Enhancing Entity Aware Machine Translation with Multi-task Learning

TL;DR

This paper introduces a multi-task learning approach that jointly optimizes entity recognition and machine translation to improve entity-aware translation performance, addressing data scarcity and contextual complexity.

Contribution

It presents a novel multi-task learning framework specifically designed for enhancing entity-aware machine translation by integrating recognition and translation tasks.

Findings

Improved translation accuracy on SemEval 2025 dataset

Effective joint training of recognition and translation tasks

Demonstrated benefits over baseline models

Abstract

Entity-aware machine translation (EAMT) is a complicated task in natural language processing due to not only the shortage of translation data related to the entities needed to translate but also the complexity in the context needed to process while translating those entities. In this paper, we propose a method that applies multi-task learning to optimize the performance of the two subtasks named entity recognition and machine translation, which improves the final performance of the Entity-aware machine translation task. The result and analysis are performed on the dataset provided by the organizer of Task 2 of the SemEval 2025 competition.