Sciweavers

Share
AAAI
2008

CRF-OPT: An Efficient High-Quality Conditional Random Field Solver

9 years 1 months ago
CRF-OPT: An Efficient High-Quality Conditional Random Field Solver
Conditional random field (CRF) is a popular graphical model for sequence labeling. The flexibility of CRF poses significant computational challenges for training. Using existing optimization packages often leads to long training time and unsatisfactory results. In this paper, we develop CRFOPT, a general CRF training package, to improve the efficiency and quality for training CRFs. We propose two improved versions of the forwardbackward algorithm that exploit redundancy and reduce the time by several orders of magnitudes. Further, we propose an exponential transformation that enforces sufficient step sizes for quasiNewton methods. The technique improves the convergence quality, leading to better training results. We evaluate CRF-OPT on a gene prediction task on pathogenic DNA sequences, and show that it is faster and achieves better prediction accuracy than both the HMM models and the original CRF model without exponential transformation.
Minmin Chen, Yixin Chen, Michael R. Brent
Added 02 Oct 2010
Updated 02 Oct 2010
Type Conference
Year 2008
Where AAAI
Authors Minmin Chen, Yixin Chen, Michael R. Brent
Comments (0)
books