Shannon invariants: A scalable approach to information decomposition

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Gutknecht, Aaron J., Rosas, Fernando E., Ehrlich, David A., Makkeh, Abdullah, Mediano, Pedro A. M., Wibral, Michael
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916702140235776
author Gutknecht, Aaron J.
Rosas, Fernando E.
Ehrlich, David A.
Makkeh, Abdullah
Mediano, Pedro A. M.
Wibral, Michael
author_facet Gutknecht, Aaron J.
Rosas, Fernando E.
Ehrlich, David A.
Makkeh, Abdullah
Mediano, Pedro A. M.
Wibral, Michael
contents Distributed systems, such as biological and artificial neural networks, process information via complex interactions engaging multiple subsystems, resulting in high-order patterns with distinct properties across scales. Investigating how these systems process information remains challenging due to difficulties in defining appropriate multivariate metrics and ensuring their scalability to large systems. To address these challenges, we introduce a novel framework based on what we call "Shannon invariants" -- quantities that capture essential properties of high-order information processing in a way that depends only on the definition of entropy and can be efficiently calculated for large systems. Our theoretical results demonstrate how Shannon invariants can be used to resolve long-standing ambiguities regarding the interpretation of widely used multivariate information-theoretic measures. Moreover, our practical results reveal distinctive information-processing signatures of various deep learning architectures across layers, which lead to new insights into how these systems process information and how this evolves during training. Overall, our framework resolves fundamental limitations in analyzing high-order phenomena and offers broad opportunities for theoretical developments and empirical analyses.
format Preprint
id arxiv_https___arxiv_org_abs_2504_15779
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Shannon invariants: A scalable approach to information decomposition
Gutknecht, Aaron J.
Rosas, Fernando E.
Ehrlich, David A.
Makkeh, Abdullah
Mediano, Pedro A. M.
Wibral, Michael
Information Theory
Artificial Intelligence
Machine Learning
Adaptation and Self-Organizing Systems
Data Analysis, Statistics and Probability
Distributed systems, such as biological and artificial neural networks, process information via complex interactions engaging multiple subsystems, resulting in high-order patterns with distinct properties across scales. Investigating how these systems process information remains challenging due to difficulties in defining appropriate multivariate metrics and ensuring their scalability to large systems. To address these challenges, we introduce a novel framework based on what we call "Shannon invariants" -- quantities that capture essential properties of high-order information processing in a way that depends only on the definition of entropy and can be efficiently calculated for large systems. Our theoretical results demonstrate how Shannon invariants can be used to resolve long-standing ambiguities regarding the interpretation of widely used multivariate information-theoretic measures. Moreover, our practical results reveal distinctive information-processing signatures of various deep learning architectures across layers, which lead to new insights into how these systems process information and how this evolves during training. Overall, our framework resolves fundamental limitations in analyzing high-order phenomena and offers broad opportunities for theoretical developments and empirical analyses.
title Shannon invariants: A scalable approach to information decomposition
topic Information Theory
Artificial Intelligence
Machine Learning
Adaptation and Self-Organizing Systems
Data Analysis, Statistics and Probability
url https://arxiv.org/abs/2504.15779