Koulik Khamaru

 

Department of Statistics, Rutgers University

About me

I am an Assistant Professor of Statistics at Rutgers University. Before joining Rutgers, I completed my PhD from University of California, Berkeley under the guidance of two amazing supervisors Professor Martin J. Wainwright and Professor Michael I. Jordan. Before coming to Berkeley I finished my Undergraduate and Masters in Statistics from Indian Statistical Institute, Kolkata. In 2019, I spent a wonderful summer at Amazon, working with Professor Dean Foster. In 2020, I was part of the Berkeley-BAIR collaboration and worked under Professor Lester Mackey.

Research interest

My research interests spans theory and application of statistics, machine learning and optimization. Specific areas include EM algorithm, Gaussian mixture models, model mis-specification, factor analysis, reinforcement learning, inference in sequential environments, non-convex optimization.

If you are interested in any of these topics / statistics and optimization in general, please feel free to chat with me.

Recent news

Select papers

  1. Adaptive Linear Estimating Equations.
    with Mufang Ying, and Cun-Hui Zhang, NeurIPS 2023

  2. Optimal variance-reduced stochastic approximation in Banach spaces.
    with Wenlong Mou, Martin J. Wainwright, Peter L. Bartlett and Michael I. Jordan. (2022) Submitted

  3. Is Temporal Difference Learning Optimal? An Instance-Dependent Analysis.
    with Ashwin Pananjady, Feng Ruan, Martin J. Wainwright and and Michael I. Jordan. (2021) SIAM Journal on Mathematics of Data Science

  4. Singularity, misspecification and the convergence rate of EM.
    with Raaz Dwivedi, Nhat Ho, Michael I. Jordan, Martin J. Wainwright and Bin Yu (2020) Annals of Statistics

Contact information

kk1241 (at) rutgers (dot) edu
403 Hill Center
Department of Statistics, Rutgers University