Metric Learning in Multilingual Sentence Similarity Measurement for Document Alignment

TL;DR

This paper introduces a supervised metric learning approach for multilingual sentence similarity, improving document alignment accuracy across languages by fine-tuning distance measurements with parallel datasets.

Contribution

It proposes a novel supervised metric learning method for multilingual sentence similarity, outperforming unsupervised techniques in document alignment tasks.

Findings

Supervised metrics outperform unsupervised ones in multilingual document alignment.

The approach is effective across languages from different language families.

Results demonstrate improved accuracy in alignment tasks.

Abstract

Document alignment techniques based on multilingual sentence representations have recently shown state of the art results. However, these techniques rely on unsupervised distance measurement techniques, which cannot be fined-tuned to the task at hand. In this paper, instead of these unsupervised distance measurement techniques, we employ Metric Learning to derive task-specific distance measurements. These measurements are supervised, meaning that the distance measurement metric is trained using a parallel dataset. Using a dataset belonging to English, Sinhala, and Tamil, which belong to three different language families, we show that these task-specific supervised distance learning metrics outperform their unsupervised counterparts, for document alignment.