TY - GEN
T1 - Using Prediction from Sentential Scope to Build a Pseudo Co-Testing Learner for Event Extraction
AU - Liao, Shasha
AU - Grishman, Ralph
N1 - Publisher Copyright:
© 2011 AFNLP
PY - 2011
Y1 - 2011
N2 - Event extraction involves the identification of instances of a type of event, along with their attributes and participants. Developing a training corpus by annotating events in text is very labor intensive, and so selecting informative instances to annotate can save a great deal of manual work. We present an active learning (AL) strategy, pseudo co-testing, based on one view from a classifier aiming to solve the original problem of event extraction, and another view from a classifier aiming to solve a coarser granularity task. As the second classifier can provide more graded matching from a wider scope, we can build a set of pseudo-contention-points which are very informative, and can speed up the AL process. Moreover, we incorporate multiple selection criteria into the pseudo co-testing, seeking training examples that are informative, representative, and varied. Experiments show that pseudo co-testing can reduce annotation labor by 81%; incorporating multiple selection criteria reduces the labor by a further 7%.
AB - Event extraction involves the identification of instances of a type of event, along with their attributes and participants. Developing a training corpus by annotating events in text is very labor intensive, and so selecting informative instances to annotate can save a great deal of manual work. We present an active learning (AL) strategy, pseudo co-testing, based on one view from a classifier aiming to solve the original problem of event extraction, and another view from a classifier aiming to solve a coarser granularity task. As the second classifier can provide more graded matching from a wider scope, we can build a set of pseudo-contention-points which are very informative, and can speed up the AL process. Moreover, we incorporate multiple selection criteria into the pseudo co-testing, seeking training examples that are informative, representative, and varied. Experiments show that pseudo co-testing can reduce annotation labor by 81%; incorporating multiple selection criteria reduces the labor by a further 7%.
UR - http://www.scopus.com/inward/record.url?scp=84876786846&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84876786846&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:84876786846
T3 - IJCNLP 2011 - Proceedings of the 5th International Joint Conference on Natural Language Processing
SP - 714
EP - 722
BT - IJCNLP 2011 - Proceedings of the 5th International Joint Conference on Natural Language Processing
A2 - Wang, Haifeng
A2 - Yarowsky, David
PB - Association for Computational Linguistics (ACL)
T2 - 5th International Joint Conference on Natural Language Processing, IJCNLP 2011
Y2 - 8 November 2011 through 13 November 2011
ER -