Neural Machine Translation (NMT) models are known to suffer from noisy inputs. To make models robust, we generate adversarial augmentation samples that attack the model and preserve the source-side semantic meaning at the same time. To generate such samples, we propose a doubly-trained architecture that pairs two NMT models of opposite translation directions with a joint loss function, which combines the target-side attack and the source-side semantic similarity constraint. The results from our experiments across three different language pairs and two evaluation metrics show that these adversarial samples improve the model robustness.
@article{arxiv.2110.05691,
title = {Doubly-Trained Adversarial Data Augmentation for Neural Machine Translation},
author = {Weiting Tan and Shuoyang Ding and Huda Khayrallah and Philipp Koehn},
journal= {arXiv preprint arXiv:2110.05691},
year = {2021}
}