BART goes multilingual: the UniTN/Essex submission to the CoNLL-2012 shared task

  • Authors:
  • Olga Uryupina;Alessandro Moschitti;Massimo Poesio

  • Affiliations:
  • University of Trento;University of Trento;University of Trento and University of Essex

  • Venue:
  • CoNLL '12 Joint Conference on EMNLP and CoNLL - Shared Task
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper describes the UniTN/Essex submission to the CoNLL-2012 Shared Task on the Multilingual Coreference Resolution. We have extended our CoNLL-2011 submission, based on BART, to cover two additional languages, Arabic and Chinese. This paper focuses on adapting BART to new languages, discussing the problems we have encountered and the solutions adopted. In particular, we propose a novel entity-mention detection algorithm that might help identify nominal mentions in an unknown language. We also discuss the impact of basic linguistic information on the overall performance level of our coreference resolution system.