Predicting Steady-State Behavior in Complex Networks with Graph Neural Networks

Priodyuti Pradhan, Amit Reza · 2025

Read Predicting Steady-State Behavior in Complex Networks with Graph Neural Networks by Priodyuti Pradhan, Amit Reza (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2502-01693.txt.

Description

Priodyuti Pradhan and Amit Reza teach graph neural networks to predict the settled behavior of a dynamical system on a network, in a 2025 paper. In complex systems, a disturbance can spread everywhere, linger weakly, or stay trapped near its source. Knowing which regime a network sits in matters for control, but solving the math case by case is slow.

Their framework joins graph convolution with attention, letting the model weigh which neighbors matter most. Trained on network examples, it sorted the three spread regimes with high accuracy. The team also tested it on real-world data and checked that the performance held. To open the black box, they derive the model's forward and backward signal flow in closed form, giving an analytical handle on why it decides as it does.

Catalog record

CREATORPriodyuti Pradhan, Amit Reza
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE48 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2502-01693

Browse the full library · About the Tower of Babel