Zihao Wang

I am interested in machine learning. In particular, I want to:

enable AI to make creative contributions to science—for instance, discovering new statistical algorithms for complex experimental data, and developing a scientific understanding of AI systems themselves;
understand the societal and economic impact of AI and what it means for humanity—where the unique value of human intelligence lies, and how we should interact and co-evolve with this new form of intelligence.

Currently, I am a Ph.D. candidate at Stanford Mathematics. I am extremely fortunate to be advised by Andrea Montanari. I also work closely with Percy Liang. I am interning at the Flatiron Institute in New York during the 2026 summer, working with Joan Bruna, Alberto Bietti, and Denny Wu.

Previously, I received my B.S. in Mathematics from Peking University in 2024. I visited Princeton in 2023. I was fortunate to be advised by Jason D. Lee and Lei Wu for my undergraduate research.

Publications

* indicates equal contribution or alphabetical author order. Representative papers are highlighted.

Phase Transitions for Feature Learning in Neural Networks
Andrea Montanari*, Zihao Wang*
arXiv:2602.01434 / paper / slides / poster / Andrea's talk

Neural Networks Learn Generic Multi-Index Models near Information-Theoretic Limit
Bohan Zhang*, Zihao Wang*, Hengyu Fu, Jason D. Lee
ICLR, 2026, arXiv:2511.15120 / paper

UQ: Assessing Language Models on Unsolved Questions
Fan Nie*, Ken Ziyu Liu*, Zihao Wang, Rui Sun, Wei Liu, Weijia Shi, Huaxiu Yao, Linjun Zhang, Andrew Y. Ng, James Zou, Sanmi Koyejo, Yejin Choi, Percy Liang, Niklas Muennighoff*
arXiv:2508.17580 / paper / slides

Learning Hierarchical Polynomials of Multiple Nonlinear Features with Three-Layer Networks
Hengyu Fu, Zihao Wang, Eshaan Nichani, Jason D. Lee
ICLR, 2025, arXiv:2411.17201 / paper

The Local Landscape of Phase Retrieval Under Limited Samples
Kaizhao Liu*, Zihao Wang*, Lei Wu
IEEE Transactions on Information Theory, arXiv:2311.15221 / paper

Learning Hierarchical Polynomials with Three-Layer Neural Networks
Zihao Wang, Eshaan Nichani, Jason D. Lee
ICLR, 2024, arXiv:2311.13774 / paper

Theoretical Analysis of the Inductive Biases in Deep Convolutional Networks
Zihao Wang, Lei Wu
NeurIPS, 2023, arXiv:2305.08404 / paper

Selected Awards

Weiming Scholar, Peking University 2024

Francis Robbins Upton Fellowship (declined), Princeton University 2024

Selected Presentations

Phase transitions for feature learning

Learning theory workshop, Tsinghua University July 2026

Spring retreat, Department of Statistics, Stanford April 2026

ML theory reading group, Stanford April 2026

ML theory reading group, Simons Institute, Berkeley March 2026

MaD seminar, New York University March 2026

CCM seminar, Flatiron Institute March 2026

ML theory reading group, Wharton, UPenn March 2026

ML seminar, Peking University December 2025

ML theory reading group, Tsinghua University December 2025

Miscellaneous

I love travelling and want to explore places off the beaten path. One of my favorite books is The Art of Travel by Alain de Botton. One of my favorite movies is Nomadland by Chloé Zhao.

I am proud to be born and raised in Qingdao, a coastal city where I grew up just a stone's throw from the sea.

Updated at June 2026

Thanks Jon Barron for this amazing template