Learning Interpretable Collective Variables of the Noisy Voter Model

Title data

Lücke, Marvin ; Winkelmann, Stefanie ; Heitzig, Jobst ; Molkenthin, Nora ; Koltai, Peter:
Learning Interpretable Collective Variables of the Noisy Voter Model.
arXiv , 2023
DOI: https://doi.org/10.48550/arXiv.2307.03491

Official URL:

Abstract in another language

We present a data-driven method to learn and understand collective variables for noisy voter model dynamics on networks. A collective variable (CV) is a projection of the high-dimensional system state into a low-dimensional space that preserves the essential dynamical information. Thus, CVs can be used to improve our understanding of complex emergent behaviors and to enable an easier analysis and prediction. We demonstrate our method using three example networks: the stochastic block model, a ring-shaped graph, and a scale-free network generated by the Albert--Barabási model. Our method combines the recent transition manifold approach with a linear regression step to produce interpretable CVs that describe the role and importance of each network node.

Further data

Item Type:	Preprint, postprint
Institutions of the University:	Faculties > Faculty of Mathematics, Physics und Computer Science > Department of Mathematics > Chair Dynamical Systems and Data > Chair Dynamical Systems and Data - Univ.-Prof. Dr. Peter Koltai
Result of work at the UBT:	No
DDC Subjects:	500 Science > 510 Mathematics 500 Science > 530 Physics
Date Deposited:	20 Jul 2023 05:24
Last Modified:	20 Jul 2023 05:24
URI:	https://eref.uni-bayreuth.de/id/eprint/86184