Lujia Shen (Zhejiang University), Yuwen Pu (Zhejiang University), Shouling Ji (Zhejiang University), Changjiang Li (Penn State), Xuhong Zhang (Zhejiang University), Chunpeng Ge (Shandong University), Ting Wang (Penn State)

Transformer-based models, such as BERT and GPT, have been widely adopted in natural language processing (NLP) due to their exceptional performance. However, recent studies show their vulnerability to textual adversarial attacks where the model's output can be misled by intentionally manipulating the text inputs. Despite various methods that have been proposed to enhance the model's robustness and mitigate this vulnerability, many require heavy consumption resources (e.g., adversarial training) or only provide limited protection (e.g., defensive dropout).

In this paper, we propose a novel method called dynamic attention, tailored for the transformer architecture, to enhance the inherent robustness of the model itself against various adversarial attacks. Our method requires no downstream task knowledge and does not incur additional costs. The proposed dynamic attention consists of two modules: *(i) attention rectification*, which masks or weakens the attention value of the chosen tokens, and *(ii) dynamic modeling*, which dynamically builds the set of candidate tokens. Extensive experiments demonstrate that dynamic attention significantly mitigates the impact of adversarial attacks, improving up to 33% compared to previous methods against widely used adversarial attacks. The model-level design of dynamic attention enables it to be easily combined with other defense methods (e.g., adversarial training) to further enhance the model's robustness. Furthermore, we demonstrate that dynamic attention preserves the state-of-the-art robustness space of the original model compared to other dynamic modeling methods.

View More Papers

Vision: Towards Fully Shoulder-Surfing Resistant and Usable Authentication for...

Tobias Länge (Karlsruhe Institute of Technology), Philipp Matheis (Karlsruhe Institute of Technology), Reyhan Düzgün (Ruhr University Bochum), Melanie Volkamer (Karlsruhe Institute of Technology), Peter Mayer (Karlsruhe Institute of Technology, University of Southern Denmark)

Read More

A Two-Layer Blockchain Sharding Protocol Leveraging Safety and Liveness...

Yibin Xu (University of Copenhagen), Jingyi Zheng (University of Copenhagen), Boris Düdder (University of Copenhagen), Tijs Slaats (University of Copenhagen), Yongluan Zhou (University of Copenhagen)

Read More

Resilient Routing for Low Earth Orbit Mega-Constellation Networks

Alexander Kedrowitsch (Virginia Tech), Jonathan Black (Virginia Tech) Daphne Yao (Virginia Tech)

Read More

DEMASQ: Unmasking the ChatGPT Wordsmith

Kavita Kumari (Technical University of Darmstadt, Germany), Alessandro Pegoraro (Technical University of Darmstadt), Hossein Fereidooni (Technische Universität Darmstadt), Ahmad-Reza Sadeghi (Technical University of Darmstadt)

Read More