Chan Li

About me

I am a third-year Ph.D. candidate in the Department of Physics at the University of California, San Diego (UCSD), advised by Prof. Nigel Goldenfeld. I am expected to graduate in June 2027.
My research lies at the intersection of statistical physics and machine learning, aiming to apply physics principles to understand the black box of deep learning and design physics-inspired learning algorithms.

I am currently working on:
(1) applying statistical and physics-inspired modeling to analyze and optimize learning dynamics in neural networks,
(2) studying stability and non-equilibrium behavior to improve the robustness and efficiency of large-scale learning systems,
(3) uncovering universal scaling laws and phase-transition–like behaviors that govern model performance, and
(4) leveraging principles from statistical physics to develop interpretable, reliable, and scalable machine learning algorithms.

Before joining UC San Diego, I conducted research at Sun Yat-sen University (Guangzhou, China) in the PMI Lab under the supervision of Prof. Haiping Huang.
My earlier research focused on developing statistical-physics-based theories of learning, covering hierarchical credit assignment, emergent representations, continual learning, and predictive coding. These works collectively proposed a unified mean-field and ensemble framework for understanding how biological and artificial neural systems assign, transfer, and preserve information across multiple levels of representation.

I received my M.S. in Theoretical Physics (2020–2023) and B.S. in Optoelectronic Information Science and Engineering (2016–2020) from Sun Yat-sen University, Guangzhou, China.

My work has been published in Physical Review Letters (the top place in physics to announce rapidly novel exciting results), Physical Review E, and Physical Review Research.

I am currently exploring Summer 2026 research internship opportunities in machine learning and data science, aiming to apply my background in theoretical physics and statistical modeling to develop physics-inspired approaches for analyzing and improving modern machine learning systems.

Link to PDF version of my CV

Contact

Feel free to reach out to me at chan.li.physics@gmail.com (personal email)
or chanli@ucsd.edu (professional email).