Cross-Attention is All You Need: Adapting Pretrained Transformers for Machine Translation

Abstract

We study the power of cross-attention in the Transformer architecture within the context of transfer learning for machine translation, and extend the findings of studies into cross-attention when training from scratch. We conduct a series of experiments through fine-tuning a translation model on data where either the source or target language has changed. These experiments reveal that fine-tuning only the cross-attention parameters is nearly as effective as fine-tuning all parameters (i.e., the entire translation model). We provide insights into why this is the case and observe that limiting fine-tuning in this manner yields cross-lingually aligned embeddings. The implications of this finding for researchers and practitioners include a mitigation of catastrophic forgetting, the potential for zero-shot translation, and the ability to extend machine translation models to several new language pairs with reduced parameter storage overhead.

arXiv commentAccepted to EMNLP 2021 Main Conference

References in corpus (6)

• Exploiting Similarities among Languages for Machine Translation
• On the Cross-lingual Transferability of Monolingual Representations
• Multilingual Denoising Pre-training for Neural Machine Translation
• Transfer Learning across Low-Resource, Related Languages for Neural Machine Translation
• Pretrained Transformers as Universal Computation Engines
• A Universal Parent Model for Low-Resource Neural Machine Translation Transfer