no code implementations • 25 Mar 2024 • Jiaojiao Zhang, Linglingzhi Zhu, Mikael Johansson
We propose a novel differentially private algorithm for online federated learning that employs temporally correlated noise to improve the utility while ensuring the privacy of the continuously released models.
no code implementations • 11 Dec 2023 • Xuyang Wu, Changxin Liu, Sindri Magnusson, Mikael Johansson
In contrast to alternatives, our algorithms can converge to the fixed point set of their synchronous counterparts using step-sizes that are independent of the delays.
no code implementations • 4 Sep 2023 • Jiaojiao Zhang, Jiang Hu, Mikael Johansson
We propose a novel algorithm for solving the composite Federated Learning (FL) problem.
no code implementations • 2 Aug 2023 • Jiaojiao Zhang, Dominik Fay, Mikael Johansson
This paper proposes a locally differentially private federated learning algorithm for strongly convex but possibly nonsmooth problems that protects the gradients of each worker against an honest but curious server.
no code implementations • 17 Feb 2022 • Xuyang Wu, Sindri Magnusson, Hamid Reza Feyzmahdavian, Mikael Johansson
In this paper, we show that it is possible to use learning rates that depend on the actual time-varying delays in the system.
no code implementations • 25 Jan 2022 • Xiaoyu Wang, Mikael Johansson
In this note, we investigate uniform boundedness properties of iterates and function values along the trajectories of the stochastic gradient descent algorithm and its important momentum variant.
no code implementations • ICLR 2022 • Vien V. Mai, Jacob Lindbäck, Mikael Johansson
In addition, we establish a linear convergence rate for our formulation of the OT problem.
no code implementations • 9 Sep 2021 • Hamid Reza Feyzmahdavian, Mikael Johansson
We introduce novel convergence results for asynchronous iterations that appear in the analysis of parallel and distributed optimization algorithms.
no code implementations • 5 Jun 2021 • Xiaoyu Wang, Mikael Johansson
Many popular learning-rate schedules for deep neural networks combine a decaying trend with local perturbations that attempt to escape saddle points and bad local minima.
no code implementations • NeurIPS 2021 • Xiaoyu Wang, Sindri Magnússon, Mikael Johansson
The convergence of stochastic gradient descent is highly dependent on the step-size, especially on non-convex problems such as neural network training.
no code implementations • 12 Feb 2021 • Vien V. Mai, Mikael Johansson
We also study the convergence of a clipped method with momentum, which includes clipped SGD as a special case, for weakly convex problems under standard assumptions.
no code implementations • 24 Jun 2020 • Mahmoud Assran, Arda Aytekin, Hamid Feyzmahdavian, Mikael Johansson, Michael Rabbat
Motivated by large-scale optimization problems arising in the context of machine learning, there have been several advances in the study of asynchronous parallel and distributed optimization methods during the past decade.
no code implementations • 13 Mar 2020 • Sarit Khirirat, Sindri Magnússon, Arda Aytekin, Mikael Johansson
With the increasing scale of machine learning tasks, it has become essential to reduce the communication between computing nodes.
no code implementations • ICML 2020 • Vien V. Mai, Mikael Johansson
This paper establishes the convergence rate of a stochastic subgradient method with a momentum term of Polyak type for a broad class of non-smooth, non-convex, and constrained optimization problems.
1 code implementation • ICML 2020 • Vien V. Mai, Mikael Johansson
We therefore propose a simple scheme for stabilization that combines the global worst-case guarantees of proximal gradient methods with the local adaptation and practical speed-up of Anderson acceleration.
1 code implementation • 23 Sep 2019 • Martin Biel, Mikael Johansson
These structure-exploiting solvers are based on variations of the classical L-shaped and progressive-hedging algorithms.
Optimization and Control Mathematical Software 90C15, 90C06, 90C90
no code implementations • 23 Sep 2019 • Sarit Khirirat, Sindri Magnússon, Mikael Johansson
Several gradient compression techniques have been proposed to reduce the communication load at the price of a loss in solution accuracy.
no code implementations • 20 Mar 2019 • Vien V. Mai, Mikael Johansson
This paper introduces an efficient algorithm for finding the dominant generalized eigenvectors of a pair of symmetric matrices.
no code implementations • 24 Jan 2019 • Vien V. Mai, Mikael Johansson
This paper introduces an efficient second-order method for solving the elastic net problem.
no code implementations • 10 Jan 2019 • Arda Aytekin, Mikael Johansson
The event-driven and elastic nature of serverless runtimes makes them a very efficient and cost-effective alternative for scaling up computations.
no code implementations • 8 Oct 2018 • Arda Aytekin, Martin Biel, Mikael Johansson
We present POLO --- a C++ library for large-scale parallel optimization research that emphasizes ease-of-use, flexibility and efficiency in algorithm design.
no code implementations • NeurIPS 2018 • Dan Alistarh, Torsten Hoefler, Mikael Johansson, Sarit Khirirat, Nikola Konstantinov, Cédric Renggli
Distributed training of massive machine learning models, in particular deep neural networks, via Stochastic Gradient Descent (SGD) is becoming commonplace.
no code implementations • 18 Jun 2018 • Sarit Khirirat, Hamid Reza Feyzmahdavian, Mikael Johansson
Asynchronous computation and gradient compression have emerged as two key techniques for achieving scalability in distributed optimization for large-scale machine learning.
no code implementations • NeurIPS 2018 • Motoya Ohnishi, Masahiro Yukawa, Mikael Johansson, Masashi Sugiyama
Motivated by the success of reinforcement learning (RL) for discrete-time tasks such as AlphaGo and Atari games, there has been a recent surge of interest in using RL for continuous-time control of physical systems (cf.
no code implementations • 18 Oct 2016 • Arda Aytekin, Hamid Reza Feyzmahdavian, Mikael Johansson
This paper presents an asynchronous incremental aggregated gradient algorithm and its implementation in a parameter server framework for solving regularized optimization problems.
no code implementations • 2 Aug 2015 • Zhenhua Zou, Anders Gidmark, Themistoklis Charalambous, Mikael Johansson
While the idea of RF-EH is appealing, it is not always beneficial to attempt to harvest energy; in environments where the ambient energy is low, nodes could consume more energy being awake with their harvesting circuits turned on than what they can extract from the ambient radio signals; it is then better to enter a sleep mode until the ambient RF energy increases.
no code implementations • 18 May 2015 • Hamid Reza Feyzmahdavian, Arda Aytekin, Mikael Johansson
Mini-batch optimization has proven to be a powerful paradigm for large-scale learning.
no code implementations • 27 Sep 2013 • M. Sadegh Talebi, Zhenhua Zou, Richard Combes, Alexandre Proutiere, Mikael Johansson
The parameters, and hence the optimal path, can only be estimated by routing packets through the network and observing the realized delays.