BibSonomy :: publication ::

tag user group author concept BibTeX key search:all search:jnothman

The blue social bookmark and publication sharing system.

entry of jnothman:

(0)

This publication has not been reviewed yet.

rating distribution

average user rating

The average rating is computed over all reviews. However, some of them may be invisible to you due to the visibility setting chosen by the reviewers.

(0.0 of 5.0 based on 0 reviews)

Syntactic Dependency Based Heuristics for Biological Event Extraction

by: Halil Kilicoglu, and Sabine Bergler

In: Proceedings of the BioNLP 2009 Workshop Companion Volume for Shared Task Boulder, Colorado: Association for Computational Linguistics (June 2009) , p. 119--127.

Resources (URL, PDF, PS...)

URL:	http://www.aclweb.org/anthology/W09-1418

Rule-based event extraction from dependency parses. 3rd in BioNLP shared task 1, and 1st in task 3. Pipeline: identify trigger; identify participants; regulation events with sub-events are recursively checked for validity. Build event trigger dictionary based on training data, and calculated maximum likelihood scores, such that words would be ignored if below a threshold. Analysed dependency relationship paths (Stanford "collapsed" Parser) from trigger to entity, implementing 27 rules based on correctness and generality. [How might one learn do this automatically?] For "require" and "involve", rules are reversed. Special rule for hyphenated terms. Rules are applied in the order of simplicity. Special handling of coordination, apposition; also rules for speculation and negation detection.

BibTeX record

@inproceedings{kilicoglu2009,
  added-at = {2009-10-22T15:17:50.000+0200},
  address = {Boulder, Colorado},
  author = {Kilicoglu, Halil and Bergler, Sabine},
  biburl = {http://www.bibsonomy.org/bibtex/2d5c1b6f86e61c373f862129cebc9b531/jnothman},
  booktitle = {Proceedings of the BioNLP 2009 Workshop Companion Volume for Shared Task},
  description = {Rule-based event extraction from dependency parses. 3rd in BioNLP shared task 1, and 1st in task 3.

Pipeline: identify trigger; identify participants; regulation events with sub-events are recursively checked for validity.

Build event trigger dictionary based on training data, and calculated maximum likelihood scores, such that words would be ignored if below a threshold.

Analysed dependency relationship paths (Stanford "collapsed" Parser) from trigger to entity, implementing 27 rules based on correctness and generality. [How might one learn do this automatically?] For "require" and "involve", rules are reversed. Special rule for hyphenated terms. Rules are applied in the order of simplicity.

Special handling of coordination, apposition; also rules for speculation and negation detection.},
  interhash = {06a4e1b24252ea79010681afe45b6058},
  intrahash = {d5c1b6f86e61c373f862129cebc9b531},
  keywords = {biomedical bionlp09 dependency_parse event_extraction rule_based speculation_detection},
  month = {June},
  pages = {119--127},
  publisher = {Association for Computational Linguistics},
  timestamp = {2009-10-22T15:17:50.000+0200},
  title = {Syntactic Dependency Based Heuristics for Biological Event Extraction},
  url = {http://www.aclweb.org/anthology/W09-1418},
  year = 2009
}

Endnote record

BibSonomy :: publication ::

Syntactic Dependency Based Heuristics for Biological Event Extraction

Resources (URL, PDF, PS...)

Description

BibTeX record

Endnote record