Research map / Graph learning / GNN architectures
On The Expressive Power of GNN Derivatives
Yam Eitan, Moshe Eliasof, Yoav Gelberg, Fabrizio Frasca, Guy Bar-Shalom, Haggai Maron
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 analytic activation in base MPNN instead of non-analytic (ablation) | HOD-GNN · synthetic substructure counting | MAE | Improved |
| use non-analytic activation instead of encoding-augmented MPNN (ablation) | HOD-GNN · synthetic substructure counting | MAE | Improved |
| increase maximum derivative order up to 4 (ablation) | HOD-GNN · synthetic substructure counting | MAE | Improved |
| increase maximum derivative order beyond 4 (ablation) | HOD-GNN · synthetic substructure counting | MAE | No clear effect |
| change backbone MPNN among GCN, GIN, GPS (ablation) | HOD-GNN · real-world datasets | accuracy/MAE | No clear effect |
| replace GIN + RWSE with HOD-GNN | GIN + RWSE · ZINC-12K | MAE: 0.128 -> 0.0666 | Improved |
| replace GCN + RWSE with HOD-GNN | GCN + RWSE · Peptides-struct | MAE: 0.2574 -> 0.2450 | Improved |
| replace GIN with HOD-GNN | GIN · ogbg-molhiv | ROC-AUC: 75.58 -> 80.86 | Improved |
| use HOD-GNN instead of Subgraphormer | Subgraphormer · ogbg-molbace | ROC-AUC: 84.35 -> 82.10 | Hurt |
| use HOD-GNN instead of sampling-based Subgraph GNNs on Peptides | HyMN · Peptides-func | AP: 68.57 -> 69.68 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- Subgraphormer — Subgraphormer: Unifying Subgraph GNNs and Graph Transformers via Graph Products (ICML 2024) · better on 0 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Beyond Weisfeiler-Lehman: A Quantitative Framework for GNN Expressiveness (ICLR 2024 · similarity 0.92)
- Towards Bridging Generalization and Expressivity of Graph Neural Networks (ICLR 2025 · similarity 0.91)
- What Expressivity Theory Misses: Message Passing Complexity for GNNs (NeurIPS 2025 · similarity 0.90)
- An Empirical Study of Realized GNN Expressiveness (ICML 2024 · similarity 0.89)
- On the Rademacher Complexity of Graph Neural Networks: Unifying Expressivity and Geometry (NeurIPS 2026 · similarity 0.89)
- Can Classic GNNs Be Strong Baselines for Graph-level Tasks? Simple Architectures Meet Excellence (ICML 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.
