Research map / Adversarial robustness and security / Attacks on text and graphs
Learning to Compress Graphs via Dual Agents for Consistent Topological Robustness Evaluation
Qisen Chai, Yansong Wang, Junjie Huang, Tao Jia
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add reward network shaping to RDA training (ablation) | Cutter RDA · RDA training | average return: 0.513 -> 0.668 | Improved |
| use Cutter instead of GEC at moderate compression | Citeseer rho=0.5 · Citeseer RPS | RPS: 0.958 -> 0.963 | No clear effect |
| use Cutter instead of MCGS at rho=0.5 | Citeseer · Citeseer RPS | RPS: 0.966 -> 0.963 | Hurt |
| use Cutter instead of GEC at extreme compression | Coauthor-Physics rho=0.1 · Coauthor-Physics RPS | RPS: 0.497 -> 0.758 | Improved |
| use Cutter instead of SparRL at extreme compression | Coauthor-Physics rho=0.1 · Coauthor-Physics RPS | RPS: 0.512 -> 0.758 | Improved |
| use Cutter instead of GEC for path length at rho=0.5 | Cora topology · Cora topology | path length difference: 1.045 -> 0.317 | Improved |
| use Cutter instead of HySamp for degree difference | Cora topology · Cora topology | degree difference: 0.145 -> 0.167 | Hurt |
| increase compression from rho=0.5 to rho=0.1 for Cutter | Cutter on Cora · Cora RPS | RPS: 0.953 -> 0.832 | Hurt |
Closest papers by meaning
- Dynamic Deep Factor Graph for Multi-Agent Reinforcement Learning (TPAMI 2025 · similarity 0.85)
- Graph Agent Network: Empowering Nodes with Inference Capabilities for Adversarial Resilience (AAAI 2025 · similarity 0.85)
- When Witnesses Defend: A Witness Graph Topological Layer for Adversarial Graph Learning (AAAI 2025 · similarity 0.85)
- Robust Graph Condensation via Classification Complexity Mitigation (NeurIPS 2025 · similarity 0.84)
- Graph-Assisted Stitching for Offline Hierarchical Reinforcement Learning (ICML 2025 · similarity 0.84)
- Towards Efficient Collaboration via Graph Modeling in Reinforcement Learning (AAAI 2025 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
