Understanding and Mitigating Uncertainty in Zero-Shot Translation

Anonymous

Abstract: Zero-shot translation is a promising direction for building a comprehensive multilingual neural machine translation~(MNMT) system. However, its quality is still not satisfactory due to off-target issues. In this paper, we aim to understand and alleviate the off-target issues from the perspective of uncertainty in zero-shot translation. By carefully examining the translation output and model confidence, we identify two uncertainties that are responsible for the off-target issues, namely, extrinsic data uncertainty and intrinsic model uncertainty. Based on the observations, we propose two lightweight and complementary approaches to denoise the training data for model training and explicitly penalize the off-target translations during model training. Extensive experiments on both balanced and imbalanced datasets show that our approaches significantly improve the performance of zero-shot translation over strong MNMT baselines. Qualitative analyses provide insights into where our approaches reduce off-target translations.

Paper Type: long

Research Area: Machine Translation