Welcome to my website.

I’m a fourth-year Ph.D. student in UMich EECS department since Fall 2021. I’m advised by Prof. Necmiye Ozay and studying some statistical fundamental limits of learning for dynamics and control and how to break these fundamental limits by changing the dynamical system structure or the control paradigm. Generally, we think about the statistical limits of a learning-based algorithm from the following three perspectives in order:

• Statistical consistency: Does the algorithm converge to the ground-truth solution with respect to sample size?[ACC 2025][Journal Submission 2025]
• Statistical optimality: If the algorithm is statistically consistent, does it achieve the minimax sample complexity lower bound?[L-CSS and CDC 2025]
• Statistical hardness: If the algorithm is statistically optimal, does the optimal sample complexity increase moderately with the system complexity?[CDC 2023]

You can check my PhD proposal slides to have a detailed overview. I’m also passionate about machine coffee and learning-based control system design.

Publications

Preprints

Noise Sensitivity of the Semidefinite Programs for Direct Data-Driven LQR[slides][poster][a thread on X]
Xiong Zeng, Laurent Bako, and Necmiye Ozay. Submitted for journal publication, January 2025.

Journal Papers

System Identification Under Bounded Noise: Optimal Rates Beyond Least Squares[codes]
Xiong Zeng, Jing Yu, and Necmiye Ozay, IEEE Control Systems Letters (L-CSS) and IEEE Conference on Decision and Control (CDC) 2025 (to appear).

Conference Papers

Noise Sensitivity of Direct Data-Driven Linear Quadratic Regulator by Semidefinite Programming
Xiong Zeng, Laurent Bako, and Necmiye Ozay, American Control Conference (ACC) 2025 (to appear).

On the Hardness of Learning to Stabilize Linear Systems[slides][poster][video]
Xiong Zeng, Zexiang Liu, Zhe Du, Necmiye Ozay, and Mario Sznaier. IEEE Conference on Decision and Control (CDC) 2023.

Academic Service

Reviewer for IEEE Transactions on Signal Processing (TSP), Nonlinear Analysis: Hybrid Systems (NAHS), Control Systems Letters (L-CSS), Conference on Decision and Control (CDC), and American Control Conference (ACC).

Bio

I obtained my bachelor’s degree from Automation Engineering Department of Huazhong University of Science and Technology in 2020, where I was advised by Prof. Ye Yuan. After graduation, I took a gap year with my undergraduate advisor and Prof. Jun Liu from University of Waterloo.

I was born and grew up in Loudi, Hunan (湖南娄底), China.

Misc

I like trail running, and people can often find me on the trails and hills in Ann Arbor, day or night. My personal best record for the full marathon is 03:28:08 in the Indianapolis Monumental Marathon 2022. I have completed 2 Rim-to-Rim-to-Rim (R2R2R) runs in the great Grand Canyon with my friends.

“士不可以不弘毅，任重而道远。非淡泊无以明志，非宁静无以致远。从群众中来，到群众中去。”

English version of the above famous Chinese quotes (by ChatGPT): “A scholar must cultivate broad-mindedness and resilience, as the responsibilities are heavy and the journey is long. Without nonattachment, one cannot clarify one’s aspirations; without tranquility, one cannot reach far. From people, for people.”