Stuart-Landau Oscillatory Graph Neural Network

Kaicheng Zhang 1 , David N. Reynolds 2 , Piero Deidda 3,4 , Francesco Tudisco 1,3,5 · 2025

Read Stuart-Landau Oscillatory Graph Neural Network by Kaicheng Zhang 1 , David N. Reynolds 2 , Piero Deidda 3,4 , Francesco Tudisco 1,3,5 (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-2511-08094.txt.

Description

Kaicheng Zhang, David Reynolds, Piero Deidda, and Francesco Tudisco introduce a graph neural network built on Stuart Landau oscillators. Deep graph networks often suffer oversmoothing, where node features blur together, and vanishing gradients. Oscillatory designs borrow physics to fight both problems, but phase only models discard amplitude information that can carry useful signals.

The Stuart Landau model keeps amplitude and phase together, with tunable settings for the Hopf parameter and coupling strength. Node features then evolve with richer behavior, including amplitude control and multistable synchronization. Across node, graph, and regression tasks, the network beats earlier oscillatory graph models and sets a grounded new framework.

Catalog record

CREATORKaicheng Zhang 1 , David N. Reynolds 2 , Piero Deidda 3,4 , Francesco Tudisco 1,3,5
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE46 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2511-08094

Browse the full library · About the Tower of Babel