Cognate-aware morphological segmentation for multilingual neural translation

TL;DR

This paper introduces Cognate Morfessor, a multilingual morphological segmentation method that enhances neural translation quality for closely related languages by improving segmentation consistency, especially benefiting resource-scarce languages like Estonian.

Contribution

The paper presents Cognate Morfessor, a novel multilingual segmentation approach that leverages cognate information to improve translation accuracy in neural models.

Findings

Improved translation quality for Estonian.

Enhanced morphological segmentation consistency.

Better handling of cognates and loanwords.

Abstract

This article describes the Aalto University entry to the WMT18 News Translation Shared Task. We participate in the multilingual subtrack with a system trained under the constrained condition to translate from English to both Finnish and Estonian. The system is based on the Transformer model. We focus on improving the consistency of morphological segmentation for words that are similar orthographically, semantically, and distributionally; such words include etymological cognates, loan words, and proper names. For this, we introduce Cognate Morfessor, a multilingual variant of the Morfessor method. We show that our approach improves the translation quality particularly for Estonian, which has less resources for training the translation model.