Paper: Active Learning for Coreference Resolution

ACL ID W12-2409
Title Active Learning for Coreference Resolution
Venue Workshop on Biomedical Natural Language Processing
Year 2012

Active learning can lower the cost of anno- tation for some natural language processing tasks by using a classifier to select informa- tive instances to send to human annotators. It has worked well in cases where the training in- stances are selected one at a time and require minimal context for annotation. However, coreference annotations often require some context and the traditional active learning ap- proach may not be feasible. In this work we explore various active learning methods for coreference resolution that fit more realisti- cally into coreference annotation workflows.