Dmitry Kovalev

PhD Student

KING ABDULLAH UNIVERSITY OF SCIENCE AND TECHNOLOGY

About

I’m a student at the King Abdullah University of Science and Technology, where I’m currently studying Computer Science and doing research in the Visual Computing Center under the supervision of Peter Richtárik. Prior to this I got my BS degree in Applied Mathematics and Physics at the Moscow Institute of Physics and Technology.

Interests

Optimization
Machine Learning

Education

PhD in Computer Science
King Abdullah University of Science and Technology
MS in Computer Science, 2019
King Abdullah University of Science and Technology
BS in Applied Mathematics and Physics, 2018
Moscow Institute of Physics and Technology

Recent Posts

ICCOPT 2019

I gave a talk "Revisiting Stochastic Extragradient Method" which was a part of "Variational Inequalities, Minimax Problems and GANs" session organised by Konstantin Mishchenko.

05 Aug 2019

Slides

Data Science Summer School 2019

I'm presenting poster at Data Science Summer School at École polytechnique, Paris, France.

26 Jun 2019

Poster

June 2019 plans

I'm staying in Moscow until June 16. I will attend "Control, Information, Optimization" summer school in Voronovo, Moscow region from June 17 to June 22. Then I will attend Data Science Summer School in Paris, France from June 24 to June 28.

11 Jun 2019

Talk: Stochastic Spectral and Conjugate Descent Methods

Today I gave a talk at the "Automatic control and Optimization Theory" seminar at the Institute for Control Problems, Moscow. The talk is based on my NeurIPS 2018 paper.

26 Mar 2019

slides

Moscow visit

I’m staying in Moscow from 18 March to 2 April. I'm going to give 2 talks and have a rest from work.

18 Mar 2019

Dmitry Kovalev

About

Interests

Education

Recent Papers

An Optimal Algorithm for Strongly Convex Minimization under Affine Constraints

ADOM: Accelerated Decentralized Optimization Method for Time-Varying Networks

IntSGD: Floatless Compression of Stochastic Gradients

Decentralized Distributed Optimization for Saddle Point Problems

A Linearly Convergent Algorithm for Decentralized Optimization: Sending Less Bits for Free!

Linearly Converging Error Compensated SGD

Towards Accelerated Rates for Distributed Optimization over Time-varying Networks

Optimal and Practical Algorithms for Smooth and Strongly Convex Decentralized Optimization

From Local SGD to Local Fixed Point Methods for Federated Learning

Acceleration for Compressed Gradient Descent in Distributed and Federated Optimization

Fast Linear Convergence of Randomized BFGS

Variance Reduced Coordinate Descent with Acceleration: New Method With a Surprising Application to Finite-Sum Problems

Distributed Fixed Point Methods with Compressed Iterates

Accelerated methods for composite non-bilinear saddle point problem

Stochastic Newton and Cubic Newton Methods with Simple Local Linear-Quadratic Rates

Stochastic Proximal Langevin Algorithm: Potential Splitting and Nonasymptotic Rates

Revisiting Stochastic Extragradient

RSN: Randomized Subspace Newton

Stochastic Distributed Learning with Gradient Quantization and Variance Reduction

Don't Jump Through Hoops and Remove Those Loops: SVRG and Katyusha are Better Without the Outer Loop

A hypothesis about the rate of global convergence for optimal methods (Newton's type) in smooth convex optimization

Stochastic Spectral and Conjugate Descent Methods