Generating Natural Language Adversarial Examples on a Large Scale with Generative Models
This addresses the vulnerability of text classifiers to adversarial attacks, enabling large-scale generation of deceptive texts, though it is incremental as it builds on existing generative techniques.
The paper tackles the problem of generating diverse and natural adversarial examples for text classification by proposing an unrestricted generation method using generative models, achieving a higher attack success rate than existing methods while maintaining acceptable human readability.
Today text classification models have been widely used. However, these classifiers are found to be easily fooled by adversarial examples. Fortunately, standard attacking methods generate adversarial texts in a pair-wise way, that is, an adversarial text can only be created from a real-world text by replacing a few words. In many applications, these texts are limited in numbers, therefore their corresponding adversarial examples are often not diverse enough and sometimes hard to read, thus can be easily detected by humans and cannot create chaos at a large scale. In this paper, we propose an end to end solution to efficiently generate adversarial texts from scratch using generative models, which are not restricted to perturbing the given texts. We call it unrestricted adversarial text generation. Specifically, we train a conditional variational autoencoder (VAE) with an additional adversarial loss to guide the generation of adversarial examples. Moreover, to improve the validity of adversarial texts, we utilize discrimators and the training framework of generative adversarial networks (GANs) to make adversarial texts consistent with real data. Experimental results on sentiment analysis demonstrate the scalability and efficiency of our method. It can attack text classification models with a higher success rate than existing methods, and provide acceptable quality for humans in the meantime.