Federated Bayesian Deep Learning: The Application of Statistical Aggregation Methods to Bayesian Models

Federated learning (FL) is an approach to training machine learning models that takes advantage of multiple distributed datasets while maintaining data privacy and reducing communication costs associated with sharing local datasets. Aggregation strategies have been developed to pool or fuse the weig...

Full description

Saved in:
Bibliographic Details
Main Authors: John Fischer, Marko Orescanin, Justin Loomis, Patrick Mcclure
Format: Article
Language:English
Published: IEEE 2024-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10781394/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850106680791007232
author John Fischer
Marko Orescanin
Justin Loomis
Patrick Mcclure
author_facet John Fischer
Marko Orescanin
Justin Loomis
Patrick Mcclure
author_sort John Fischer
collection DOAJ
description Federated learning (FL) is an approach to training machine learning models that takes advantage of multiple distributed datasets while maintaining data privacy and reducing communication costs associated with sharing local datasets. Aggregation strategies have been developed to pool or fuse the weights and biases of distributed deterministic models; however, modern deterministic deep learning (DL) models are often poorly calibrated and lack the ability to communicate a measure of epistemic uncertainty in prediction, which is desirable for remote sensing platforms and safety-critical applications. Conversely, Bayesian DL models are often well calibrated and capable of quantifying and communicating a measure of epistemic uncertainty along with a competitive prediction accuracy. Unfortunately, because the weights and biases in Bayesian DL models are defined by a probability distribution, simple application of the aggregation methods associated with FL schemes for deterministic models is either impossible or results in sub-optimal performance. In this work, we use independent and identically distributed (IID) and non-IID partitions of the CIFAR-10 dataset and a fully variational ResNet-20 architecture to analyze six different aggregation strategies for Bayesian DL models. Additionally, we analyze the traditional federated averaging approach applied to an approximate Bayesian Monte Carlo dropout model as a lightweight alternative to more complex variational inference methods in FL. We show that aggregation strategy is a key hyperparameter in the design of a Bayesian FL system with downstream effects on accuracy, calibration, uncertainty quantification, training stability, and client compute requirements.
format Article
id doaj-art-f302002ef75a45b5a20be919c57c52f5
institution OA Journals
issn 2169-3536
language English
publishDate 2024-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-f302002ef75a45b5a20be919c57c52f52025-08-20T02:38:46ZengIEEEIEEE Access2169-35362024-01-011218579018580610.1109/ACCESS.2024.351325310781394Federated Bayesian Deep Learning: The Application of Statistical Aggregation Methods to Bayesian ModelsJohn Fischer0https://orcid.org/0000-0002-2565-6706Marko Orescanin1https://orcid.org/0000-0003-3305-8412Justin Loomis2https://orcid.org/0009-0000-8435-7020Patrick Mcclure3https://orcid.org/0000-0002-2439-971XDepartment of Computer Science, Naval Postgraduate School, Monterey, CA, USADepartment of Computer Science, Naval Postgraduate School, Monterey, CA, USADepartment of Computer Science, Naval Postgraduate School, Monterey, CA, USADepartment of Computer Science, Naval Postgraduate School, Monterey, CA, USAFederated learning (FL) is an approach to training machine learning models that takes advantage of multiple distributed datasets while maintaining data privacy and reducing communication costs associated with sharing local datasets. Aggregation strategies have been developed to pool or fuse the weights and biases of distributed deterministic models; however, modern deterministic deep learning (DL) models are often poorly calibrated and lack the ability to communicate a measure of epistemic uncertainty in prediction, which is desirable for remote sensing platforms and safety-critical applications. Conversely, Bayesian DL models are often well calibrated and capable of quantifying and communicating a measure of epistemic uncertainty along with a competitive prediction accuracy. Unfortunately, because the weights and biases in Bayesian DL models are defined by a probability distribution, simple application of the aggregation methods associated with FL schemes for deterministic models is either impossible or results in sub-optimal performance. In this work, we use independent and identically distributed (IID) and non-IID partitions of the CIFAR-10 dataset and a fully variational ResNet-20 architecture to analyze six different aggregation strategies for Bayesian DL models. Additionally, we analyze the traditional federated averaging approach applied to an approximate Bayesian Monte Carlo dropout model as a lightweight alternative to more complex variational inference methods in FL. We show that aggregation strategy is a key hyperparameter in the design of a Bayesian FL system with downstream effects on accuracy, calibration, uncertainty quantification, training stability, and client compute requirements.https://ieeexplore.ieee.org/document/10781394/Bayesian deep learningfederated learningMonte Carlo dropoutuncertainty decompositionuncertainty quantificationvariational inference
spellingShingle John Fischer
Marko Orescanin
Justin Loomis
Patrick Mcclure
Federated Bayesian Deep Learning: The Application of Statistical Aggregation Methods to Bayesian Models
IEEE Access
Bayesian deep learning
federated learning
Monte Carlo dropout
uncertainty decomposition
uncertainty quantification
variational inference
title Federated Bayesian Deep Learning: The Application of Statistical Aggregation Methods to Bayesian Models
title_full Federated Bayesian Deep Learning: The Application of Statistical Aggregation Methods to Bayesian Models
title_fullStr Federated Bayesian Deep Learning: The Application of Statistical Aggregation Methods to Bayesian Models
title_full_unstemmed Federated Bayesian Deep Learning: The Application of Statistical Aggregation Methods to Bayesian Models
title_short Federated Bayesian Deep Learning: The Application of Statistical Aggregation Methods to Bayesian Models
title_sort federated bayesian deep learning the application of statistical aggregation methods to bayesian models
topic Bayesian deep learning
federated learning
Monte Carlo dropout
uncertainty decomposition
uncertainty quantification
variational inference
url https://ieeexplore.ieee.org/document/10781394/
work_keys_str_mv AT johnfischer federatedbayesiandeeplearningtheapplicationofstatisticalaggregationmethodstobayesianmodels
AT markoorescanin federatedbayesiandeeplearningtheapplicationofstatisticalaggregationmethodstobayesianmodels
AT justinloomis federatedbayesiandeeplearningtheapplicationofstatisticalaggregationmethodstobayesianmodels
AT patrickmcclure federatedbayesiandeeplearningtheapplicationofstatisticalaggregationmethodstobayesianmodels