Chaowei Xiao

Papers

Total Citations

H-Index

About

Chaowei Xiao is a leading researcher in the security and robustness of deep learning systems, with a particular focus on adversarial machine learning. His work has fundamentally advanced our understanding of how neural networks can be attacked and defended, especially in safety-critical domains. Xiao’s seminal paper, “Characterizing Attacks on Deep Reinforcement Learning” (2019, 52 citations), was among the first to systematically demonstrate that Deep Reinforcement Learning (DRL) models are vulnerable to adversarial perturbations in their observations, exposing critical weaknesses in autonomous decision-making systems. More recently, he has pioneered defenses for 3D point cloud recognition, a key technology for autonomous driving and robotics. His 2022 work, “PointDP: Diffusion-driven Purification against Adversarial Attacks on 3D Point Cloud Recognition” (10 citations), introduces an innovative diffusion-based purification method that effectively removes adversarial noise from 3D data. By bridging the gap between theoretical attack strategies and practical, deployable defenses, Xiao’s research has become essential reading for anyone working on trustworthy AI, and his contributions continue to shape the field’s approach to building resilient machine learning models for real-world applications.