Federated ADMM from Bayesian Duality

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Möllenhoff, Thomas, Swaroop, Siddharth, Doshi-Velez, Finale, Khan, Mohammad Emtiyaz
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914366729748480
author Möllenhoff, Thomas
Swaroop, Siddharth
Doshi-Velez, Finale
Khan, Mohammad Emtiyaz
author_facet Möllenhoff, Thomas
Swaroop, Siddharth
Doshi-Velez, Finale
Khan, Mohammad Emtiyaz
contents We propose a new Bayesian approach to generalize the federated Alternating Direction Method of Multipliers (ADMM). We show that the solutions of variational-Bayesian (VB) objectives are associated with a duality structure that not only resembles the structure of ADMM's fixed-points but also generalizes it. For example, ADMM-like updates are recovered when the VB objective is optimized over the isotropic-Gaussian family, and new non-trivial extensions are obtained for other exponential-family distributions. These extensions include a Newton-like variant that converges in one step on quadratic objectives and an Adam-like variant that yields up to 7% accuracy boosts for deep heterogeneous cases. Our work opens a new Bayesian way to generalize ADMM and other primal-dual methods.
format Preprint
id arxiv_https___arxiv_org_abs_2506_13150
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Federated ADMM from Bayesian Duality
Möllenhoff, Thomas
Swaroop, Siddharth
Doshi-Velez, Finale
Khan, Mohammad Emtiyaz
Machine Learning
Optimization and Control
We propose a new Bayesian approach to generalize the federated Alternating Direction Method of Multipliers (ADMM). We show that the solutions of variational-Bayesian (VB) objectives are associated with a duality structure that not only resembles the structure of ADMM's fixed-points but also generalizes it. For example, ADMM-like updates are recovered when the VB objective is optimized over the isotropic-Gaussian family, and new non-trivial extensions are obtained for other exponential-family distributions. These extensions include a Newton-like variant that converges in one step on quadratic objectives and an Adam-like variant that yields up to 7% accuracy boosts for deep heterogeneous cases. Our work opens a new Bayesian way to generalize ADMM and other primal-dual methods.
title Federated ADMM from Bayesian Duality
topic Machine Learning
Optimization and Control
url https://arxiv.org/abs/2506.13150