Learning to Compress Graphs via Dual Agents for Consistent Topological Robustness Evaluation
Qisen Chai, Yansong Wang, Junjie Huang, Tao Jia
TL;DR
This work tackles the scalability challenge of evaluating robustness on large graphs by introducing Cutter, a dual-agent reinforcement learning framework that compresses graphs while preserving both topological structure and robustness profiles. A Vital Detection Agent and a Redundancy Detection Agent share a graph-convolutional encoder and employ task-specific heads, enhanced by trajectory-level reward shaping, prototype-based state–action shaping, and cross-agent active–follow exploration to improve learning efficiency and coordination. Empirical results on five real-world graphs show compressed graphs retain key topology and closely follow the original robustness degradation under multiple attacks, enabling faster, faithful robustness analysis. The authors also provide theoretical support via policy invariance proofs for reward-shaping schemes and demonstrate safe, transferable exploration strategies that facilitate scalable robustness-preserving graph compression.
Abstract
As graph-structured data grow increasingly large, evaluating their robustness under adversarial attacks becomes computationally expensive and difficult to scale. To address this challenge, we propose to compress graphs into compact representations that preserve both topological structure and robustness profile, enabling efficient and reliable evaluation. We propose Cutter, a dual-agent reinforcement learning framework composed of a Vital Detection Agent (VDA) and a Redundancy Detection Agent (RDA), which collaboratively identify structurally vital and redundant nodes for guided compression. Cutter incorporates three key strategies to enhance learning efficiency and compression quality: trajectory-level reward shaping to transform sparse trajectory returns into dense, policy-equivalent learning signals; prototype-based shaping to guide decisions using behavioral patterns from both high- and low-return trajectories; and cross-agent imitation to enable safer and more transferable exploration. Experiments on multiple real-world graphs demonstrate that Cutter generates compressed graphs that retain essential static topological properties and exhibit robustness degradation trends highly consistent with the original graphs under various attack scenarios, thereby significantly improving evaluation efficiency without compromising assessment fidelity.