Picture for Mert Gurbuzbalaban

Mert Gurbuzbalaban

Accelerated gradient methods for nonconvex optimization: Escape trajectories from strict saddle points and convergence to local minima

Add code
Jul 13, 2023
Viaarxiv icon

Uniform-in-Time Wasserstein Stability Bounds for (Noisy) Stochastic Gradient Descent

Add code
May 20, 2023
Viaarxiv icon

Heavy-Tail Phenomenon in Decentralized SGD

Add code
May 16, 2022
Figure 1 for Heavy-Tail Phenomenon in Decentralized SGD
Figure 2 for Heavy-Tail Phenomenon in Decentralized SGD
Figure 3 for Heavy-Tail Phenomenon in Decentralized SGD
Figure 4 for Heavy-Tail Phenomenon in Decentralized SGD
Viaarxiv icon

A Variance-Reduced Stochastic Accelerated Primal Dual Algorithm

Add code
Feb 19, 2022
Figure 1 for A Variance-Reduced Stochastic Accelerated Primal Dual Algorithm
Figure 2 for A Variance-Reduced Stochastic Accelerated Primal Dual Algorithm
Figure 3 for A Variance-Reduced Stochastic Accelerated Primal Dual Algorithm
Viaarxiv icon

Differentially Private Accelerated Optimization Algorithms

Add code
Aug 05, 2020
Figure 1 for Differentially Private Accelerated Optimization Algorithms
Figure 2 for Differentially Private Accelerated Optimization Algorithms
Figure 3 for Differentially Private Accelerated Optimization Algorithms
Figure 4 for Differentially Private Accelerated Optimization Algorithms
Viaarxiv icon

Fractional moment-preserving initialization schemes for training fully-connected neural networks

Add code
Jun 08, 2020
Figure 1 for Fractional moment-preserving initialization schemes for training fully-connected neural networks
Figure 2 for Fractional moment-preserving initialization schemes for training fully-connected neural networks
Figure 3 for Fractional moment-preserving initialization schemes for training fully-connected neural networks
Figure 4 for Fractional moment-preserving initialization schemes for training fully-connected neural networks
Viaarxiv icon

The Heavy-Tail Phenomenon in SGD

Add code
Jun 08, 2020
Figure 1 for The Heavy-Tail Phenomenon in SGD
Figure 2 for The Heavy-Tail Phenomenon in SGD
Figure 3 for The Heavy-Tail Phenomenon in SGD
Figure 4 for The Heavy-Tail Phenomenon in SGD
Viaarxiv icon

Non-Convex Stochastic Optimization via Non-Reversible Stochastic Gradient Langevin Dynamics

Add code
Apr 06, 2020
Figure 1 for Non-Convex Stochastic Optimization via Non-Reversible Stochastic Gradient Langevin Dynamics
Figure 2 for Non-Convex Stochastic Optimization via Non-Reversible Stochastic Gradient Langevin Dynamics
Figure 3 for Non-Convex Stochastic Optimization via Non-Reversible Stochastic Gradient Langevin Dynamics
Viaarxiv icon

Robust Distributed Accelerated Stochastic Gradient Methods for Multi-Agent Networks

Add code
Nov 15, 2019
Figure 1 for Robust Distributed Accelerated Stochastic Gradient Methods for Multi-Agent Networks
Figure 2 for Robust Distributed Accelerated Stochastic Gradient Methods for Multi-Agent Networks
Figure 3 for Robust Distributed Accelerated Stochastic Gradient Methods for Multi-Agent Networks
Figure 4 for Robust Distributed Accelerated Stochastic Gradient Methods for Multi-Agent Networks
Viaarxiv icon

A Universally Optimal Multistage Accelerated Stochastic Gradient Method

Add code
Jan 25, 2019
Figure 1 for A Universally Optimal Multistage Accelerated Stochastic Gradient Method
Figure 2 for A Universally Optimal Multistage Accelerated Stochastic Gradient Method
Figure 3 for A Universally Optimal Multistage Accelerated Stochastic Gradient Method
Figure 4 for A Universally Optimal Multistage Accelerated Stochastic Gradient Method
Viaarxiv icon