A Survey of Adversarial Defences and Robustness in NLP

Published in "ACM Computing Surveys"
Shreya Goyal , Sumanth Doddapaneni , Mitesh M. Khapra , Balaraman Ravindran

In the recent years, it has been seen that deep neural networks are lacking robustness and are vulnerable in case of adversarial perturbations in input data. Strong adversarial attacks are proposed by various authors for tasks under computer vision and Natural Language Processing (NLP). As a counter-effort, several defense mechanisms are also proposed to save these networks from failing. Defending the neural networks from adversarial attacks has its own importance, where the goal is to ensure that the model’s prediction doesn’t change if input data is perturbed. Numerous methods for adversarial defense in NLP are proposed of late, for different NLP tasks such as text classification, named entity recognition, natural language inferencing, etc. Some of these methods are not just used for defending neural networks from adversarial attacks, but also used as a regularization mechanism during training, saving the model from overfitting. The proposed survey is an attempt to review different methods proposed for adversarial defenses in NLP in recent years by proposing a novel taxonomy. This survey also highlights the fragility of the advanced deep neural networks in NLP and the challenges in defending them.