Jason D. Lee

Jason D. Lee is an associate professor of Electrical Engineering and Computer Sciences (EECS) and Statistics at the University of California, Berkeley. Previously, he held academic positions at Princeton University as an associate professor, and was a research scientist at Google DeepMind. He completed his Ph.D. in Computational and Mathematical Engineering at Stanford University under the advisement of Trevor Hastie and Jonathan Taylor. For students and collaborators, his primary contact email is jasonlee@princeton.edu, though prospective students and postdocs are asked to include "filter_student" in the subject line.

• Ph.D., Computational and Mathematical Engineering, Stanford University (2015)
• B.Sc., Mathematics, Duke University (2010)

Lee's research lies at the intersection of machine learning, statistics, and optimization, focusing on the theoretical foundations of artificial intelligence. His work addresses fundamental questions in deep learning, including optimization landscapes, representation learning, and reinforcement learning theory. He has made significant contributions to understanding how gradient descent operates in neural network training and has developed provably efficient algorithms for various learning scenarios.

His ten most recent publications represent a diverse yet coherent body of work across machine learning theory, focusing on topics such as Gaussian multi-index models, transformer learning capabilities, shallow neural networks, and optimization techniques. These publications appear in top venues including COLT, ICML, NeurIPS, and JMLR.

Among his notable accolades are:

• Samsung AI Researcher of the Year Award (2023)
• NSF Career Award (2022)
• ONR Young Investigator Award (2021)
• Sloan Research Fellow in Computer Science (2019)
• NIPS Best Student Paper Award (2016)
• Princeton Commendation for Outstanding Teaching (ECE538B)

Lee has advised numerous students including Alex Damian, Wenhao Zhan, Eshaan Nichani, Tianle Cai, Zixuan Wang, and Yunwei Ren. Former advisees have gone on to positions at institutions like NYU Courant, Facebook AI Research, UW, MIT, Duke, and Microsoft Research.

His research group and collaborators span multiple institutions, working on theoretical and applied aspects of machine learning and artificial intelligence, with a particular focus on the optimization and learning dynamics of neural networks and transformer models.