Research map / Graph learning / GNN architectures
Counting Graph Substructures with Graph Neural Networks
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Characterizes the ability of GNNs to count and list graph substructures, giving a theoretical account of representation and generalization power in terms of substructure counting.
Problem
Open questions remain about whether GNNs can count and list substructures, which matter in biological and social networks.
Approach
Characterizes GNN representation and generalization power through their ability to count and list substructures, filling a gap in the theory of GNN expressivity. Full method details are unavailable beyond the abstract.
Closest papers by meaning
- Can Graph Neural Networks Count Substructures? (NeurIPS 2020 · similarity 0.95)
- Homomorphism Counts for Graph Neural Networks: All About That Basis (ICML 2024 · similarity 0.92)
- A Flexible, Equivariant Framework for Subgraph GNNs via Graph Products and Graph Coarsening (NeurIPS 2024 · similarity 0.89)
- Deep Homomorphism Networks (NeurIPS 2024 · similarity 0.89)
- Union Subgraph Neural Networks (AAAI 2024 · similarity 0.88)
- Beyond Message Passing: Neural Graph Pattern Machine (ICML 2025 · similarity 0.88)
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.