Saved in:
Bibliographic Details
Main Authors: Hasan, Mohsin, Zhang, Guojun, Guo, Kaiyang, Chen, Xi, Poupart, Pascal
Format: Preprint
Published: 2023
Subjects:
Online Access:https://arxiv.org/abs/2312.09817
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911753349103616
author Hasan, Mohsin
Zhang, Guojun
Guo, Kaiyang
Chen, Xi
Poupart, Pascal
author_facet Hasan, Mohsin
Zhang, Guojun
Guo, Kaiyang
Chen, Xi
Poupart, Pascal
contents Federated Learning (FL) involves training a model over a dataset distributed among clients, with the constraint that each client's dataset is localized and possibly heterogeneous. In FL, small and noisy datasets are common, highlighting the need for well-calibrated models that represent the uncertainty of predictions. The closest FL techniques to achieving such goals are the Bayesian FL methods which collect parameter samples from local posteriors, and aggregate them to approximate the global posterior. To improve scalability for larger models, one common Bayesian approach is to approximate the global predictive posterior by multiplying local predictive posteriors. In this work, we demonstrate that this method gives systematically overconfident predictions, and we remedy this by proposing $β$-Predictive Bayes, a Bayesian FL algorithm that interpolates between a mixture and product of the predictive posteriors, using a tunable parameter $β$. This parameter is tuned to improve the global ensemble's calibration, before it is distilled to a single model. Our method is evaluated on a variety of regression and classification datasets to demonstrate its superiority in calibration to other baselines, even as data heterogeneity increases. Code available at https://github.com/hasanmohsin/betaPredBayesFL
format Preprint
id arxiv_https___arxiv_org_abs_2312_09817
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Calibrated One Round Federated Learning with Bayesian Inference in the Predictive Space
Hasan, Mohsin
Zhang, Guojun
Guo, Kaiyang
Chen, Xi
Poupart, Pascal
Machine Learning
Federated Learning (FL) involves training a model over a dataset distributed among clients, with the constraint that each client's dataset is localized and possibly heterogeneous. In FL, small and noisy datasets are common, highlighting the need for well-calibrated models that represent the uncertainty of predictions. The closest FL techniques to achieving such goals are the Bayesian FL methods which collect parameter samples from local posteriors, and aggregate them to approximate the global posterior. To improve scalability for larger models, one common Bayesian approach is to approximate the global predictive posterior by multiplying local predictive posteriors. In this work, we demonstrate that this method gives systematically overconfident predictions, and we remedy this by proposing $β$-Predictive Bayes, a Bayesian FL algorithm that interpolates between a mixture and product of the predictive posteriors, using a tunable parameter $β$. This parameter is tuned to improve the global ensemble's calibration, before it is distilled to a single model. Our method is evaluated on a variety of regression and classification datasets to demonstrate its superiority in calibration to other baselines, even as data heterogeneity increases. Code available at https://github.com/hasanmohsin/betaPredBayesFL
title Calibrated One Round Federated Learning with Bayesian Inference in the Predictive Space
topic Machine Learning
url https://arxiv.org/abs/2312.09817