Research map / Graph learning / GNN architectures
GNNEvaluator: Evaluating GNN Performance On Unseen Graphs Without Labels
Xin Zheng, Miao Zhang, Chunyang Chen, Soheila Molaei, Chuan Zhou, Shirui Pan
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 |
|---|---|---|---|
| GNNEvaluator instead of ATC-MC | ATC-MC · ACMv9->Citationv2 avg | MAE: 14.00 -> 10.71 | Improved |
| GNNEvaluator instead of AutoEval-G | AutoEval-G · ACMv9->DBLPv8 avg | MAE: 18.24 -> 12.86 | Improved |
| GNNEvaluator instead of ATC-MC | ATC-MC · Citationv2->ACMv9 avg | MAE: 22.02 -> 16.38 | Improved |
| GNNEvaluator instead of Thres. tau=0.9 | Thres. · Citationv2->DBLPv8 avg | MAE: 19.89 -> 10.71 | Improved |
| GNNEvaluator instead of AutoEval-G | AutoEval-G · DBLPv8->ACMv9 avg | MAE: 13.06 -> 6.79 | Improved |
| GNNEvaluator instead of AutoEval-G | AutoEval-G · DBLPv8->Citationv2 avg | MAE: 17.22 -> 7.80 | Improved |
| GNNEvaluator instead of ATC-MC-c on GCN, ACMv9->Citationv2 | ATC-MC-c · ACMv9->Citationv2 GCN | MAE: 2.41 -> 4.85 | Hurt |
| GNNEvaluator instead of ATC-MC-c on GCN, ACMv9->DBLPv8 | ATC-MC-c · ACMv9->DBLPv8 GCN | MAE: 31.15 -> 11.80 | Improved |
| replace DiscGraph set with meta-graph set (remove discrepancy node attributes) (ablation) | GNNEvaluator · GAT evaluation (5 seeds) | closeness to ground-truth accuracy | Hurt |
| use K=100 DiscGraphs instead of 200-400 | GNNEvaluator · Citationv2->ACMv9/DBLPv8 | MAE | Hurt |
| vary number of DiscGraphs K among 200, 300, 400 | GNNEvaluator · Citationv2->ACMv9/DBLPv8 | MAE | No clear effect |
Closest papers by meaning
- Online GNN Evaluation Under Test-time Graph Distribution Shifts (ICLR 2024 · similarity 0.92)
- Classic GNNs are Strong Baselines: Reassessing GNNs for Node Classification (NeurIPS 2024 · similarity 0.91)
- Can Classic GNNs Be Strong Baselines for Graph-level Tasks? Simple Architectures Meet Excellence (ICML 2025 · similarity 0.89)
- Improving GNN Calibration with Discriminative Ability: Insights and Strategies (AAAI 2024 · similarity 0.89)
- GC4NC: A Benchmark Framework for Graph Condensation on Node Classification with New Insights (NeurIPS 2025 · similarity 0.89)
- Generalization of Graph Neural Networks Is Robust to Model Mismatch (AAAI 2025 · similarity 0.89)
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.
