Lei Wu (吴磊)
 |
Office: 静园6院 205 |
About Me
I am an Assistant Professor in the School of Mathematical Sciences and Center for Machine Learning Research at Peking University.
Previously, I was a postdoctoral researcher at PACM, Princeton University and at the Wharton School, University of Pennsylvania. I received my Ph.D. in Computational Mathematics from Peking University in 2018, advised by Prof. Weinan E, and my B.S. in Mathematics from Nankai University in 2012.
I aim to develop a predictive science of learning that bridges tractable models and real-world training, from kernel methods to neural networks and modern LLM pretraining. See Research for my research vision and selected highlights.
Recruiting
We are actively seeking self-motivated postdocs, PhD, and undergraduate students interested in the scientific foundations of modern machine learning, particularly approximation, optimization, scaling laws, and LLM pretraining.
If you are interested in joining our group, please email me your CV, transcript (for student applicants), and a brief description of your background and research interests.
Recent News
2026-02: Constant-depth network with smooth activations released on arXiv.
Establishes that smooth activations (e.g., GELU, SiLU) enable smoothness adaptivity in constant-depth neural networks, achieving optimal approximation and statistical rates.
2026-02: Optimal learning-rate schedules under FSL released on arXiv.
Proves that the optimal learning-rate schedule depends on task difficulty; in the hard-task regime, the optimal schedule exhibits a warmup–stable–decay (WSD) structure.
2026-02: Fast catch-up, late switching accepted to ICLR 2026.
Studies batch-size scheduling under FSL, revealing a fast catch-up effect, which holds across linear regression and LLM pretraining.
2025-09: Functional Scaling Laws accepted to NeurIPS 2025 (Spotlight).
Introduces a functional scaling law (FSL) framework that—in contrast to classical scaling laws, which only describe final-step behavior—characterizes the entire loss trajectory, spanning from linear regression to LLM pretraining.