Deep learning method reveals overlapping groups in complex networks
Diffusion-Induced Spatial Attention Overlapping Community Detection
Social and Information NetworksMachine Learning
Summary
Many networks have nodes that belong to multiple groups at the same time, which is tricky to figure out. The authors created a new deep learning method called DISCO that looks beyond immediate connections to find these overlapping groups more accurately. It uses a way to focus attention on important parts of the network based on how influence spreads and can work with different types of node information. They tested DISCO against other methods and found it works well, and also showed it can spot unusual changes in network structure, such as in computer security settings.
What this means in practice
- •For network security teams: Detect structural changes between network snapshots to identify anomalies and locate suspicious devices in communication systems.
- •For social network analysts: Identify overlapping user groups by modeling influence spreading beyond direct connections for improved community insights.
Authors
Kosti Koistinen, Vesa Kuikka, Joni Herttuainen, Matthew Hendren, Brian Holt, Kimmo K. Kaski
Abstract
Detection of overlapping communities is essential for modelling networks in which nodes participate simultaneously in multiple structural or functional groups. Existing graph neural network approaches commonly rely on local message passing, which can obscure community boundaries through smoothing and limit the representation of structurally relevant long-range dependencies. We introduce Diffusion-Induced Spatial Attention Community Detection (DISCO), a deep-learning framework that combines a structural prior derived from influence spreading dynamics, sparse multi-head attention, and non-negative community-affiliation learning. The prior identifies candidate interactions beyond immediate graph neighbours and biases attention according to their structural proximity, while a Bernoulli-Poisson edge-reconstruction objective enables overlapping community inference from node attributes and structural profiles, or both. Benchmark experiments show that DISCO performs competitively against established graph convolutional and graph attention approaches across different input configurations. To demonstrate its practical applicability, we present a proof-of-concept cybersecurity use case in which changes between community assignments inferred from consecutive communication-network snapshots provide an interpretable anomaly signal. Temporal community similarity identifies structural deviations, while node-level contributions help locate the devices associated with them. DISCO therefore provides both a flexible method for overlapping community detection and a foundation for analysing structural change in dynamic networks.