Colored Markov Random Fields for Probabilistic Topological Modeling

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Marinucci, Lorenzo, Di Nino, Leonardo, D'Acunto, Gabriele, Pandolfo, Mario Edoardo, Di Lorenzo, Paolo, Barbarossa, Sergio
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866917122023620608
author Marinucci, Lorenzo
Di Nino, Leonardo
D'Acunto, Gabriele
Pandolfo, Mario Edoardo
Di Lorenzo, Paolo
Barbarossa, Sergio
author_facet Marinucci, Lorenzo
Di Nino, Leonardo
D'Acunto, Gabriele
Pandolfo, Mario Edoardo
Di Lorenzo, Paolo
Barbarossa, Sergio
contents Probabilistic Graphical Models (PGMs) encode conditional dependencies among random variables using a graph -nodes for variables, links for dependencies- and factorize the joint distribution into lower-dimensional components. This makes PGMs well-suited for analyzing complex systems and supporting decision-making. Recent advances in topological signal processing highlight the importance of variables defined on topological spaces in several application domains. In such cases, the underlying topology shapes statistical relationships, limiting the expressiveness of canonical PGMs. To overcome this limitation, we introduce Colored Markov Random Fields (CMRFs), which model both conditional and marginal dependencies among Gaussian edge variables on topological spaces, with a theoretical foundation in Hodge theory. CMRFs extend classical Gaussian Markov Random Fields by including link coloring: connectivity encodes conditional independence, while color encodes marginal independence. We quantify the benefits of CMRFs through a distributed estimation case study over a physical network, comparing it with baselines with different levels of topological prior.
format Preprint
id arxiv_https___arxiv_org_abs_2512_03727
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Colored Markov Random Fields for Probabilistic Topological Modeling
Marinucci, Lorenzo
Di Nino, Leonardo
D'Acunto, Gabriele
Pandolfo, Mario Edoardo
Di Lorenzo, Paolo
Barbarossa, Sergio
Machine Learning
Signal Processing
Methodology
Probabilistic Graphical Models (PGMs) encode conditional dependencies among random variables using a graph -nodes for variables, links for dependencies- and factorize the joint distribution into lower-dimensional components. This makes PGMs well-suited for analyzing complex systems and supporting decision-making. Recent advances in topological signal processing highlight the importance of variables defined on topological spaces in several application domains. In such cases, the underlying topology shapes statistical relationships, limiting the expressiveness of canonical PGMs. To overcome this limitation, we introduce Colored Markov Random Fields (CMRFs), which model both conditional and marginal dependencies among Gaussian edge variables on topological spaces, with a theoretical foundation in Hodge theory. CMRFs extend classical Gaussian Markov Random Fields by including link coloring: connectivity encodes conditional independence, while color encodes marginal independence. We quantify the benefits of CMRFs through a distributed estimation case study over a physical network, comparing it with baselines with different levels of topological prior.
title Colored Markov Random Fields for Probabilistic Topological Modeling
topic Machine Learning
Signal Processing
Methodology
url https://arxiv.org/abs/2512.03727