Incorporating non-local information into information extraction systems by Gibbs sampling

DMCA

Incorporating non-local information into information extraction systems by Gibbs sampling (2005)

Cached

Download Links

Other Repositories/Bibliography

DBLP

by Jenny Rose Finkel , Trond Grenager , Christopher Manning

Venue:	IN ACL
Citations:	695 - 25 self

BibTeX

@INPROCEEDINGS{Finkel05incorporatingnon-local,
    author = {Jenny Rose Finkel and Trond Grenager and Christopher Manning},
    title = {Incorporating non-local information into information extraction systems by Gibbs sampling},
    booktitle = {IN ACL},
    year = {2005},
    pages = {363--370},
    publisher = {}
}

OpenURL

Abstract

Most current statistical natural language processing models use only local features so as to permit dynamic programming in inference, but this makes them unable to fully account for the long distance structure that is prevalent in language use. We show how to solve this dilemma with Gibbs sampling, a simple Monte Carlo method used to perform approximate inference in factored probabilistic models. By using simulated annealing in place of Viterbi decoding in sequence models such as HMMs, CMMs, and CRFs, it is possible to incorporate non-local structure while preserving tractable inference. We use this technique to augment an existing CRF-based information extraction system with long-distance dependency models, enforcing label consistency and extraction template consistency constraints. This technique results in an error reduction of up to 9 % over state-of-the-art systems on two established information extraction tasks.

Keyphrases

gibbs sampling non-local information information extraction system label consistency crf-based information extraction system state-of-the-art system language use simple monte carlo method extraction template consistency constraint long distance structure long-distance dependency model simulated annealing factored probabilistic model tractable inference approximate inference sequence model dynamic programming technique result information extraction task error reduction non-local structure local feature

DMCA