Multilingual Representation Distillation with Contrastive Learning

TL;DR

This paper introduces a contrastive learning approach for multilingual representation distillation, significantly improving the quality of cross-lingual sentence similarity tasks, especially for low-resource languages.

Contribution

It presents a novel contrastive learning-based method for multilingual representation distillation, enhancing semantic similarity estimation across languages.

Findings

Outperforms LASER, LASER3, and LaBSE in low-resource language tasks

Improves multilingual similarity search accuracy

Effective for corpus filtering and translation quality estimation

Abstract

Multilingual sentence representations from large models encode semantic information from two or more languages and can be used for different cross-lingual information retrieval and matching tasks. In this paper, we integrate contrastive learning into multilingual representation distillation and use it for quality estimation of parallel sentences (i.e., find semantically similar sentences that can be used as translations of each other). We validate our approach with multilingual similarity search and corpus filtering tasks. Experiments across different low-resource languages show that our method greatly outperforms previous sentence encoders such as LASER, LASER3, and LaBSE.