About I am a PhD student in Electrical and Computer Engineering (ECE) at the University of California San Diego (UCSD), where I am advised by Prof. Yang Zheng. I obtained my MS in Communication Engineering from National Taiwan University (NTU), where I was advised by Prof. See-May Phoong. I received my BS in ECE from National Yang Ming Chiao Tung University.

Research I am broadly interested in the theory of optimization and sequential decision-making under uncertainty, with applications in stochastic, robust, and nonstochastic control. Some of my research directions include: Online learning and control. I am investigating the problem of adaptive and nonstationary online non-stochastic control using tools from online convex optimization, aiming to develop algorithms that adapt across three layers: adversarial, nonstationary, and benign environments. I also design and analyze data-driven predictive control algorithms for Koopman-linearizable nonlinear systems with provable dynamic regret guarantees. Convex and nonconvex optimization in control. I study optimization landscapes of classical optimal and robust control problems from a modern policy optimization perspective, including linear–quadratic regulator (LQR) and \( \mathcal{H}_\infty \) robust control, etc. We build a unified Extended Convex Lifting framework to reveal hidden convexity, which offers a bridge between nonconvex policy optimization and convex reformulations. Previously, I worked on reward-free exploration in reinforcement learning, specifically the problem of online active model estimation for Markov decision processes. I also worked on signal processing for communication systems in my MS study.

Experience

I worked as an Audio R&D Intern at Qualcomm during Summer 2025, focusing on design and analysis of adaptive active noise cancellation (ANC) algorithms for wearables and AR audio systems. I combined ideas from optimization, control theory, and signal processing to develop practical solutions.

---

Talks

• Problem-dependent Dynamic Regret for Online Convex Optimization
  Presented at the SOC reading groupg, Oct. 2025.
• Online Non-stochastic Control: a Regret-minimizing Approach
  Invited presentation in DSC291: Online Learning (Instructor: Prof. Yoav Freund), March 2025.
• Policy Optimization for Mixed \( \mathcal{H}_2 \)/\( \mathcal{H}_\infty \) Control: Benign Nonconvexity and Global Optimality
  Presented at the 44th Southern California Control Workshop, University of Southern California, Nov. 1, 2024.

Publications

• Semidefinite Programming Duality in Infinite-Horizon Linear Quadratic Differential Games
  Yuto Watanabe, Chih-Fan Pai, and Yang Zheng. Preprint 2025
• Extended Convex Lifting for Policy Optimization of Optimal and Robust Control
  Yang Zheng, Chih-Fan Pai, and Yujie Tang. To appear in L4DC 2025
• Low Complexity Estimation of LTV Channels for OFDM Systems with Uniformly Spaced Pilots
  Chih-Fan Pai, and See-May Phoong. 2024 32nd European Signal Processing Conference (EUSIPCO)
• Benign Nonconvex Landscapes in Optimal and Robust Control, Part II: Extended Convex Lifting
  Yang Zheng, Chih-Fan Pai, and Yujie Tang. Preprint 2024
• Benign Nonconvex Landscapes in Optimal and Robust Control, Part I: Global Optimality
  Yang Zheng, Chih-Fan Pai, and Yujie Tang. Preprint 2023
• Depth-L Nyquist(M) Filters and Biorthogonal Partners
  Chih-Fan Pai, Tien-Chiu Hung, and See-May Phoong. IEEE Access 2020

Teaching

I worked as a teaching assistant for the following courses.

• ECE 228: Machine Learning for Physical Applications, UCSD (Instructor: Prof. Yuanyuan Shi)
• ECE 171A: Linear Control System Theory, UCSD (Instructor: Prof. Yang Zheng)
• ECE 285: Semidefinite and Sum-of-squares Optimization, UCSD (Instructor: Prof. Yang Zheng)
• ECE 101: Linear Systems Fundamentals, UCSD (Instructor: Prof. Paul Siegel)
• Calculus, NTU MATH
• Linear Algebra, NTU EE
• Digital Signal Processing, NTU CommE
• Multirate Signal Processing, NTU CommE