Bayesian Nonparametric Federated Learning of Neural Networks

2019/05/28 by Mikhail Yurochkin, Yurochkin, Mikhail, Mayank Agarwal +9 · 29 citations
Computer Science · Mathematics · #Bayesian Methods and Mixture Models #FOS: Computer and information sciences #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Privacy-Preserving Technologies in Data #Statistical Methods and Inference

paper · pdf · doi:10.48550/arxiv.1905.12022

openalex publication_date 2019/05/28 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

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.

Cited by

• FedDPC : Handling Data Heterogeneity and Partial Client Participation in Federated Learning
• Geometry-Aligned Differential Privacy for Location-Safe Federated Radio Map Construction
• Merging without Forgetting: Continual Fusion of Task-Specific Models via Optimal Transport
• Bayesian Coreset Optimization for Personalized Federated Learning
• Federated Learning Based on Dynamic Regularization
• Federated Learning on Non-IID Data Silos: An Experimental Study
• SecureBoost: A Lossless Federated Learning Framework
• FedTopo: Relation-Level Topology Sharing for Model-Heterogeneous Federated Learning
• Learning Reconfigurable Representations for Multimodal Federated Learning with Missing Data
• FLAMMABLE: A Multi-Model Federated Learning Framework with Multi-Model Engagement and Adaptive Batch Sizes
• Private Federated Multiclass Post-hoc Calibration
• N-Parties Private Structure and Parameter Learning for Sum-Product Networks
• Local Learning Matters: Rethinking Data Heterogeneity in Federated Learning
• Beyond Aggregation: Guiding Clients in Heterogeneous Federated Learning
• FedCF: Fair Federated Conformal Prediction
• Ringleader ASGD: The First Asynchronous SGD with Optimal Time Complexity under Data Heterogeneity
• Adaptive Dual-Mode Distillation with Incentive Schemes for Scalable, Heterogeneous Federated Learning on Non-IID Data
• Task-Agnostic Federated Continual Learning via Replay-Free Gradient Projection
• Probabilistic Federated Learning of Neural Networks Incorporated with Global Posterior Information
• Federated CycleGAN for Privacy-Preserving Image-to-Image Translation
• FedDAF: Federated Domain Adaptation Using Model Functional Distance
• FedBiF: Communication-Efficient Federated Learning via Bits Freezing
• Cost-Free Personalization via Information-Geometric Projection in Bayesian Federated Learning
• Federated Learning Meets Multi-objective Optimization
• Variational Federated Multi-Task Learning
• Towards Instance-wise Personalized Federated Learning via Semi-Implicit Bayesian Prompt Tuning
• DOPA: Stealthy and Generalizable Backdoor Attacks from a Single Client under Challenging Federated Constraints
• An Expectation-Maximization Perspective on Federated Learning
• A Field Guide to Federated Optimization
• HeteRo-Select: Informativeness as the Participation Driver in Heterogeneous Federated Learning

Related

• Probabilistic Federated Learning of Neural Networks Incorporated with Global Posterior Information
• How to Combine Variational Bayesian Networks in Federated Learning
• Bayesian Federated Neural Matching that Completes Full Information
• Federated Learning on Stochastic Neural Networks
• Architecture Agnostic Federated Learning for Neural Networks