Exploring Fine-tuning Techniques for Pre-trained Cross-lingual Models via Continual Learning

Abstract

Recently, fine-tuning pre-trained language models (e.g., multilingual BERT) to downstream cross-lingual tasks has shown promising results. However, the fine-tuning process inevitably changes the parameters of the pre-trained model and weakens its cross-lingual ability, which leads to sub-optimal performance. To alleviate this problem, we leverage continual learning to preserve the original cross-lingual ability of the pre-trained model when we fine-tune it to downstream tasks. The experimental result shows that our fine-tuning methods can better preserve the cross-lingual ability of the pre-trained model in a sentence retrieval task. Our methods also achieve better performance than other fine-tuning baselines on the zero-shot cross-lingual part-of-speech tagging and named entity recognition tasks.

References in corpus (5)

• Word Translation Without Parallel Data
• Unicoder: A Universal Language Encoder by Pre-training with Multiple Cross-lingual Tasks
• XGLUE: A New Benchmark Dataset for Cross-lingual Pre-training, Understanding and Generation
• Attention-Informed Mixed-Language Training for Zero-shot Cross-lingual Task-oriented Dialogue Systems
• Coach: A Coarse-to-Fine Approach for Cross-domain Slot Filling

Cited by in corpus (3)

• mT5: A massively multilingual pre-trained text-to-text transformer
• Cross-lingual Spoken Language Understanding with Regularized Representation Alignment
• Soft Layer Selection with Meta-Learning for Zero-Shot Cross-Lingual Transfer