Choosing Transfer Languages for Cross-Lingual Learning

TL;DR

This paper develops a model to automatically rank and select optimal transfer languages for cross-lingual NLP tasks, outperforming ad hoc methods by considering multiple linguistic and data features.

Contribution

It introduces a ranking model that predicts effective transfer languages by integrating various linguistic and data features, improving over heuristic selection methods.

Findings

Model predicts transfer languages better than baselines.

Features like phylogenetic similarity and data size are most informative.

Insights guide future manual selection of transfer languages.

Abstract

Cross-lingual transfer, where a high-resource transfer language is used to improve the accuracy of a low-resource task language, is now an invaluable tool for improving performance of natural language processing (NLP) on low-resource languages. However, given a particular task language, it is not clear which language to transfer from, and the standard strategy is to select languages based on ad hoc criteria, usually the intuition of the experimenter. Since a large number of features contribute to the success of cross-lingual transfer (including phylogenetic similarity, typological properties, lexical overlap, or size of available data), even the most enlightened experimenter rarely considers all these factors for the particular task at hand. In this paper, we consider this task of automatically selecting optimal transfer languages as a ranking problem, and build models that consider the aforementioned features to perform this prediction. In experiments on representative NLP tasks, we demonstrate that our model predicts good transfer languages much better than ad hoc baselines considering single features in isolation, and glean insights on what features are most informative for each different NLP tasks, which may inform future ad hoc selection even without use of our method. Code, data, and pre-trained models are available at https://github.com/neulab/langrank