Adapting Monolingual Models: Data can be Scarce when Language Similarity is High

TL;DR

This paper explores how high language similarity enables effective zero-shot transfer learning for low-resource languages using minimal data, by retraining lexical layers of BERT models.

Contribution

It demonstrates that high language similarity allows substantial transfer performance with only 10MB of data by retraining lexical layers and fine-tuning Transformer layers.

Findings

High language similarity enables effective transfer with minimal data.

Retraining lexical layers improves monolingual model performance.

Monolingual BERT models outperform multilingual BERT in low-resource scenarios.

Abstract

For many (minority) languages, the resources needed to train large models are not available. We investigate the performance of zero-shot transfer learning with as little data as possible, and the influence of language similarity in this process. We retrain the lexical layers of four BERT-based models using data from two low-resource target language varieties, while the Transformer layers are independently fine-tuned on a POS-tagging task in the model's source language. By combining the new lexical layers and fine-tuned Transformer layers, we achieve high task performance for both target languages. With high language similarity, 10MB of data appears sufficient to achieve substantial monolingual transfer performance. Monolingual BERT-based models generally achieve higher downstream task performance after retraining the lexical layer than multilingual BERT, even when the target language is included in the multilingual model.