“That Is a Suspicious Reaction!”: Interpreting Logits Variation to Detect NLP Adversarial Attacks
E. Mosca, Shreyash Agarwal, Javier Rando Ramírez, Georg Groh · Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2022
Adversarial attacks are a major challenge faced by current machine learning research.These purposely crafted inputs fool even the most advanced models, precluding their deployment in safety-critical applications.Extensive research in computer vision has been carried to develop reliable defense strategies.However, the same issue remains less explored in natural language processing.Our work presents a model-agnostic detector of adversarial text examples.The approach identifies patterns in the logits of the target classifier when perturbing the input text.The proposed detector improves the current state-ofthe-art performance in recognizing adversarial inputs and exhibits strong generalization capabilities across different NLP models, datasets, and word-level attacks.