Rahul Kidambi

I am a graduate student of Prof. Sham M. Kakade studying Machine Learning at the University of Washington, Seattle.

E-mail: rkidambi AT uw DOT edu


Research:

My research revolves around the design and analysis of practical Algorithms for Machine Learning, as viewed through the lens of Optimization and Statistics.

I hold an active interest in the empirical aspects of deep learning and non-convex optimization.

Previously, I worked on problems at the intersection of Structured Prediction, Semi-Supervised Learning and Active Learning.


Recent Papers/Preprints:

Asterisk* indicates alphabetical ordering of authors.

  • On the insufficiency of existing Momentum schemes for Stochastic Optimization,
    Rahul Kidambi, Praneeth Netrapalli, Prateek Jain and Sham M. Kakade.
    Appeared in International Conference on Learning Representations (ICLR), 2018. (Selected for Oral Presentation; 2% Acceptance Rate)
    ArXiv manuscript, abs/1803.05591, March 2018.
    [Open Review] [Code].

  • Leverage Score Sampling for Faster Accelerated Regression and ERM, [*]
    Naman Agarwal, Sham M. Kakade, Rahul Kidambi, Yin Tat Lee, Praneeth Netrapalli and Aaron Sidford.
    ArXiv manuscript, abs/1711.08426, November 2017.

  • A Markov Chain Theory Approach to Characterizing the Minimax Optimality of Stochastic Gradient Descent (for Least Squares), [*]
    Prateek Jain, Sham M. Kakade, Rahul Kidambi, Praneeth Netrapalli, Venkata Krishna Pillutla and Aaron Sidford.
    Invited paper at FSTTCS 2017.
    ArXiv manuscript, abs/1710.09430, October 2017.

  • Accelerating Stochastic Gradient Descent, [*]
    Prateek Jain, Sham M. Kakade, Rahul Kidambi, Praneeth Netrapalli and Aaron Sidford.
    To Appear, Conference on Learning Theory (COLT), 2018.
    ArXiv manuscript, abs/1704.08227, April 2017.
    [Sham's Talk at MSR].

  • Parallelizing Stochastic Gradient Descent for Least Squares Regression: mini-batching, averaging, and model misspecification1, [*]
    Prateek Jain, Sham M. Kakade, Rahul Kidambi, Praneeth Netrapalli and Aaron Sidford.
    Accepted for publication in Journal of Machine Learning Research (JMLR), March 2018.
    ArXiv manuscript, abs/1610.03774, October 2016. Latest version, April 2018.

    The dblp listing provides a complete set of my papers.


    1. Previously titled "Parallelizing Stochastic Approximation Through Mini-Batching and Tail Averaging."

    Academic Service:

  • Conference Reviewing: ISMB 2012, NIPS 2016, COLT 2017, 2018.
  • Journal Reviewing: Journal of Machine Learning Research (JMLR) - 2015, Electronic Journal of Statistics (EJS) - 2017.


    Teaching:

    Spring 2018: Teaching Assistant for CSE 547/STAT 548: Machine Learning for Big Data.

    I have been a Teaching Assistant for the following classes:

  • EE 514a: Information Theory-I (Autumn 2015).
  • EE 215: Fundamentals of Electrical Engineering (Autumn 2014, Winter 2015).


    Contact Information:

    Rahul Kidambi,
    Department of Electrical Engineering,
    185 Stevens Way, AE100R Campus Box 352500,
    University of Washington,
    Seattle, WA 98195-2500, USA.