Federated Generalized Bayesian Learning via Distributed Stein Variational Gradient Descent

TL;DR

This paper presents DSVGD, a federated Bayesian inference method using particles to efficiently approximate the global posterior, balancing communication and accuracy.

Contribution

Introduces Distributed Stein Variational Gradient Descent (DSVGD), a novel non-parametric Bayesian federated learning framework with flexible communication and improved scalability.

Findings

Outperforms benchmark federated learning methods in accuracy.

Provides well-calibrated, trustworthy predictions.

Offers a flexible trade-off between communication load and convergence.

Abstract

This paper introduces Distributed Stein Variational Gradient Descent (DSVGD), a non-parametric generalized Bayesian inference framework for federated learning. DSVGD maintains a number of non-random and interacting particles at a central server to represent the current iterate of the model global posterior. The particles are iteratively downloaded and updated by one of the agents with the end goal of minimizing the global free energy. By varying the number of particles, DSVGD enables a flexible trade-off between per-iteration communication load and number of communication rounds. DSVGD is shown to compare favorably to benchmark frequentist and Bayesian federated learning strategies, also scheduling a single device per iteration, in terms of accuracy and scalability with respect to the number of agents, while also providing well-calibrated, and hence trustworthy, predictions.