Identifying Nominals with No Head Match Co-references Using Deep Learning

TL;DR

This paper introduces a novel neural network architecture that significantly improves the identification of nominals with no head match in coreference resolution, outperforming existing systems on the CoNLL 2012 dataset.

Contribution

It presents a new neural network model combining domain-specific embeddings and feature engineering, achieving state-of-the-art results in coreference resolution.

Findings

Outperforms previous models on CoNLL 2012 dataset

Increases F1 score by nearly 4% with additional features

Demonstrates effectiveness of combining submodels and features

Abstract

Identifying nominals with no head match is a long-standing challenge in coreference resolution with current systems performing significantly worse than humans. In this paper we present a new neural network architecture which outperforms the current state-of-the-art system on the English portion of the CoNLL 2012 Shared Task. This is done by using a logistic regression on features produced by two submodels, one of which is has the architecture proposed in [CM16a] while the other combines domain specific embeddings of the antecedent and the mention. We also propose some simple additional features which seem to improve performance for all models substantially, increasing F1 by almost 4% on basic logistic regression and other complex models.