Research map / Graph learning / GNN architectures
Explainability Methods for Graph Convolutional Neural Networks
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 CAM/Grad-CAM instead of CG for class-specific explanations | GCNN explanations · Country vs. Urban; Indoor vs. Outdoor; BBBP; BACE; TOX21 NR-ER | contrastivity: CAM/Grad-CAM: 99.99 ± 0.39, 100.00 ± 0.11, 99.99± 0.11, 100.0± 0.0, 99.99± 0.29; CG: 0.02 ± 0.45, 0.18 ± 1.88, 0.45± 2.1 | Helped |
| Use CG rather than CAM/Grad-CAM | GCNN explanations · Country vs. Urban; Indoor vs. Outdoor; BBBP; BACE; TOX21 NR-ER | fidelity: CG: 0.40, 0.40, 0.19, 0.38, 0.53; CAM/Grad-CAM: 0.25, 0.25, 0.17, 0.36, 0.11 | Helped |
| Use c-EB when sparse explanations are the priority | GCNN explanations · Country vs. Urban; Indoor vs. Outdoor; BBBP; BACE; TOX21 NR-ER | sparsity: c-EB: 60.92 ± 23.28, 64.31 ± 22.48, 40.35±22.11, 51.4±13.97, 30.12±23.04 | Helped |
| Use CG when occlusion-based fidelity is the priority | GCNN explanations · Across the reported datasets | fidelity | Helped |
| Use Grad-CAM on graph classification explanations | GCNN explanations · Across the reported datasets | contrastivity | Helped |
| Use c-EB for large graphs where localization matters | GCNN explanations · Across the reported datasets | sparsity | Helped |
Closest papers by meaning
- Graph Segmentation and Contrastive Enhanced Explainer for Graph Neural Networks (AAAI 2025 · similarity 0.88)
- Global Concept-Based Interpretability for Graph Neural Networks via Neuron Analysis (AAAI 2023 · similarity 0.88)
- Network Dissection: Quantifying Interpretability of Deep Visual Representations (CVPR 2017 · similarity 0.88)
- Self-Interpretable Graph Learning with Sufficient and Necessary Explanations (AAAI 2024 · similarity 0.88)
- Towards Robust Fidelity for Evaluating Explainability of Graph Neural Networks (ICLR 2024 · similarity 0.87)
- Empowering CAM-Based Methods with Capability to Generate Fine-Grained and High-Faithfulness Explanations (AAAI 2024 · similarity 0.87)
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.