Improving Coreference Resolution by Learning Entity-Level Distributed Representations
This work addresses a core problem in natural language processing for tasks like text understanding, offering a novel approach that advances the field beyond incremental improvements.
The paper tackles the challenge of incorporating entity-level information in coreference resolution by introducing a neural network system that learns vector representations for coreference clusters and uses a learning-to-search algorithm to optimize cluster merges. It achieves substantial improvements, outperforming the state-of-the-art on the CoNLL 2012 Shared Task datasets for English and Chinese with minimal hand-engineered features.
A long-standing challenge in coreference resolution has been the incorporation of entity-level information - features defined over clusters of mentions instead of mention pairs. We present a neural network based coreference system that produces high-dimensional vector representations for pairs of coreference clusters. Using these representations, our system learns when combining clusters is desirable. We train the system with a learning-to-search algorithm that teaches it which local decisions (cluster merges) will lead to a high-scoring final coreference partition. The system substantially outperforms the current state-of-the-art on the English and Chinese portions of the CoNLL 2012 Shared Task dataset despite using few hand-engineered features.