Communication-Aware Collaborative Learning

TL;DR

This paper develops communication-efficient collaborative PAC learning algorithms that maintain low sample complexity, even in noisy environments, by leveraging distributed boosting techniques and noise-robust adaptations.

Contribution

It introduces novel algorithms that significantly reduce communication costs in collaborative PAC learning while handling classification noise effectively.

Findings

Algorithms achieve low communication overhead without increasing sample complexity.

Proposed methods are robust to classification noise in collaborative settings.

Distributed boosting is effectively used to optimize communication efficiency.

Abstract

Algorithms for noiseless collaborative PAC learning have been analyzed and optimized in recent years with respect to sample complexity. In this paper, we study collaborative PAC learning with the goal of reducing communication cost at essentially no penalty to the sample complexity. We develop communication efficient collaborative PAC learning algorithms using distributed boosting. We then consider the communication cost of collaborative learning in the presence of classification noise. As an intermediate step, we show how collaborative PAC learning algorithms can be adapted to handle classification noise. With this insight, we develop communication efficient algorithms for collaborative PAC learning robust to classification noise.