On Adversarial Examples for Text Classification by Perturbing Latent Representations

arXiv:2405.03789v1 Announce Type: new
Abstract: Recently, with the advancement of deep learning, several applications in text classification have advanced significantly. However, this improvement comes with a cost because deep learning is vulnerable to adversarial examples. This weakness indicates that deep learning is not very robust. Fortunately, the input of a text classifier is discrete. Hence, it can prevent the classifier from state-of-the-art attacks. Nonetheless, previous works have generated black-box attacks that successfully manipulate the discrete values of the input to …

abstract advanced advancement adversarial adversarial examples applications arxiv classification cost cs.ai cs.lg deep learning examples however improvement robust text text classification type vulnerable