Training Conditional Random Fields using Virtual Evidence Boosting
Proc. of the International Joint Conference on
Artificial Intelligence (IJCAI), 2007
Abstract
While conditional random fields (CRFs) have been
applied successfully in a variety of domains, their
training remains a challenging task. In this paper,
we introduce a novel training method for CRFs,
called virtual evidence boosting, which simulta-
neously performs feature selection and parameter
estimation. To achieve this, we extend standard
boosting to handle virtual evidence, where an ob-
servation can be specified as a distribution rather
than a single number. This extension allows us to
develop a unified framework for learning both local
and compatibility features in CRFs. In experiments
on synthetic data as well as real activity classifi-
cation problems, our new training algorithm out-
performs other training approaches including max-
imum likelihood, maximum pseudo-likelihood, and
the most recent boosted random fields.