Active Community Detection with Maximal Expected Model Change

TL;DR

This paper introduces an active learning algorithm for community detection in networks that selects nodes to query based on their potential to significantly alter community assignment models, improving detection accuracy especially in challenging sparse networks.

Contribution

The paper proposes the Maximal Expected Model Change (MEMC) algorithm for active community detection, with theoretical analysis and empirical validation on SBM and real networks.

Findings

MEMC outperforms random and state-of-the-art active learning methods.

Super-linear error reduction in sparse and below-detection-threshold SBMs.

Effective in both binary and multi-class community detection.

Abstract

We present a novel active learning algorithm for community detection on networks. Our proposed algorithm uses a Maximal Expected Model Change (MEMC) criterion for querying network nodes label assignments. MEMC detects nodes that maximally change the community assignment likelihood model following a query. Our method is inspired by detection in the benchmark Stochastic Block Model (SBM), where we provide sample complexity analysis and empirical study with SBM and real network data for binary as well as for the multi-class settings. The analysis also covers the most challenging case of sparse degree and below-detection-threshold SBMs, where we observe a super-linear error reduction. MEMC is shown to be superior to the random selection baseline and other state-of-the-art active learners.