Reproduce “Evading Overlapping Community Detection via Proxy Node Injection” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Dario Loi, Matteo Silvestri, Fabrizio Silvestri, Gabriele Tolomei
Addressing the challenge of community membership hiding in overlapping graphs, the ODRL deep reinforcement learning framework employs strategic proxy node injection to alter network connections. The method successfully obscures target node affiliations, showing superior performance against baselines and robust transferability across different community detection algorithms.
Key insights
- Naïve proxy node injection strategies, effective for non-overlapping community hiding, significantly fail when applied to overlapping community structures.
- Deep Reinforcement Learning, when combined with strategic proxy node injection, provides a principled and effective mechanism for obscuring community membership in complex overlapping networks.
- A factored policy architecture is essential for managing the large action space in graph rewiring, enabling scalability and efficient learning for DRL agents.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.