Research map / Graph learning / GNN architectures
PlanE: Representation Learning over Planar Graphs
Radoslav Dimitrov, Zeyang Zhao, Ralph Abboud, Ismail Ilkan Ceylan
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 planar decomposition (BasePlanE) instead of standard GCN | message passing GNN · EXP | accuracy (%): 50.0 -> 100 | Improved |
| BasePlanE instead of random-node-initialized GCN | GCN-RNI(N) · EXP | accuracy (%): 98.0 -> 100 | Improved |
| BasePlanE instead of GIN | GIN · P3R | accuracy (%): 11.1 -> 100 | Improved |
| BasePlanE instead of 2-WL expressive PPGN | PPGN · P3R | accuracy (%): 100 -> 100 | No clear effect |
| BasePlanE instead of GIN | GIN · QM9_CC | MAE: 0.0612 -> 0.0023 | Improved |
| BasePlanE instead of ESAN | ESAN · QM9_CC | MAE: 0.0038 -> 0.0023 | Improved |
| GCN instead of constant prediction | constant baseline · QM9_CC | MAE: 0.1627 -> 0.1275 | Improved |
| E-BasePlanE instead of GIN | GIN · MolHIV | ROC-AUC: 77.07 -> 80.04 | Improved |
| E-BasePlanE instead of CIN | CIN · MolHIV | ROC-AUC: 80.94 -> 80.04 | No clear effect |
| E-BasePlanE instead of R-SPN k=10 | R-SPN · QM9 mu | MAE: 2.21 -> 1.97 | Improved |
| E-BasePlanE instead of R-SPN k=10 | R-SPN · QM9 ZPVE | MAE: 2.58 -> 2.81 | Hurt |
| E-BasePlanE instead of R-SPN k=10 | R-SPN · QM9 G | MAE: 0.83 -> 0.88 | Hurt |
Closest papers by meaning
- A Fair Comparison of Graph Neural Networks for Graph Classification (ICLR 2020 · similarity 0.87)
- PF-GNN: Differentiable particle filtering based approximation of universal graph representations (ICLR 2022 · similarity 0.87)
- Homomorphism Expressivity of Spectral Invariant Graph Neural Networks (ICLR 2025 · similarity 0.87)
- From Distribution to Geometry: Stable Graph Generalization via Invariant Barycenters (ICML 2026 · similarity 0.87)
- Exploring Consistency in Graph Representations: from Graph Kernels to Graph Neural Networks (NeurIPS 2024 · similarity 0.86)
- Homomorphism Counts for Graph Neural Networks: All About That Basis (ICML 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.
