Source-linked AI summary
Noise bridges dynamical correlation and topology in coupled oscillator networks
Jie Ren, Wen-Xu Wang, Baowen Li, Ying-Cheng Lai
TL;DR
Inferring interaction patterns from dynamical time series remains difficult without knowledge of nodal dynamics, particularly under strong synchronization. This paper derives a noise-induced correspondence between dynamical correlation and network topology, then validates topology reconstruction across diverse dynamics and networks. The method accurately predicts full topology for undirected networks, while directed networks permit inference of local in-degree rather than global structure.
Problem
Inferring interaction patterns from measured dynamics is challenging without knowledge of nodal dynamics, and strong synchronization can make interactions impossible to extract from measurements.
Method
The paper derives a noise-induced relationship between dynamical correlation and the connection matrix, reconstructing the Laplacian from time-series correlation using a pseudo inverse.
Results
The method predicts full network topology with uniformly high success rates across four dynamical systems and model and real-world networks, particularly for undirected networks.
Takeaways & Limitations
Noise enables accurate and efficient network inference from measured correlations without controlling nodal dynamics or knowing the nodal dynamics.
Takeaways & Limitations
For directed networks, correlation alone cannot determine the global structure, but it can infer each node’s in-degree.
Abstract
from arXiv · showhide
We study the relationship between dynamical properties and interaction patterns in complex oscillator networks in the presence of noise. A striking finding is that noise leads to a general, one-to-one correspondence between the dynamical correlation and the connections among oscillators for a variety of node dynamics and network structures. The universal finding enables an accurate prediction of the full network topology based solely on measuring the dynamical correlation. The power of the method for network inference is demonstrated by the high success rate in identifying links for distinct dynamics on both model and real-life networks. The method can have potential applications in various fields due to its generality, high accuracy and efficiency.