Optimizing to arbitrary NLP metrics using ensemble selection
HLT '05 Proceedings of the conference on Human Language Technology and Empirical Methods in Natural Language Processing
Maximum metric score training for coreference resolution
COLING '10 Proceedings of the 23rd International Conference on Computational Linguistics
Hi-index | 0.00 |
Human speakers generally have no difficulty in determining which noun phrases in a text or dialogue refer to the same real-world entity. This task of identifying co-referring noun phrases—noun phrase coreference resolution—can present a serious challenge to a natural language processing system, however. Indeed, it is one of the critical problems that currently limits the performance for many practical natural language processing tasks. State-of-the-art coreference resolution systems operate by relying on a set of hand-crafted heuristics that requires a lot of time and linguistic expertise to develop. Recently, machine learning techniques have been used to circumvent both of these problems by automating the acquisition of coreference resolution heuristics, yielding coreference systems that offer performance comparable to their heuristic-based counterparts. In this dissertation, we present a machine learning-based solution to noun phrase coreference that extends earlier work in the area and out-performs the best existing learning-based coreference engine on a suite of standard coreference data sets. Performance gains accrue from more effective use of the available training data via a set of linguistic and extra-linguistic extensions to the standard machine learning framework for coreference resolution.