Jerry Liu

PhD student at Stanford ICME · Hazy Research · DOE CSGF Fellow · jerrywliu@stanford.edu

prof_pic.png

About

I’m a 4th year PhD student in the Institute of Computational & Mathematical Engineering at Stanford, advised by Chris Ré. I am supported by the DOE Computational Science Graduate Fellowship. Previously, I completed my undergraduate degrees in Math and Computer Science at Duke, where I was advised by Cynthia Rudin. My path has been shaped by many kind and brilliant researchers, including Atri Rudra, Michael Mahoney, Jin Yao, and Kenny Weiss.

Research Interests

My recent work studies memory in language models: how knowledge is encoded in model weights and how architectural choices determine what models can remember and at what cost. As LLM workloads become more recall-intensive, I’m interested in the tradeoffs between learning in weights and learning in context, and in using these insights to build more parameter- and compute-efficient systems.

I’m also interested in machine learning for numerical tasks, especially in scientific settings such as differential equations. My previous work examined why standard architectures struggle with high-precision numerical computation, and developed methods to improve precision in PDEs and continuous-valued regression.

News

Apr 07, 2026 Featured in a DEIXIS profile on my work at the intersection of machine learning, numerics, and scientific reasoning.
March 2026 Constructing Machine-Precision Neural Networks with Quasi-Interpolants was selected for an oral presentation (8/113) at the AI&PDE Workshop at ICLR 2026.
Jun 30, 2025 Presented BWLer at the Theory of AI for Scientific Computing (TASC) Workshop at COLT, where it received the Best Paper Award.

Selected Publications

  1. Constructing Efficient Fact-Storing MLPs for Transformers
    Owen Dugan*, Roberto Garcia*, Ronny Junkins*, Jerry Liu*, Dylan Zinsley, Sabri Eyuboglu, Atri Rudra, and Chris Ré
    In review, 2025
    TL;DR: We show how MLPs can store facts as key-value mappings within Transformers, giving an explicit, information-theoretically optimal construction that improves on prior constructions by up to two orders of magnitude. Our analysis precisely characterizes how embedding geometry and query noise limit fact-storage capacity.
  2. Constructing Machine-Precision Neural Networks with Quasi-Interpolants
    Catherine Deng*, Junmiao Hu*, Milan Rohatgi, Jerry Weihong Liu, and Christopher Ré
    In AI&PDE: ICLR 2026 Workshop on AI and Partial Differential Equations, 2026
    Oral (8/113 accepted papers)
    TL;DR: We give a closed-form construction for MLP interpolants based on quasi-interpolation theory, reaching machine precision for the first time and showing that optimization, rather than expressivity, is the main bottleneck for high-precision scientific ML.
  3. BWLer: Barycentric Weight Layer Elucidates a Precision-Conditioning Tradeoff for PINNs
    Jerry Weihong Liu, Yasa Baig, Denise Hui Jean Lee, Rajat Vadiraj Dwaraknath, Atri Rudra, and Christopher Re
    In Workshop on the Theory of AI for Scientific Computing @ COLT, 2025
    Best Paper Award
    TL;DR: BWLer uses barycentric interpolation within physics-informed neural networks to decouple solution parameterization from derivative computation, improving accuracy by up to three orders of magnitude on benchmarks and revealing a precision-conditioning tradeoff.