Dynamic Conditional Random Fields: Factorized Probabilistic Models for Labeling and Segmenting Sequence Data

Charles Sutton, Khashayar Rohanimanesh, Andrew McCallum

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

In sequence modeling, we often wish to represent complex interaction between labels, such as when performing multiple, cascaded labeling tasks on the same sequence, or when long-range dependencies exist. We present dynamic conditional random fields (DCRFs), a generalization of linear-chain conditional random fields (CRFs) in which each time slice contains a set of state variables and edges---a distributed state representation as in dynamic Bayesian networks (DBNs)---and parameters are tied across slices. Since exact inference can be intractable in such models, we perform approximate inference using several schedules for belief propagation, including tree-based reparameterization (TRP). On a natural-language chunking task, we show that a DCRF performs better than a series of linear-chain CRFs, achieving comparable performance using only half the training data.
Original languageEnglish
Title of host publicationProceedings of the Twenty-first International Conference on Machine Learning
Place of PublicationNew York, NY, USA
PublisherACM
Number of pages8
ISBN (Print)1-58113-838-5
DOIs
Publication statusPublished - 2004

Fingerprint Dive into the research topics of 'Dynamic Conditional Random Fields: Factorized Probabilistic Models for Labeling and Segmenting Sequence Data'. Together they form a unique fingerprint.

Cite this