A Markov Chain Theory Approach to Characterizing the Minimax Optimality
of Stochastic Gradient Descent (for Least Squares)

v1v2 (latest)

A Markov Chain Theory Approach to Characterizing the Minimax Optimality of Stochastic Gradient Descent (for Least Squares)

25 October 2017

Praneeth Netrapalli

Krishna Pillutla

ArXiv (abs)PDF HTML

Papers citing "A Markov Chain Theory Approach to Characterizing the Minimax Optimality of Stochastic Gradient Descent (for Least Squares)"

5 / 5 papers shown

Title
The Implicit Regularization of Stochastic Gradient Flow for Least Squares Alnur Ali Yan Sun Robert Tibshirani 68 77 0 17 Mar 2020
Parallelizing Stochastic Gradient Descent for Least Squares Regression: mini-batching, averaging, and model misspecification Prateek Jain Sham Kakade Rahul Kidambi Praneeth Netrapalli Aaron Sidford MoMe 71 36 0 12 Oct 2016
Competing with the Empirical Risk Minimizer in a Single Pass Roy Frostig Rong Ge Sham Kakade Aaron Sidford 69 100 0 20 Dec 2014
Non-parametric Stochastic Approximation with Large Step sizes Aymeric Dieuleveut Francis R. Bach 56 170 0 02 Aug 2014
Adaptivity of averaged stochastic gradient descent to local strong convexity for logistic regression Francis R. Bach 93 165 0 25 Mar 2013