Spiking label propagation for community detection

Abstract

In this paper we present results from a method of community detection using label propagation in undirected, unweighted graphs which incorporates elements of neural computing and spike-based data. Using a fully connected, edge-weighted system of spiking neurons driven by external currents, we generate spike responses that are decoded into a binary signal. The similarity between pairs of signals is quantified using a Hamming-distance based metric and is used to classify vertices into communities. We test our approach on a set of graph instances, each with 128 vertices and either homogeneous or heterogeneous community size distributions. We present our method as a candidate for a split-computing workflow that incorporates neuromorphic hardware and does not require extensive pre-training of network parameters.

Version 2: 8 pages, 6 figures

References in corpus (7)

• Near linear time algorithm to detect community structures in large-scale networks
• Resolution limit in community detection
• Statistical Mechanics of Community Detection
• Benchmarks for testing community detection algorithms on directed and weighted graphs with overlapping communities
• Convolutional Networks for Fast, Energy-Efficient Neuromorphic Computing
• Detecting network communities by propagating labels under constraints
• Note on the equivalence of the label propagation method of community detection and a Potts model approach