Research map / Graph learning / GNN architectures
Weisfeiler and Lehman Go Topological: Message Passing Simplicial Networks
Read the paperCheck 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 triangle-aware simplicial message passing instead of a GNN unaware of triangles | MPSN versus GNN L0-inv · Synthetic Flow test | accuracy: 61.0±4.2 -> 85.3±5.8 | Helped |
| Use triangle-aware simplicial message passing instead of a GNN unaware of triangles | MPSN versus GNN L0-inv · Ocean Drifters test | accuracy: 63.5±6.0 -> 71.5±4.1 | Helped |
| Add clique-complex simplicial processing | SIN · TUDatasets graph classification | accuracy: 76.5 ± 3.4 vs 77.2±4.7 (Proteins); 82.8 ± 2.2 vs 86.0±1.8 (NCI1); 75.6 ± 3.2 vs 77.8 ± 3.3 (IMDB-B); 52.5 ± 3.0 vs 54.3 | Mixed |
| Use an orientation-equivariant Id activation model instead of MPSN L0-inv (ablation) | MPSN · Synthetic Flow test | accuracy: 85.3±5.8 -> 82.6±3.0 | No clear effect |
| Use an orientation-equivariant Id activation model instead of MPSN L0-inv (ablation) | MPSN · Ocean Drifters test | accuracy: 71.5±4.1 -> 73.0±2.7 | No clear effect |
| Use Tanh activation in the orientation-equivariant model (ablation) | MPSN · Synthetic Flow test | accuracy: 85.3±5.8 -> 95.2±1.8 | Helped |
| Use Tanh activation in the orientation-equivariant model (ablation) | MPSN · Ocean Drifters test | accuracy: 71.5±4.1 -> 72.5±0.0 | No clear effect |
| Replace an odd orientation-equivariant activation with ReLU (ablation) | MPSN · Synthetic Flow test | accuracy: 95.2±1.8 -> 50.0±0.0 | Hurt |
| Replace an odd orientation-equivariant activation with ReLU (ablation) | MPSN · Ocean Drifters test | accuracy: 72.5±0.0 -> 46.5±5.7 | Hurt |
Closest papers by meaning
- Demystifying Topological Message-Passing with Relational Structures: A Case Study on Oversquashing in Simplicial Message-Passing (ICLR 2025 · similarity 0.90)
- Weisfeiler and Lehman Go Paths: Learning Topological Features via Path Complexes (AAAI 2024 · similarity 0.89)
- Weisfeiler-Leman at the margin: When more expressivity matters (ICML 2024 · similarity 0.88)
- Topological Blindspots: Understanding and Extending Topological Deep Learning Through the Lens of Expressivity (ICLR 2025 · similarity 0.87)
- Homomorphism Expressivity of Spectral Invariant Graph Neural Networks (ICLR 2025 · similarity 0.87)
- TopoTune: A Framework for Generalized Combinatorial Complex Neural Networks (ICML 2025 · 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.