About
Linglong Meng
Hi! My name is Linglong Meng, and my Chinese name is “孟令茏”. In Chinese, Linglong conveys intelligence and a quick, agile mind.
I am in the Honors Program with a major in Mathematics at the University of Michigan. I am very fortunate to be advised by Prof. Laura Balzano in the SPADA Lab and Prof. Gabriel Poesia in the Poetry Lab, both from the Department of Electrical Engineering and Computer Science.
News
News will appear here.
Research
My research interests grew out of my early work with classical sparse coding methods in machine learning. I hope to use modern mathematical perspectives to understand the internal mechanisms of learning systems and translate those insights into the design of future model architectures through a deeper understanding of training dynamics, optimization strategies, and representation structure. My goal is to develop models that are more efficient, capable, and interpretable.
Recently, my research has focused on three main directions:
-
Low-Dimensional Structures
I study different forms of low-rank structure that emerge during model training and adaptation, with reference to this IEEE Signal Processing Magazine article.
-
Generative Model Theory
I investigate mathematical and mechanistic explanations for diffusion models and large language models.
-
Modern Optimization
I am especially interested in Riemannian optimization methods such as geodesic modeling and in optimization dynamics for deep learning. My goal is to understand and explain the nature of overparameterization in deep learning.