Graph Refinement for Coreference Resolution

TL;DR

This paper introduces a graph-based, document-level approach for coreference resolution that iteratively refines predictions to incorporate global context, leading to improved performance over traditional mention pair-wise models.

Contribution

It presents a novel graph refinement method that models coreference at the document level and predicts relationships non-autoregressively, enhancing resolution accuracy.

Findings

Improved coreference resolution performance over baselines

Document-level modeling enhances decision accuracy

Iterative refinement captures global dependencies

Abstract

The state-of-the-art models for coreference resolution are based on independent mention pair-wise decisions. We propose a modelling approach that learns coreference at the document-level and takes global decisions. For this purpose, we model coreference links in a graph structure where the nodes are tokens in the text, and the edges represent the relationship between them. Our model predicts the graph in a non-autoregressive manner, then iteratively refines it based on previous predictions, allowing global dependencies between decisions. The experimental results show improvements over various baselines, reinforcing the hypothesis that document-level information improves conference resolution.