Source-linked AI summary
Small But Slow World: How Network Topology and Burstiness Slow Down Spreading
M. Karsai, M. Kivelä, R. K. Pan, K. Kaski, J. Kertész, A. -L. Barabási, J. Saramäki
TL;DR
Short topological paths do not ensure rapid spreading in communication networks, whose heterogeneous structure and temporal activity patterns complicate standard spreading models. Using time-stamped communication events, SI simulations, and null models that selectively destroy correlations, the paper finds that weight-topology correlations and bursty link activity are the main sources of slowing.
Problem
Short paths suggest rapid influence transmission, but spreading in real small-world networks is surprisingly slow, motivating analysis of structural and temporal heterogeneity.
Method
The study simulates SI spreading on time-stamped human communication event sequences and compares null models that selectively destroy structural and temporal correlations.
Results
Community structure with correlated link weights and bursty single-link activity are the main contributors to slow spreading, while daily patterns and inter-link event correlations play minor overall roles.
Takeaways & Limitations
Spreading speed in communication networks depends substantially on weight-topology correlations and inhomogeneous, bursty link activity rather than on short paths alone.
Takeaways & Limitations
The mean-field calculation illustrating burstiness has limitations.
Abstract
from arXiv · showhide
Communication networks show the small-world property of short paths, but the spreading dynamics in them turns out slow. We follow the time evolution of information propagation through communication networks by using the SI model with empirical data on contact sequences. We introduce null models where the sequences are randomly shuffled in different ways, enabling us to distinguish between the contributions of different impeding effects. The slowing down of spreading is found to be caused mostly by weight-topology correlations and the bursty activity patterns of individuals.