TWINFUZZ: Dual-Model Fuzzing for Robustness Generalization in Deep Learning

Deep learning (DL) models are increasingly deployed in safety-critical applications such as face recognition, autonomous driving, and medical diagnosis. Despite their impressive accuracy, they remain vulnerable to adversarial examples - subtle perturbations that can cause incorrect predictions, i.e., the robustness issues. While adversarial training improves robustness against known attacks, it often fails to generalize to unseen or stronger threats, revealing a critical gap in robustness generalization. In this work, we propose a dual-model fuzzing framework to enhance generalized robustness in DL models. Central to our method is a lightweight metric, the Lagrangian Information Bottleneck (LIB), which guides entropy-based mutation toward semantically meaningful and high-risk regions of the input space. The executor uses a resistant model and a more error-prone vulnerable model; their prediction consistency forms the basis of agreement mining, a label-free oracle for isolating decision-boundary samples. To ensure fuzzing effectiveness, we further introduce a task-driven seed selection strategy (e.g., SSIM for vision) that filters out low-quality inputs. We implement a prototype, TWINFUZZ, and evaluate it on six benchmark datasets and nine DL models. Compared with state-of-the-art testing approaches, TWINFUZZ achieves superior improvements in both training-specific and generalized robustness.

Authors

• Enze Dai Shenzhen International Graduate School, Tsinghua University
• Wentao Mo Wangxuan Institute of Computer Technology, Peking University
• Kun Hu The University of Sydney
• Xiaogang Zhu The University of Adelaide
• Xi Xiao Tsinghua Shenzhen International Graduate School, Tsinghua University, Shenzhen, China Peng Cheng Laboratory, Shenzhen, China Key Laboratory of Cyberspace Security, Ministry of Education of China, Zhengzhou, China
• Sheng Wen Swinburne University of Technology
• Shaohua Wang Central University of Finance and Economics
• Yang Xiang Xiaobing.ai

Keywords

No keywords are indexed for this paper.