Random Multi-Type Spanning Forests for Synchronization on Sparse Graphs

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Jaquard, Hugo, Amblard, Pierre-Olivier, Barthelmé, Simon, Tremblay, Nicolas
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913543825129472
author Jaquard, Hugo
Amblard, Pierre-Olivier
Barthelmé, Simon
Tremblay, Nicolas
author_facet Jaquard, Hugo
Amblard, Pierre-Olivier
Barthelmé, Simon
Tremblay, Nicolas
contents Random diffusions are a popular tool in Monte-Carlo estimations, with well established algorithms such as Walk-on-Spheres (WoS) going back several decades. In this work, we introduce diffusion estimators for the problems of angular synchronization and smoothing on graphs, in the presence of a rotation associated to each edge. Unlike classical WoS algorithms that are point-wise estimators, our diffusion estimators allow for global estimations by propagating along the branches of random spanning subgraphs called multi-type spanning forests. Building upon efficient samplers based on variants of Wilson's algorithm, we show that our estimators outperform standard numerical-linear-algebra solvers in challenging instances, depending on the topology and density of the graph.
format Preprint
id arxiv_https___arxiv_org_abs_2403_19300
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Random Multi-Type Spanning Forests for Synchronization on Sparse Graphs
Jaquard, Hugo
Amblard, Pierre-Olivier
Barthelmé, Simon
Tremblay, Nicolas
Probability
Data Structures and Algorithms
Statistics Theory
Random diffusions are a popular tool in Monte-Carlo estimations, with well established algorithms such as Walk-on-Spheres (WoS) going back several decades. In this work, we introduce diffusion estimators for the problems of angular synchronization and smoothing on graphs, in the presence of a rotation associated to each edge. Unlike classical WoS algorithms that are point-wise estimators, our diffusion estimators allow for global estimations by propagating along the branches of random spanning subgraphs called multi-type spanning forests. Building upon efficient samplers based on variants of Wilson's algorithm, we show that our estimators outperform standard numerical-linear-algebra solvers in challenging instances, depending on the topology and density of the graph.
title Random Multi-Type Spanning Forests for Synchronization on Sparse Graphs
topic Probability
Data Structures and Algorithms
Statistics Theory
url https://arxiv.org/abs/2403.19300