Get our free extension to see links to code for papers anywhere online!

Chrome logo Add to Chrome

Firefox logo Add to Firefox

Picture for Mikael Johansson

Asynchronous Iterations in Optimization: New Sequence Results and Sharper Algorithmic Guarantees


Sep 09, 2021
Hamid Reza Feyzmahdavian, Mikael Johansson

* 44 pages, 1 Figure 

  Access Paper or Ask Questions

Bandwidth-based Step-Sizes for Non-Convex Stochastic Optimization


Jun 05, 2021
Xiaoyu Wang, Mikael Johansson


  Access Paper or Ask Questions

On the Convergence of Step Decay Step-Size for Stochastic Optimization


Feb 18, 2021
Xiaoyu Wang, Sindri MagnĂșsson, Mikael Johansson


  Access Paper or Ask Questions

Stability and Convergence of Stochastic Gradient Clipping: Beyond Lipschitz Continuity and Smoothness


Feb 12, 2021
Vien V. Mai, Mikael Johansson


  Access Paper or Ask Questions

Advances in Asynchronous Parallel and Distributed Optimization


Jun 24, 2020
Mahmoud Assran, Arda Aytekin, Hamid Feyzmahdavian, Mikael Johansson, Michael Rabbat

* 33 pages, 4 figures 

  Access Paper or Ask Questions

Communication Efficient Sparsification for Large Scale Machine Learning


Mar 13, 2020
Sarit Khirirat, Sindri MagnĂșsson, Arda Aytekin, Mikael Johansson

* 15 pages, 5 figures, 1 table 

  Access Paper or Ask Questions

Convergence of a Stochastic Gradient Method with Momentum for Nonsmooth Nonconvex Optimization


Feb 13, 2020
Vien V. Mai, Mikael Johansson


  Access Paper or Ask Questions

Anderson Acceleration of Proximal Gradient Methods


Oct 18, 2019
Vien V. Mai, Mikael Johansson

* A preliminary version of this work was published in the proceedings of ICASSP 2019 

  Access Paper or Ask Questions

Noisy Accelerated Power Method for Eigenproblems with Applications


Mar 20, 2019
Vien V. Mai, Mikael Johansson

* Accepted for publication in the IEEE Transaction on Signal Processing 

  Access Paper or Ask Questions

Curvature-Exploiting Acceleration of Elastic Net Computations


Jan 24, 2019
Vien V. Mai, Mikael Johansson

* 34 pages, 2 figures 

  Access Paper or Ask Questions

Harnessing the Power of Serverless Runtimes for Large-Scale Optimization


Jan 10, 2019
Arda Aytekin, Mikael Johansson

* 9 pages, double column, 9 figures 

  Access Paper or Ask Questions

Continuous-time Value Function Approximation in Reproducing Kernel Hilbert Spaces


Oct 26, 2018
Motoya Ohnishi, Masahiro Yukawa, Mikael Johansson, Masashi Sugiyama

* NIPS 2018 - Advances in Neural Information Processing Systems (with the supplementary document) 

  Access Paper or Ask Questions

POLO: a POLicy-based Optimization library


Oct 08, 2018
Arda Aytekin, Martin Biel, Mikael Johansson

* 25 pages, 7 figures 

  Access Paper or Ask Questions

The Convergence of Sparsified Gradient Methods


Sep 27, 2018
Dan Alistarh, Torsten Hoefler, Mikael Johansson, Sarit Khirirat, Nikola Konstantinov, CĂ©dric Renggli

* NIPS 2018 - Advances in Neural Information Processing Systems; Authors in alphabetic order 

  Access Paper or Ask Questions

Distributed learning with compressed gradients


Jun 18, 2018
Sarit Khirirat, Hamid Reza Feyzmahdavian, Mikael Johansson

* 29 pages, 4 figures, 2 tables 

  Access Paper or Ask Questions

Stochastic Online Shortest Path Routing: The Value of Feedback


Jan 18, 2017
M. Sadegh Talebi, Zhenhua Zou, Richard Combes, Alexandre Proutiere, Mikael Johansson

* 18 pages 

  Access Paper or Ask Questions

Analysis and Implementation of an Asynchronous Optimization Algorithm for the Parameter Server


Oct 18, 2016
Arda Aytekin, Hamid Reza Feyzmahdavian, Mikael Johansson

* 10 pages, 3 figures 

  Access Paper or Ask Questions

Optimal Radio Frequency Energy Harvesting with Limited Energy Arrival Knowledge


Aug 02, 2015
Zhenhua Zou, Anders Gidmark, Themistoklis Charalambous, Mikael Johansson


  Access Paper or Ask Questions

An Asynchronous Mini-Batch Algorithm for Regularized Stochastic Optimization


May 18, 2015
Hamid Reza Feyzmahdavian, Arda Aytekin, Mikael Johansson


  Access Paper or Ask Questions

Ergodic Mirror Descent


Aug 01, 2012
John C. Duchi, Alekh Agarwal, Mikael Johansson, Michael I. Jordan

* 35 pages, 2 figures 

  Access Paper or Ask Questions