Research map / Adversarial robustness and security / Attacks on text and graphs
ATEX-CF: Attack-Informed Counterfactual Explanations for Graph Neural Networks
Yu Zhang, Sean Bin Yang, Arijit Khan, Cuneyt Gurcan Akcora
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 |
|---|---|---|---|
| use ATEX-CF instead of Nettack | GCN · Cora | misclassification rate: 0.53 -> 0.72 | Improved |
| use ATEX-CF instead of GOttack | GCN · Cora | plausibility: 0.10 -> 0.75 | Improved |
| use ATEX-CF instead of PGExplainer | GCN · BA-SHAPES | misclassification rate: 0.73 -> 0.83 | Improved |
| use ATEX-CF instead of PGExplainer | GCN · Cora | time per explanation (s) | Hurt |
| use ATEX-CF instead of CF-GNNExplainer | 6 datasets · average rank | overall average rank: 4.67 -> 2.40 | Improved |
| raise perturbation budget from 1 to 15 (ablation) | ATEX-CF · Cora | misclassification rate: 0.46 -> 0.76 | Improved |
| raise addition cost C from 1.0 to 20 (ablation) | ATEX-CF · Cora, kappa=20 | misclassification rate: 0.70 -> 0.54 | Hurt |
| raise addition cost C from 1.0 to 21 (ablation) | ATEX-CF · Cora, kappa=20 | plausibility: 0.71 -> 0.62 | Hurt |
| add candidate-edge pruning (ablation) | ATEX-CF · Cora | time per explanation (s): 6.12 -> 3.00 | Improved |
| add candidate-edge pruning (ablation) | ATEX-CF · Cora | edge edits per explanation: 1.71 -> 1.62 | Improved |
| add candidate-edge pruning (ablation) | ATEX-CF · Cora | plausibility: 0.75 -> 0.76 | No clear effect |
| vary alpha_deg and alpha_motif from 0.5 to 1.5 (ablation) | ATEX-CF · Cora | fidelity and plausibility | No clear effect |
Closest papers by meaning
- COMRECGC: Global Graph Counterfactual Explainer through Common Recourse (ICML 2025 · similarity 0.88)
- Structure Your Data: Towards Semantic Graph Counterfactuals (ICML 2024 · similarity 0.87)
- Graph Neural Network Explanations are Fragile (ICML 2024 · similarity 0.87)
- Provably Robust Explainable Graph Neural Networks against Graph Perturbation Attacks (ICLR 2025 · similarity 0.86)
- Graph Inverse Style Transfer for Counterfactual Explainability (ICML 2025 · similarity 0.86)
- UNR-Explainer: Counterfactual Explanations for Unsupervised Node Representation Learning Models (ICLR 2024 · similarity 0.86)
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.
