Yuheng Bu

Yuheng Bu is an Assistant Professor in the Department of Computer Science at the University of California, Santa Barbara (UCSB). Previously, he held positions at the University of Florida and MIT. He earned his Ph.D. from the University of Illinois at Urbana-Champaign (2019) and B.S. from Tsinghua University (2014). His research focuses on foundational machine learning, integrating tools from information theory and signal processing to address challenges in fair learning, uncertainty quantification, and watermarking generative AI. Notable contributions include work on generalization error analysis, adaptive watermarking techniques, and fairness-aware algorithms.

• Affiliations: UCSB (current), University of Florida (2020-2023), MIT (Postdoc, 2019-2020)
• Education: Ph.D. Electrical & Computer Engineering, UIUC (2019); B.S. Electronic Engineering, Tsinghua University (2014)

Research Interests: Machine Learning Theory, AI Ethics, Information-Theoretic Bounds, Signal Processing, and Applications in Generative AI Security. His work emphasizes rigorous theoretical foundations for trustworthy algorithms with guarantees on generalization, fairness, and robustness.

Key Achievements: Developed adaptive watermarking schemes for LLMs, derived exact generalization error characterizations for Gibbs algorithms, and advanced understanding of fairness overfitting. His work has been recognized with awards including the Abraham Wald Prize (2023) and the MLSP Best Student Paper (2022).

• Grants & Collaborations: Collaborations with institutions like MIT, Turing Institute, and INRIA. Research supported by NSF and industry partnerships.
• Labs/Teams: Leads the Information-Theoretic Learning Group at UCSB, fostering interdisciplinary research on AI foundations.