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

MASTERKEY: Automated Jailbreaking of Large Language Model Chatbots

Gelei Deng (Nanyang Technological University), Yi Liu (Nanyang Technological University), Yuekang Li (University of New South Wales), Kailong Wang (Huazhong University of Science and Technology), Ying Zhang (Virginia Tech), Zefeng Li (Nanyang Technological University), Haoyu Wang (Huazhong University of Science and Technology), Tianwei Zhang (Nanyang Technological University), Yang Liu (Nanyang Technological University)

Read More

Don't Interrupt Me – A Large-Scale Study of On-Device...

Marian Harbach (Google), Igor Bilogrevic (Google), Enrico Bacis (Google), Serena Chen (Google), Ravjit Uppal (Google), Andy Paicu (Google), Elias Klim (Google), Meggyn Watkins (Google), Balazs Engedy (Google)

Read More

Cyclops: Binding a Vehicle’s Digital Identity to its Physical...

Lewis William Koplon, Ameer Ghasem Nessaee, Alex Choi (University of Arizona, Tucson), Andres Mentoza (New Mexico State University, Las Cruces), Michael Villasana, Loukas Lazos, Ming Li (University of Arizona, Tucson)

Read More

GTrans: Graph Transformer-Based Obfuscation-resilient Binary Code Similarity Detection

Yun Zhang (Hunan University), Yuling Liu (Hunan University), Ge Cheng (Xiangtan University), Bo Ou (Hunan University)

Read More