Yu-Xiang Wang

Yu-Xiang Wang is Associate Professor at UC San Diego's Halıcıoğlu Data Science Institute and Department of Computer Science and Engineering. He directs the S2ML Lab, focusing on machine learning theory and applications in privacy, reinforcement learning, and generative AI.

His research develops provable algorithms for watermarking AI content, practical differential privacy, offline RL, and adaptive deep learning. Current NSF projects investigate adaptivity in neural networks and privacy-preserving ML.

Publications address watermark robustness, differential privacy optimization, neural collapse phenomena, and sparse model training. Recent advances include privacy-preserving generative AI detection and adaptive algorithms for nonstationary environments.

Honored with NSF CAREER award and paper awards (NeurIPS Spotlight, ICML Oral), he teaches graduate courses in machine learning and differential privacy. Research applications span healthcare, finance, and AI safety.