Changlong Wu

Assistant Professor of Electrical and Computer Engineering

Links

Changlong Wu is an assistant professor in the Department of Electrical and Computer Engineering at the University of Arizona. He received his PhD in electrical engineering from the University of Hawaii at Manoa. Prior to joining the University of Arizona, he served as a visiting assistant professor in the Department of Computer Science at Purdue University and as a postdoctoral research associate at the NSF Center for Science of Information (CSoI).

Dr. Wu's research lies at the intersection of information theory, machine learning, and the foundations of artificial intelligence. He is particularly interested in understanding when learning is possible and in characterizing the fundamental limits of learnability across a wide range of settings—including online learning, sequential decision-making, and robust learning under constraints such as privacy, fault tolerance, and model hallucination. His work has appeared in leading conferences and journals, including COLT, ICML, NeurIPS, ICLR, and IEEE Transactions on Information Theory.

Degrees

PhD Electrical Engineering, University of Hawaii at Manoa, 2021
BE Computer Science, Wuhan University, 2015

Teaching Interests

Machine learning and artificial intelligence, information theory, theory of computation, and engineering mathematics

Research Interests

Algorithmic and statistical learning theory, online learning, information theory, and trustworthy machine learning

Selected Publications

C. Wu, A. Grama, and W. Szpankowski. No Free Lunch: Fundamental Limits of Learning Non-Hallucinating Generative Models, ICLR 2025.
C. Wu, Y. Wang, and A. Grama. A Theory of Fault-Tolerant Learning, ICML 2024 (Spotlight).
C. Wu, J. Sima, and W. Szpankowski. Oracle-Efficient Hybrid Online Learning with Unknown Distribution, COLT 2024.
C. Wu, A. Grama, and W. Szpankowski. Online Learning in Dynamically Changing Environments, COLT 2023.
C. Wu, M. Heidari, A. Grama, and W. Szpankowski. Precise Regret Bounds for Log-loss via a Truncated Bayesian Algorithm, NeurIPS 2022 (Oral presentation).