Optimal Distributed Online Prediction using Mini-Batches

TL;DR

This paper introduces a distributed mini-batch algorithm for online prediction that converts serial gradient methods into distributed ones, achieving optimal regret bounds and linear speed-up, suitable for web-scale problems.

Contribution

The paper presents a novel distributed mini-batch algorithm with proven regret bounds, explicitly accounting for communication latency, and demonstrates its effectiveness on large-scale prediction tasks.

Findings

Achieves asymptotically optimal regret bounds for smooth convex loss functions.

Provides a method for distributed stochastic optimization with linear speed-up.

Demonstrates effectiveness on web-scale online prediction problems.

Abstract

Online prediction methods are typically presented as serial algorithms running on a single processor. However, in the age of web-scale prediction problems, it is increasingly common to encounter situations where a single processor cannot keep up with the high rate at which inputs arrive. In this work, we present the \emph{distributed mini-batch} algorithm, a method of converting many serial gradient-based online prediction algorithms into distributed algorithms. We prove a regret bound for this method that is asymptotically optimal for smooth convex loss functions and stochastic inputs. Moreover, our analysis explicitly takes into account communication latencies between nodes in the distributed environment. We show how our method can be used to solve the closely-related distributed stochastic optimization problem, achieving an asymptotically linear speed-up over multiple processors. Finally, we demonstrate the merits of our approach on a web-scale online prediction problem.