Lightweight Distributed Gaussian Process Regression for Online Machine Learning

TL;DR

This paper introduces a lightweight distributed Gaussian process regression algorithm enabling multiple agents with limited resources to collaboratively learn a static function from streaming data, improving predictions through efficient communication.

Contribution

It proposes a novel distributed GPR method tailored for resource-constrained agents, combining local and global predictions for enhanced accuracy.

Findings

Limited communication can enhance learning performance.

The algorithm achieves accurate predictions with reduced resource usage.

Monte Carlo simulations validate the effectiveness of the approach.

Abstract

In this paper, we study the problem where a group of agents aim to collaboratively learn a common static latent function through streaming data. We propose a lightweight distributed Gaussian process regression (GPR) algorithm that is cognizant of agents' limited capabilities in communication, computation and memory. Each agent independently runs agent-based GPR using local streaming data to predict test points of interest; then the agents collaboratively execute distributed GPR to obtain global predictions over a common sparse set of test points; finally, each agent fuses results from distributed GPR with agent-based GPR to refine its predictions. By quantifying the transient and steady-state performances in predictive variance and error, we show that limited inter-agent communication improves learning performances in the sense of Pareto. Monte Carlo simulation is conducted to evaluate the developed algorithm.