How effective is Multi-source pivoting for Translation of Low Resource Indian Languages?

TL;DR

This paper investigates multi-source pivoting for translating low-resource Indian languages, showing marginal improvements over existing methods, which can be boosted with synthetic data, indicating potential for future research in low-resource MT.

Contribution

It explores multi-source pivoting using both source and pivot sentences for low-resource Indian language translation, highlighting its limited but improvable effectiveness.

Findings

Multi-source pivoting yields marginal improvements over state-of-the-art.

Synthetic target data can enhance translation quality.

Multi-source pivoting remains a promising approach despite limited gains.

Abstract

Machine Translation (MT) between linguistically dissimilar languages is challenging, especially due to the scarcity of parallel corpora. Prior works suggest that pivoting through a high-resource language can help translation into a related low-resource language. However, existing works tend to discard the source sentence when pivoting. Taking the case of English to Indian language MT, this paper explores the 'multi-source translation' approach with pivoting, using both source and pivot sentences to improve translation. We conducted extensive experiments with various multi-source techniques for translating English to Konkani, Manipuri, Sanskrit, and Bodo, using Hindi, Marathi, and Bengali as pivot languages. We find that multi-source pivoting yields marginal improvements over the state-of-the-art, contrary to previous claims, but these improvements can be enhanced with synthetic target language data. We believe multi-source pivoting is a promising direction for Low-resource translation.