Soft Layer Selection with Meta-Learning for Zero-Shot Cross-Lingual Transfer

TL;DR

This paper introduces a meta-learning approach to selectively freeze layers of multilingual models during fine-tuning, enhancing zero-shot cross-lingual transfer performance.

Contribution

It proposes a novel meta-optimizer that dynamically determines which layers to freeze, improving transfer effectiveness over standard fine-tuning methods.

Findings

Improved zero-shot transfer accuracy on natural language inference tasks.

Outperforms baseline fine-tuning and existing meta-learning approaches.

Demonstrates the effectiveness of layer-wise selection in multilingual models.

Abstract

Multilingual pre-trained contextual embedding models (Devlin et al., 2019) have achieved impressive performance on zero-shot cross-lingual transfer tasks. Finding the most effective fine-tuning strategy to fine-tune these models on high-resource languages so that it transfers well to the zero-shot languages is a non-trivial task. In this paper, we propose a novel meta-optimizer to soft-select which layers of the pre-trained model to freeze during fine-tuning. We train the meta-optimizer by simulating the zero-shot transfer scenario. Results on cross-lingual natural language inference show that our approach improves over the simple fine-tuning baseline and X-MAML (Nooralahzadeh et al., 2020).