Event schedules / ICONIP 2026

Trustworthy AI & Security 1: Adversarial Attacks, Backdoors & Robustness

Tuesday, November 24, 2026, – (Australia/Sydney)
Clarendon Room D · Conference

Paper 159 - SASP: Semantic-Aware Structural Perturbation for Transferable Adversarial Attacks
Authors: Si Bing Guo, Jing Zhu, Qin Yan and Zhao Ma

Paper 226 - Dynamic Frequency-Semantic Guidance for Diffusion-Based Adversarial Purification
Authors: Yanshuo Zhang, Guangxi Lu, Jinchuan Zhang, Lei Gao, Chong Mu and Lizong Zhang

Paper 356 - DiffAttack-ME: Energy-Guided Attention Deception for Diffusion-Based Transferable Attacks
Authors: Lin Wei, Xiang Zhang, Yucheng Shi, Mengyang He and Yufei Gao

Paper 359 - Exploring Ensemble Adversarial Attacks against Deep Reinforcement Learning Agents
Authors: Ryan Horache, Mehdi Hennequin, Juba Agoun, Mohand-Saïd Hacid and Abdelkrim Zitouni

Paper 913 - Inter-Class Directional Divergence: A Geometry-Based Regularization for Adversarial Robustness
Authors: Junfeng Liu, Bosheng Chen, Zheng Yang, Xinghao Ding, Yue Huang and Huimin Guo

Paper 1062 - M-SAGE: A Multi-Scale Attention Guided Targeted Attack for Vision-Based DRL Driving Policies
Authors: Yingyuan Wu, Runwen Chen, Shuo Feng, Zhao Jin, Yizhe Luo, Xiaorong Hu, Yucheng Shi, Mingliang Xu and Yin Ma