Wenlong Mou (牟文龙)

alt text 

PhD student,
Department of EECS,
UC Berkeley
Office: 264 Cory Hall, Berkeley, CA
Phone: +1 510 409 5625
E-mail: wmou [@] berkeley [DOT] edu
Google Scholar

About me

I am a fourth-year Ph.D. student at Department of EECS, UC Berkeley. I'm very fortunate to be advised by Prof. Martin Wainwright and Prof. Peter Bartlett. Prior to Berkeley, I received B.S. in Computer Science from Peking University in 2017, where I was fortunate to work with Prof. Liwei Wang. In 2016, I spent a wonderful summer at CMU, working with Prof. Nina Balcan. In summer 2019, I worked as a research intern with Dr. Zheng Wen at Adobe Research.

Research

My research interests are broadly in statistics, machine learning theory, optimization and applied probability. The goal of my research is to push the frontier of computational/statistical possibilities for big data analysis, under minimal assumptions. Currently, I'm particularly interested in the following topics:

  • Reinforcement learning theory, statistical complexity of exploration, geometry of Bellman equations, model-free RL and model misspecification.

  • Diffusion process and SDEs, high-dimensional sampling algorithm, interplay between optimization and sampling, discretization of SDEs.

  • Theory of Bayesian inference, asymptotic and non-asymptotic analysis of posterior distributions, mean-field variational approximation.

  • Stochastic approximation algorithms and associated statistical inference problems, stochastic optimization.

Papers

  1. Wenlong Mou, Ashwin Pananjady, Martin Wainwright, "Optimal Oracle Inequalities for Solving Projected Fixed-point Equations", preprint (arXiv)

  2. Wenlong Mou, Nicolas Flammarion, Martin Wainwright, Peter Bartlett, " Improved Bounds for Discretization of Langevin Diffusions: Near-optimal rates without Convexity ", under minor revision at Bernoulli Journal (arXiv)

  3. Wenlong Mou, Yi-An Ma (equal contribution), Martin Wainwright, Peter Bartlett, Michael Jordan, " High-Order Langevin Diffusion Yields an Accelerated MCMC Algorithm ", Journal of Machine Learning Research (2020+) (arXiv)

  4. Chris Junchi Li, Wenlong Mou (equal contribution), Martin Wainwright, Michael Jordan "ROOT-SGD: Sharp Nonasymptotics and Asymptotic Efficiency in a Single Algorithm", submitted. (arXiv)

  5. Wenlong Mou, Nhat Ho, Martin Wainwright, Peter Bartlett, Michael Jordan, " A Diffusion Process Perspective on Posterior Contraction Rates for Parameters", under revision (arXiv)

  6. Wenlong Mou, Zheng Wen, Xi Chen, " On the Sample Complexity of Reinforcement Learning with Policy Space Generalization ", preprint (arXiv)

  7. Wenlong Mou, Chris Junchi Li, Martin Wainwright, Peter Bartlett, Michael Jordan, "On Linear Stochastic Approximation: Fine-grained Polyak–Ruppert and Non-Asymptotic Concentration", In COLT 2020 (arXiv)

  8. Wenlong Mou, Nicolas Flammarion, Martin Wainwright, Peter Bartlett, " An Efficient Sampling Algorithm for Non-smooth Composite Potentials", under revision (arXiv)

  9. Wenlong Mou, Nhat Ho, Martin Wainwright, Peter Bartlett, Michael Jordan, " Sampling for Bayesian Mixture Models: MCMC with Polynomial-Time Mixing", under revision (arXiv)

  10. Wenlong Mou, Liwei Wang, Xiyu Zhai, Kai Zheng, "Generalization Bounds of SGLD for Non-convex Learning: Two Theoretical Viewpoints", In COLT 2018 (alphabetical order)

  11. Wenlong Mou, Yuchen Zhou, Jun Gao, Liwei Wang "Dropout Training, Data-dependent Regularization and Generalization Bounds ", In ICML 2018

  12. Maria-Florina Balcan, Travis Dick, Yingyu Liang, Wenlong Mou and Hongyang Zhang, "Differentially Private Clustering in High-dimensional Euclidean Space", In ICML 2017 (alphabetical order)

  13. Kai Zheng, Wenlong Mou (equal contribution), Liwei Wang, "Collect at Once, Use Effec-tively: Making Non-interactive Locally Private Learning Possible", In ICML 2017