Joint Coreference Resolution for Zeros and non-Zeros in Arabic

TL;DR

This paper introduces two architectures for jointly resolving zero and non-zero coreferences in Arabic, addressing the lack of datasets and the differences between argument types, and evaluates them on a newly annotated dataset.

Contribution

It presents the first joint coreference resolution models for zeros and non-zeros in Arabic, along with a new annotated dataset for this task.

Findings

Two architectures successfully jointly resolve AZPs and non-AZPs in Arabic.

Created a new annotated Arabic coreference dataset including zeros and non-zeros.

Demonstrated the effectiveness of joint resolution over separate models.

Abstract

Most existing proposals about anaphoric zero pronoun (AZP) resolution regard full mention coreference and AZP resolution as two independent tasks, even though the two tasks are clearly related. The main issues that need tackling to develop a joint model for zero and non-zero mentions are the difference between the two types of arguments (zero pronouns, being null, provide no nominal information) and the lack of annotated datasets of a suitable size in which both types of arguments are annotated for languages other than Chinese and Japanese. In this paper, we introduce two architectures for jointly resolving AZPs and non-AZPs, and evaluate them on Arabic, a language for which, as far as we know, there has been no prior work on joint resolution. Doing this also required creating a new version of the Arabic subset of the standard coreference resolution dataset used for the CoNLL-2012 shared task (Pradhan et al.,2012) in which both zeros and non-zeros are included in a single dataset.