Piotr Gaiński


2023

pdf bib
Step by Step Loss Goes Very Far: Multi-Step Quantization for Adversarial Text Attacks
Piotr Gaiński | Klaudia Bałazy
Proceedings of the 17th Conference of the European Chapter of the Association for Computational Linguistics

We propose a novel gradient-based attack against transformer-based language models that searches for an adversarial example in a continuous space of tokens probabilities. Our algorithm mitigates the gap between adversarial loss for continuous and discrete text representations by performing multi-step quantization in a quantization-compensation loop. Experiments show that our method significantly outperforms other approaches on various natural language processing (NLP) tasks.
Search
Co-authors
Venues