Coreference Resolution for Vietnamese Texts Towards Relation Extraction Applications

EasyChair Preprint 10099

Abstract

Relation extraction is one of the important tasks in natural language processing. However, the performance of this task is greatly influenced by the performance of coreference resolution, which is the task of identifying different mentions of the same entity. This paper presents a method for coreference resolution for Vietnamese texts that takes advantage of the available coreference models for the English language. Our proposed method combines a translation model and a word alignment model. The experimental results proved that the proposed method using a word-level dataset is effective, with F1 scores of 66.5%, 82.7% and 76.3% in MUC, B³, and CEAF respectively.

Keyphrases: Relation Extraction, Vietnamese language, coreference resolution

@booklet{EasyChair:10099,
  author    = {Man Minh Pham and Ngan Luu-Thuy Nguyen},
  title     = {Coreference Resolution for Vietnamese Texts Towards Relation Extraction Applications},
  howpublished = {EasyChair Preprint 10099},
  year      = {EasyChair, 2023}}