34
11

Inference algorithms for pattern-based CRFs on sequence data

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...xnx_1...x_n is the sum of terms over intervals [i,j][i,j] where each term is non-zero only if the substring xi...xjx_i...x_j equals a prespecified pattern α\alpha. 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(n L), O(nLmax)O(n L \ell_{max}) and O(nLmin{D,log(max+1)})O(n L \min\{|D|,\log (\ell_{max}+1)\}) where LL is the combined length of input patterns, max\ell_{max} is the maximum length of a pattern, and DD is the input alphabet. This improves on the previous algorithms of (Ye et al., 2009) whose complexities are respectively O(nLD)O(n L |D|), O(nΓL2max2)O(n |\Gamma| L^2 \ell_{max}^2) and O(nLD)O(n L |D|), where Γ|\Gamma| 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)O(n L) algorithm for computing the MAP. We present a modification that has the same worst-case complexity but can beat it in the best case.

View on arXiv
Comments on this paper