Final-year Ph.D. student [on job market]
Department of Computer Science
Princeton University

Email: dingliy [at] cs [dot] princeton [dot] edu

[Google Scholar]

About

I am a Ph.D. student in the Department of Computer Science at Princeton University. I am very glad to be supervised by Sanjeev Arora. My research interests lie in the interface of Theoretical Computer Science and Machine Learning. My work focuses on the optimization and evaluation of large models through the lens of theoretical understanding. 

Before that, I was a Yao Class student studying Computer Science at Institute for Interdisciplinary Information Science, Tsinghua University. 

Publications 

News

• Depth-µP and Skill-Mix are accepted by ICLR 2024!
• We release Skill-Mix, a new type of evaluation of LLMs on their capability to combine basic skills. (See the demo here.) Skill-Mix resists data contaminations and can detect “cramming for leaderboard“! We also find GPT-4 is beyond “stochastic parrot” behavior based on its good performance on Skill-Mix!
• Tensor Program VI is here! Our new parametrization, Depth-µP, can scale up networks to infinite depth! You also get hyperparameter transfer for free, meaning using Depth-µP, optimal hyperparameters in a shallow network also work for a deep network.