider the case of non-positive weights. (Komodakis & Paragios, 2009) gave an O(nL) algorithm for computing the MAP. We present a modification that has the same worst-case complexity but can beat it in the best case.
Publishing Year
Date Published
2013-06-01
Proceedings Title
ICML'13 Proceedings of the 30th International Conference on International
Volume
28
Issue
3
Page
145 - 153
Conference
ICML: International Conference on Machine Learning
Conference Location
Atlanta, GA, USA
Conference Date
2013-06-16 – 2013-06-21
IST-REx-ID

Cite this

Takhanov R, Kolmogorov V. Inference algorithms for pattern-based CRFs on sequence data. In: ICML’13 Proceedings of the 30th International Conference on International. Vol 28. International Machine Learning Society; 2013:145-153.
Takhanov, R., & Kolmogorov, V. (2013). Inference algorithms for pattern-based CRFs on sequence data. In ICML’13 Proceedings of the 30th International Conference on International (Vol. 28, pp. 145–153). Atlanta, GA, USA: International Machine Learning Society.
Takhanov, Rustem, and Vladimir Kolmogorov. “Inference Algorithms for Pattern-Based CRFs on Sequence Data.” In ICML’13 Proceedings of the 30th International Conference on International, 28:145–53. International Machine Learning Society, 2013.
R. Takhanov and V. Kolmogorov, “Inference algorithms for pattern-based CRFs on sequence data,” in ICML’13 Proceedings of the 30th International Conference on International, Atlanta, GA, USA, 2013, vol. 28, no. 3, pp. 145–153.
Takhanov R, Kolmogorov V. 2013. Inference algorithms for pattern-based CRFs on sequence data. ICML’13 Proceedings of the 30th International Conference on International. ICML: International Conference on Machine Learning, JMLR, vol. 28. 145–153.
Takhanov, Rustem, and Vladimir Kolmogorov. “Inference Algorithms for Pattern-Based CRFs on Sequence Data.” ICML’13 Proceedings of the 30th International Conference on International, vol. 28, no. 3, International Machine Learning Society, 2013, pp. 145–53.

Export

Marked Publications

Open Data IST Research Explorer

Search this title in

Google Scholar