A Differentially Private Probabilistic Framework for Modeling the Variability Across Federated Datasets of Heterogeneous Multi-View Observations

TL;DR

This paper introduces a hierarchical Bayesian federated learning framework with differential privacy guarantees for modeling data variability across heterogeneous multi-view datasets, demonstrated on medical imaging data for Alzheimer's disease.

Contribution

It presents a novel probabilistic federated learning approach combining hierarchical Bayesian modeling with differential privacy, optimized via EM, for multi-view heterogeneous data.

Findings

Robust performance on iid and non-iid data distributions

Effective quantification of data and view variability

High-quality data reconstruction comparable to state-of-the-art methods

Abstract

We propose a novel federated learning paradigm to model data variability among heterogeneous clients in multi-centric studies. Our method is expressed through a hierarchical Bayesian latent variable model, where client-specific parameters are assumed to be realization from a global distribution at the master level, which is in turn estimated to account for data bias and variability across clients. We show that our framework can be effectively optimized through expectation maximization (EM) over latent master's distribution and clients' parameters. We also introduce formal differential privacy (DP) guarantees compatibly with our EM optimization scheme. We tested our method on the analysis of multi-modal medical imaging data and clinical scores from distributed clinical datasets of patients affected by Alzheimer's disease. We demonstrate that our method is robust when data is distributed either in iid and non-iid manners, even when local parameters perturbation is included to provide DP guarantees. Moreover, the variability of data, views and centers can be quantified in an interpretable manner, while guaranteeing high-quality data reconstruction as compared to state-of-the-art autoencoding models and federated learning schemes. The code is available at https://gitlab.inria.fr/epione/federated-multi-views-ppca.