Towards Robust Neural Machine Translation

TL;DR

This paper introduces adversarial stability training to enhance the robustness of neural machine translation models against input perturbations, leading to improved translation quality and stability across multiple language pairs.

Contribution

It proposes a novel adversarial stability training method that makes NMT models more resilient to input perturbations, improving both robustness and translation performance.

Findings

Significant improvements over strong NMT baselines

Enhanced robustness against input perturbations

Effective across multiple language pairs

Abstract

Small perturbations in the input can severely distort intermediate representations and thus impact translation quality of neural machine translation (NMT) models. In this paper, we propose to improve the robustness of NMT models with adversarial stability training. The basic idea is to make both the encoder and decoder in NMT models robust against input perturbations by enabling them to behave similarly for the original input and its perturbed counterpart. Experimental results on Chinese-English, English-German and English-French translation tasks show that our approaches can not only achieve significant improvements over strong NMT systems but also improve the robustness of NMT models.