Global Posterior

Global posterior estimation in federated learning aims to collaboratively infer a shared model from decentralized data held by multiple clients, while preserving data privacy. Current research focuses on improving the efficiency and accuracy of this process, employing Bayesian methods like expectation propagation and approximate message passing, often within a variational inference framework, to address challenges posed by data heterogeneity and communication limitations. These advancements are crucial for scaling machine learning to large, distributed datasets and enhancing the robustness and reliability of predictions in diverse applications.

Papers

February 12, 2024

Bayesian Deep Learning Via Expectation Maximization and Turbo Deep Approximate Message Passing
Wei Xu, An Liu, Yiting Zhang, Vincent Lau
Model Compression Bayesian Deep Learning Approximate Message Bayesian Deep Global Posterior Sparse Bayesian

December 15, 2023

Calibrated One Round Federated Learning with Bayesian Inference in the Predictive Space
Mohsin Hasan, Guojun Zhang, Kaiyang Guo, Xi Chen, Pascal Poupart
Bayesian Inference Posterior Predictive Global Posterior

October 17, 2023

Federated Learning with Nonvacuous Generalisation Bounds
Pierre Jobic, Maxime Haddouche, Benjamin Guedj
Generalization Bound PAC Bayesian Private Prediction Non Vacuous Generalization Bound Global Posterior

June 21, 2023

Hierarchical Neural Simulation-Based Inference Over Event Ensembles
Lukas Heinrich, Siddharth Mishra-Sharma, Chris Pollard, Philipp Windischhofer
Neural Network Simulation Based Inference Neural Estimator Hierarchical Forecasting Global Posterior

February 8, 2023

October 31, 2022

Federated Averaging Langevin Dynamics: Toward a unified theory and new algorithms
Vincent Plassier, Alain Durmus, Eric Moulines
Bayesian Inference Federated Averaging Unified Model New Algorithm McMc Algorithm Global Posterior

June 20, 2022

Robust One Round Federated Learning with Predictive Space Bayesian Inference
Mohsin Hasan, Zehao Zhang, Kaiyang Guo, Mahdi Karami, Guojun Zhang, Xi Chen, Pascal Poupart
Robust Version Posterior Predictive Bayesian Posterior Model Space Ensemble Weather Global Posterior

December 17, 2021

Robust Distributed Bayesian Learning with Stragglers via Consensus Monte Carlo
Hari Hara Suthan Chittoor, Osvaldo Simeone
Robust Version Bayesian Perspective Bayesian Learning Straggler Mitigation Probabilistic Consensus Resilient Algorithm Global Posterior