Log In Sign Up

Katyusha: The First Direct Acceleration of Stochastic Gradient Methods

by   Zeyuan Allen-Zhu, et al.

Nesterov's momentum trick is famously known for accelerating gradient descent, and has been proven useful in building fast iterative algorithms. However, in the stochastic setting, counterexamples exist and prevent Nesterov's momentum from providing similar acceleration, even if the underlying problem is convex. We introduce Katyusha, a direct, primal-only stochastic gradient method to fix this issue. It has a provably accelerated convergence rate in convex (off-line) stochastic optimization. The main ingredient is Katyusha momentum, a novel "negative momentum" on top of Nesterov's momentum. It can be incorporated into a variance-reduction based algorithm and speed it up, both in terms of sequential and parallel performance. Since variance reduction has been successfully applied to a growing list of practical problems, our paper suggests that in each of such cases, one could potentially try to give Katyusha a hug.


Direct Acceleration of SAGA using Sampled Negative Momentum

Variance reduction is a simple and effective technique that accelerates ...

How Does Momentum Help Frank Wolfe?

We unveil the connections between Frank Wolfe (FW) type algorithms and t...

An Accelerated Stochastic Gradient for Canonical Polyadic Decomposition

We consider the problem of structured canonical polyadic decomposition. ...

Momentum Tracking: Momentum Acceleration for Decentralized Deep Learning on Heterogeneous Data

SGD with momentum acceleration is one of the key components for improvin...

Accelerated Stochastic Power Iteration

Principal component analysis (PCA) is one of the most powerful tools in ...

Accelerating Convergence of Replica Exchange Stochastic Gradient MCMC via Variance Reduction

Replica exchange stochastic gradient Langevin dynamics (reSGLD) has show...