Skip to yearly menu bar Skip to main content


Bayesian Nonparametric Federated Learning of Neural Networks

Mikhail Yurochkin · Mayank Agarwal · Soumya Ghosh · Kristjan Greenewald · Nghia Hoang · Yasaman Khazaeni

Pacific Ballroom #20

Keywords: [ Parallel and Distributed Learning ] [ Bayesian Nonparametrics ] [ Bayesian Deep Learning ]


In federated learning problems, data is scattered across different servers and exchanging or pooling it is often impractical or prohibited. We develop a Bayesian nonparametric framework for federated learning with neural networks. Each data server is assumed to provide local neural network weights, which are modeled through our framework. We then develop an inference approach that allows us to synthesize a more expressive global network without additional supervision, data pooling and with as few as a single communication round. We then demonstrate the efficacy of our approach on federated learning problems simulated from two popular image classification datasets.

Live content is unavailable. Log in and register to view live content