Bio
I am an Assistant Professor at the Machine Learning Department in Carnegie Mellon University. Prior to that, I was a Norbert Wiener Research Fellow jointly in the Applied Math department and IDSS at MIT. I received my PhD in the Computer Science Department at Princeton University under the advisement of Sanjeev Arora.
My research interests lie in the intersection of machine learning, statistics, and theoretical computer science, spanning topics like (probabilistic) generative models, algorithmic tools for learning and inference, representation and self-supervised learning, out-of-distribution generalization and applications of neural approaches to natural language processing and scientific domains. More broadly, the goal of my research is principled and mathematical understanding of statistical and algorithmic problems arising in modern machine learning paradigms.
I am the recipient of an NSF CAREER Award, an Amazon Research Award and a Google Research Award. I am also in part supported by several NSF and DoD awards, as well as an OpenAI Superalignment grant.
News & Events
Invited talk at IMSI Workshop on Statistical and Computational Challenges in Probabilistic Scientific Machine Learning (SciML)
Stephen Huan awarded DoE CSGF fellowship
Stephen has been awarded a Department of Energy (DOE) Computational Science Graduate Fellowship for the 2025–26 academic year. He is one of only 30 students nationwide selected for the fellowship, which supports researchers applying high-performance computing to problems in science and engineering.
Link for more info: https://www.cs.cmu.edu/news/2025/huan-doe-fellowship
COLT workshop on Theory of AI for Scientific Computing (TaSC)
Myself, former lab member Tanya Marwah, Misha Khodak, Nick Boffi and Jianfeng Lu are co-organizing a COLT workshop on Theory of AI for Scientific Computing. We are soliciting contributed work at the interface of learning theory, statistics, numerical methods, algorithm design, and the physical sciences.
Tanya Marwah successfully depended her thesis, and will join Polymathic AI and Flatiron Institute as a Research Fellow. Congrats Tanya!
Academic Positions
Assistant Professor
Carnegie Mellon University: Machine Learning Department
2019 - present
Norbert Wiener Fellow
MIT: IDSS and Applied Mathematics
2017 - 2019
Education
Ph.D. in Computer Science
Princeton University
Advised by Sanjeev Arora
2012 - 2017
B.Sc. in Computer Science
Princeton University
2008 - 2012
Service
Senior Area Chair
NeurIPS 2024
Area Chair
ICLR 2023, 2024
NeurIPS 2020, 2021, 2022, 2023
COLT 2022
UAI 2022
Program Committee
FOCS 2020, 2022
Awards
Google Research Award
2024Algorithmic Foundations for Generative AI: Inference, Distillation and Non-Autoregressive Generation
NSF CAREER Award
2023Theoretical Foundations of Modern Machine Learning Paradigms: Generative and Out-of-Distribution
OpenAI Superalignment Award
2023Research support for AI alignment work
Amazon Research Award
2022Causal + Deep Out-of-Distribution Learning
Research Areas
Generative Models
Algorithmic and statistical foundations of generative models, from GANs to diffusion models.
Representation learning
Methodological foundations for learning, understanding and interpreting representations.
Out-of-Distribution Generalization
Robustness and adaptation to benign and adversarial distribution shifts.
Neural Language Models
Theoretical and practical aspects of deep learning approaches to natural language processing.
AI for Scientific Applications
Machine learning approaches for the sciences.
Sampling and Optimization
Algorithms and theory for computationally and statistically efficient sampling and optimization.
Publications
Curated selection of key publications across research areas:
Kaiyue Wen, Yuchen Li and Bingbin Liu
NeurIPS 2023
With Frederic Koehler and Alexander Heckett
ICLR 2023, Oral Equivalent, Top 5% of Papers
With Tanya Marwah and Zachary C. Lipton
NeurIPS 2021, Spotlight
With Rong Ge and Holden Lee
NeurIPS 2018
With Sanjeev Arora and Yi Zhang
ICLR 2018
With Sanjeev Arora, Yuanzhi Li, Yingyu Liang and Tengyu Ma
Transactions of the Association for Computational Linguistics (TACL), 2016
Talks
Talks from the last 2 years (2023-2025):
Theoretical perspectives on modern machine learning paradigms: generative, scientific and out-of-distribution
2025UC Berkeley CS seminar, 03/2025
The statistical cost of score-based losses
2024Simons Institute for the Theory of Computing, Moderns Paradigms in Generalization Boot Camp, 09/2024
Neural networks for PDEs: representational power and inductive biases
2024Workshop on Mathematics of Data, Institute for Mathematical Sciences at NUS, 01/2024
From algorithms to neural networks and back
2024ETH Seminar, 07/2024
The statistical cost of score-based losses
2024Oxford University, Computational Statistics and Machine Learning Seminar, 07/2024
Discernible patterns in trained Transformers: a view from simple linguistic sandboxes
2024Theory of Interpretable AI seminar, 11/2024
Neural networks for PDEs: representational power and inductive biases
2024CNLS Annual Conference on Physics-Informed Machine Learning, 10/2024
The statistical cost of score-based losses
2024Duke University Computer Science Department Colloquium, 11/2024
The statistical cost of score matching
2024Joint Mathematics Meetings, 01/2024
Neural networks for PDEs: representational power and inductive biases
2024Meeting in Mathematical Statistics, CIRM, 12/2024
The statistical cost of score-based losses
2024Newton Institute program 'Diffusion Models in Machine Learning: Foundations, generative models and non-convex optimisation', 07/2024
The statistical cost of score-based losses
2024SIAM Mathematics of Data Science (MDS), 10/2024
The statistical cost of score-based losses
2024University of College London (UCL), 07/2024
The statistical cost of score-based losses
2024Webinar for the Section on Statistical Learning and Data Science (SLDS) of the American Statistical Association (ASA), 05/2024
From algorithms to neural networks and back
2023Mathematics of Modern Machine Learning (M3L) Workshop at NeurIPS 2023, 12/2023 (talk starts at 22:15)
The statistical cost of score matching
2023IDSS Stochastics and Statistics Seminar, MIT, 02/2023
Fit like you sample: sample-efficient score matching from fast mixing diffusions
2023Allerton Conference, 09/2023
The statistical cost of score matching
2023Microsoft Research Redmond, 08/2023
The statistical cost of score matching
2023Yale FDS Workshop 'Theory and Practice of Foundation Models', 10/2023
The statistical cost of score matching
2023Meeting in Mathematical Statistics, CIRM, 12/2023
Students
Current Students
PhD in Machine Learning
PhD in Computer Science
PhD in Machine Learning
PhD in Machine Learning
Alumni
PhD in Machine Learning (2025)
Research Fellow at the Flatiron Institute and Polymathic AI
Visiting undergraduate student from Tsinghua University (Spring 2023)
PhD student at Stanford University
Masters in Machine Learning (MSML)
Masters in Machine Learning (MSML)
Masters in Machine Learning (MSML)
Apple
Masters in Machine Learning (MSML)
PhD student at University of Washington
