Iterative Nearest Neighbour Machine Translation for Unsupervised Domain Adaptation

Unsupervised domain adaptation of machine translation, which adapts a pre-trained translation model to a specific domain without indomain parallel data, has drawn extensive attention in recent years. However, most existing methods focus on the fine-tuning based techniques, which is non-extensible. In this paper, we propose a new method to perform unsupervised domain adaptation in a non-parametric manner. Employing only indomain monolingual data, this method jointly perform nearest neighbour inference on both forward and backward translation directions. The forward translation model creates nearest neighbour datastore for the backward direction, and vice versa, strengthening each other in an iterative style. Experiments on multidomain datasets demonstrate that our method significantly improves the in-domain translation performance and achieves state-of-the-art results among non-parametric methods.