Learning to Compress Graphs via Dual Agents for Consistent Topological Robustness Evaluation

TL;DR

This paper introduces Cutter, a dual-agent reinforcement learning framework that compresses large graphs into smaller, topologically faithful representations, enabling efficient and reliable robustness evaluation against adversarial attacks.

Contribution

The paper presents a novel dual-agent RL approach with innovative reward shaping and imitation strategies for graph compression that preserves robustness profiles.

Findings

Cutter produces compressed graphs maintaining key topological features.

Compressed graphs exhibit robustness degradation similar to original graphs under attacks.

The method significantly improves evaluation efficiency without losing fidelity.

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.