Research

Research Interests

• high-dimensional and non-parametric statistics

• statistical inference

• reinforcement learning

• bioinformatics and statistical genomics

Preprints

• Denoising diffusion probabilistic models are optimally adaptive to unknown low dimensionality.

  • Zhihan Huang, Yuting Wei, Yuxin Chen

• Statistical inference for temporal difference learning with linear function approximation.

  • Weichen Wu, Gen Li, Yuting Wei, Alessandro Rinaldo

• Stochastic Runge-Kutta methods: Provable acceleration of diffusion models.

  • Yuchen Wu, Yuxin Chen, Yuting Wei

• Towards a mathematical theory for consistency training in diffusion models.

  • Gen Li∗, Zhihan Huang∗, Yuting Wei (∗=equal contribution)

• A sharp convergence theory for the probability flow ODEs of diffusion models.

  • Gen Li, Yuting Wei, Yuejie Chi, Yuxin Chen

• Theoretical insights for diffusion guidance: A case study for Gaussian mixture models.
  

  • Yuchen Wu, Minshuo Chen, Zihao Li, Mengdi Wang, Yuting Wei

  • Short version accepted to ICML, 2024

• A non-asymptotic distributional theory of approximate message passing for sparse and robust regression.
  

  • Gen Li, Yuting Wei

• A non-asymptotic framework for approximate message passing in spiked models.
  

  • Gen Li, Yuting Wei

• Mitigating multiple descents: A model-agnostic framework for risk monotonization.
  

  • Pratik Patil, Arun Kumar Kuchibhotla, Yuting Wei, Alessandro Rinaldo

• Minimum L1 interpolators: Precise asymptotics and multiple descent.
  

  • Yue Li, Yuting Wei

Publications: Theory and Methodology

• Hybrid reinforcement learning breaks sample size barriers in linear MDPs.

  • Kevin Tan, Wei Fan, and Yuting Wei

  • NeurIPS, 2024

• Federated natural policy gradient methods for multi-task reinforcement learning.
  

  • Tong Yang, Shicong Cen, Yuting Wei, Yuxin Chen, Yuejie Chi

  • NeurIPS, 2024

• Accelerating convergence of score-based diffusion models, provably.
  

  • Gen Li∗, Yu Huang∗, Timofey Efimov, Yuting Wei, Yuejie Chi, Yuxin Chen (∗=equal contribution)

  • ICML, 2024

• Towards faster non-asymptotic convergence for diffusion-based generative models.
  

  • Gen Li, Yuting Wei, Yuxin Chen, Yuejie Chi

  • Short version accepted to ICLR, 2024

• High-probability sample complexities for policy evaluation with linear function approximation.

  • Gen Li∗, Weichen Wu∗, Yuejie Chi, Cong Ma, Alessandro Rinaldo, Yuting Wei (∗=equal contribution)

  • IEEE Transactions on Information Theory, 2024

• The Lasso with general Gaussian designs with applications to hypothesis testing.
  

  • Michael Celentano, Andrea Montanari, Yuting Wei (alphabetical order) (slides)

  • Annals of Statistics, 2023

• Settling the sample complexity of model-based offline reinforcement learning.
  

  • Gen Li, Laixi Shi, Yuxin Chen, Yuejie Chi, Yuting Wei

  • Annals of Statistics, 2024

• Approximate message passing from random initialization with applications to Z2 synchronization.
  

  • Gen Li, Wei Fan, Yuting Wei

  • Proceedings of the National Academy of Sciences (PNAS), 2023

• The curious price of distributional robustness in reinforcement learning with a generative model.
  

  • Laixi Shi, Gen Li, Yuting Wei, Yuxin Chen, Matthieu Geist, Yuejie Chi

  • Short version accepted to NeurIPS, 2023

• Fast policy extragradient methods for competitive games with entropy regularization.
  

  • Shicong Cen, Yuting Wei, Yuejie Chi

  • Short version accepted to NeurIPS, 2021

  • Journal of Machine Learning Research, 2024

• Is Q-Learning minimax optimal? A tight sample complexity analysis.
  

  • Gen Li, Changxiao Cai, Yuxin Chen, Yuting Wei, Yuejie Chi

  • Short version accepted to ICML, 2021

  • Operations Research, 2024

• Breaking the sample size barrier in model-based reinforcement learning with a generative model.
  

  • Gen Li, Yuting Wei, Yuejie Chi, Yuxin Chen

  • Short version accepted to NeurIPS, 2020 (slides)

  • Operations Research, 2024

• Softmax policy gradient methods can take exponential time to converge.
  

  • Gen Li, Yuting Wei, Yuejie Chi, Yuxin Chen

  • Short version accepted to Conference on Learning Theory (COLT), 2021

  • Mathematical Programming, 2023

• Derandomizing Knockoffs.
  

  • Zhimei Ren, Yuting Wei, Emmanuel Candès

  • Journal of the American Statistical Association, 2023 (code)

• Minimax-optimal multi-agent RL in zero-sum Markov games with a generative model.
  

  • Gen Li, Yuejie Chi, Yuting Wei, Yuxin Chen

  • NeurIPS, oral presentation, 2022

• Pessimistic Q-Learning for offline reinforcement learning: Towards optimal sample complexity.
  

  • Laixi Shi, Gen Li, Yuting Wei, Yuxin Chen, Yuejie Chi

  • ICML, 2022

• Uniform consistency of cross validation estimators for high-dimensional ridge regression.
  

  • Pratik Patil, Yuting Wei, Alessandro Rinaldo, Ryan Tibshirani

  • AISTATS, oral presentation, 2021

• Sample-efficient reinforcement learning is feasible for linearly realizable MDPs with limited revisiting.
  

  • Gen Li, Yuxin Chen, Yuejie Chi, Yuantao Gu, Yuting Wei

  • NeurIPS, 2021 (slides)

• Integration and transfer learning of single-cell transcriptomes via cFIT.
  

  • Minshi Peng, Yue Li, Brie Wamsley, Yuting Wei, Kathryn Roeder

  • Proceedings of the National Academy of Sciences (PNAS), 2021 (code)

• Fast global convergence of natural policy gradient methods with entropy regularization.
  

  • Shicong Cen, Chen Cheng, Yuxin Chen, Yuting Wei, Yuejie Chi

  • Operations Research, 2022 (INFORMS George Nicholson award finalist)

• Sample complexity of asynchronous Q-learning: sharper analysis and variance reduction.
  

  • Gen Li, Yuting Wei, Yuejie Chi, Yuantao Gu, Yuxin Chen

  • Short version accepted to NeurIPS, 2020 (slides)

  • IEEE Transactions on Information Theory, 2021

• Tackling small eigen-gaps: Fine-grained eigenvector estimation and inference under heteroscedastic noise.
  

  • Chen Cheng, Yuting Wei, Yuxin Chen

  • IEEE Transactions on Information Theory, 2021

• Debiasing evaluations that are biased by evaluations.
  

  • Jingyan Wang, Ivan Stelmakh, Yuting Wei, Nihar B. Shah

  • Short version accepted to AAAI Conference on Artificial Intelligence, 2021

• Randomized tests for high-dimensional regression: A more efficient and powerful solution.
  

  • Yue Li, Ilmun Kim, Yuting Wei

  • NeurIPS, 2020

• Sharp statistical guarantees for adversarially robust Gaussian classification.
  

  • Chen Dan, Yuting Wei, Pradeep Ravikumar

  • International Conference on Machine Learning (ICML), 2020

• From Gauss to Kolmogorov: Localized measures of complexity for ellipses.
  

  • Yuting Wei, Billy Fang and Martin Wainwright

  • Electronic Journal of Statistics, 2020

• The local geometry of testing in ellipses: Tight control via localized Kolmogorov widths.
  

  • Yuting Wei and Martin Wainwright

  • IEEE Transactions on Information Theory, 2020

• Early stopping for kernel boosting algorithms: A general analysis with localized complexities.
  

  • Yuting Wei∗, Fanny Yang∗ and Martin Wainwright (∗=equal contribution)

  • Short version accepted to Neural Information Processing Systems (NeurIPS), spotlight presentation, 2017

  • IEEE Transactions on Information Theory, 2019

• The geometry of hypothesis testing over convex cones: Generalized likelihood tests and minimax radii.
  

  • Yuting Wei, Martin Wainwright and Adityanand Guntuboyina

  • The Annals of Statistics, 2019

• Adaptive estimation of planar convex sets.
  

  • Tony Cai, Adityanand Guntuboyina and Yuting Wei (alphabetical order).

  • The Annals of Statistics, 2018

• Sharp minimax bounds for testing discrete monotone distributions.
  

  • Yuting Wei and Martin Wainwright

  • International Symposium on Information Theory (ISIT), 2016

Publications: Applications

• Cell type hierarchy reconstruction via reconciliation of multi-resolution cluster tree.
  

  • Minshi Peng, Brie Wamsley, Andrew Elkins, Daniel M Geschwind, Yuting Wei, Kathryn Roeder

  • Nucleic Acids Research, 2021 (code)

• MHC binding prediction with KernelRLSpan and its variations.

  • Wen-Jun Shen∗, Yuting Wei∗, Xin Guo∗, Stephen Smale, Hau-San Wong and Shuai Cheng Li (∗=equal contribution)

  • Journal of Immunological Methods, 2014

Thesis

• Thesis: A geometric perspective on some topics in statistical learning

• Advisor: Martin Wainwright, Adityanand Guntuboyina