Contributions of Transformer Attention Heads in Multi- and Cross-lingual Tasks (2108.08375v1)

Abstract: This paper studies the relative importance of attention heads in Transformer-based models to aid their interpretability in cross-lingual and multi-lingual tasks. Prior research has found that only a few attention heads are important in each mono-lingual NLP task and pruning the remaining heads leads to comparable or improved performance of the model. However, the impact of pruning attention heads is not yet clear in cross-lingual and multi-lingual tasks. Through extensive experiments, we show that (1) pruning a number of attention heads in a multi-lingual Transformer-based model has, in general, positive effects on its performance in cross-lingual and multi-lingual tasks and (2) the attention heads to be pruned can be ranked using gradients and identified with a few trial experiments. Our experiments focus on sequence labeling tasks, with potential applicability on other cross-lingual and multi-lingual tasks. For comprehensiveness, we examine two pre-trained multi-lingual models, namely multi-lingual BERT (mBERT) and XLM-R, on three tasks across 9 languages each. We also discuss the validity of our findings and their extensibility to truly resource-scarce languages and other task settings.