2025
ROME is Forged in Adversity: Robust Distilled Datasets via Information Bottleneck
Zheng Zhou, Wenquan Feng, Qiaosheng Zhang, Shuchang Lyu, Qi Zhao, Guangliang Cheng
International Conference on Machine Learning (ICML) 2025 Poster (Acceptance Rate: 26.9%, 3,260/12,107)
We introduce ROME, a method that enhances the adversarial robustness of dataset distillation by leveraging the information bottleneck principle, leading to significant improvements in robustness against both white-box and black-box attacks.
ROME is Forged in Adversity: Robust Distilled Datasets via Information Bottleneck
Zheng Zhou, Wenquan Feng, Qiaosheng Zhang, Shuchang Lyu, Qi Zhao, Guangliang Cheng
International Conference on Machine Learning (ICML) 2025 Poster (Acceptance Rate: 26.9%, 3,260/12,107)
We introduce ROME, a method that enhances the adversarial robustness of dataset distillation by leveraging the information bottleneck principle, leading to significant improvements in robustness against both white-box and black-box attacks.
2024
BEARD: Benchmarking the Adversarial Robustness for Dataset Distillation
Zheng Zhou, Wenquan Feng, Shuchang Lyu, Guangliang Cheng, Xiaowei Huang, Qi Zhao
Under review. 2024
BEARD is a unified benchmark for evaluating the adversarial robustness of dataset distillation methods, providing standardized metrics and tools to support reproducible research.
BEARD: Benchmarking the Adversarial Robustness for Dataset Distillation
Zheng Zhou, Wenquan Feng, Shuchang Lyu, Guangliang Cheng, Xiaowei Huang, Qi Zhao
Under review. 2024
BEARD is a unified benchmark for evaluating the adversarial robustness of dataset distillation methods, providing standardized metrics and tools to support reproducible research.
BACON: Bayesian Optimal Condensation Framework for Dataset Distillation
Zheng Zhou, Hongbo Zhao, Guangliang Cheng, Xiangtai Li, Shuchang Lyu, Wenquan Feng, Qi Zhao
Under review. 2024
This work presents BACON, the first Bayesian framework for Dataset Distillation, offering strong theoretical support to improve performance.
BACON: Bayesian Optimal Condensation Framework for Dataset Distillation
Zheng Zhou, Hongbo Zhao, Guangliang Cheng, Xiangtai Li, Shuchang Lyu, Wenquan Feng, Qi Zhao
Under review. 2024
This work presents BACON, the first Bayesian framework for Dataset Distillation, offering strong theoretical support to improve performance.
2023
MVPatch: More Vivid Patch for Adversarial Camouflaged Attacks on Object Detectors in the Physical World
Zheng Zhou, Hongbo Zhao, Ju Liu, Qiaosheng Zhang, Liwei Geng, Shuchang Lyu, Wenquan Feng
Under review. 2023
This work proposes MVPatch, an adversarial patch designed to be both highly transferable across detectors and visually stealthy.
MVPatch: More Vivid Patch for Adversarial Camouflaged Attacks on Object Detectors in the Physical World
Zheng Zhou, Hongbo Zhao, Ju Liu, Qiaosheng Zhang, Liwei Geng, Shuchang Lyu, Wenquan Feng
Under review. 2023
This work proposes MVPatch, an adversarial patch designed to be both highly transferable across detectors and visually stealthy.
2022
Adversarial Examples Are Closely Relevant to Neural Network Models - A Preliminary Experiment Explore
Zheng Zhou, Ju Liu, Yanyang Han
International Conference on Swarm Intelligence (ICSI) 2022
This study explores how adversarial examples impact neural networks by analyzing their sensitivity to different architectures, activation functions, and loss functions.
Adversarial Examples Are Closely Relevant to Neural Network Models - A Preliminary Experiment Explore
Zheng Zhou, Ju Liu, Yanyang Han
International Conference on Swarm Intelligence (ICSI) 2022
This study explores how adversarial examples impact neural networks by analyzing their sensitivity to different architectures, activation functions, and loss functions.