From Community Detection to Community Deception

TL;DR

This paper introduces the community deception problem, proposing algorithms to rewire network connections to hide communities from detection algorithms, highlighting vulnerabilities and inspiring the development of more robust detection methods.

Contribution

It formalizes the community deception problem and presents algorithms to effectively rewire connections, demonstrating how existing detection methods can be deceived.

Findings

Existing community detection algorithms can be deceived using proposed rewiring techniques

A deception score quantifies the effectiveness of deception methods

The study raises awareness for designing detection algorithms robust to deception

Abstract

The community deception problem is about how to hide a target community C from community detection algorithms. The need for deception emerges whenever a group of entities (e.g., activists, police enforcements) want to cooperate while concealing their existence as a community. In this paper we introduce and formalize the community deception problem. To solve this problem, we describe algorithms that carefully rewire the connections of C's members. We experimentally show how several existing community detection algorithms can be deceived, and quantify the level of deception by introducing a deception score. We believe that our study is intriguing since, while showing how deception can be realized it raises awareness for the design of novel detection algorithms robust to deception techniques.