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 distributed mini-batch algorithm, a method of converting any serial gradient-based online prediction algorithm into a distributed algorithm. 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. Our method can also be used to solve the closely-related distributed stochastic optimization problem, attaining an asymptotically linear speedup. We demonstrate the merits of our approach on a web-scale online prediction problem.
Technical Report, Microsoft Research. 1 Microsoft Way, Redmond, WA 98052.