Research map / Graph learning / GNN architectures
Accurate Learning of Graph Representations with Graph Multiset Pooling
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 |
|---|---|---|---|
| Replace the strongest listed pooling baseline with GMT | GCN graph classifier · D&D | accuracy: 78.45 -> 78.72 | Helped |
| Replace the strongest listed pooling baseline with GMT | GCN graph classifier · PROTEINS | accuracy: 75.16 -> 75.09 | Hurt |
| Replace the strongest listed baseline with GMT | GCN graph classifier · MUTAG | accuracy: 81.39 -> 83.44 | Helped |
| Replace the strongest listed baseline with GMT | GCN graph classifier · HIV | ROC-AUC: 76.81 -> 77.56 | Helped |
| Replace the strongest listed pooling baseline with GMT | GCN graph classifier · Tox21 | ROC-AUC: 75.43 -> 77.30 | Helped |
| Replace the strongest listed pooling baseline with GMT | GCN graph classifier · ToxCast | ROC-AUC: 62.48 -> 65.44 | Helped |
| Replace the strongest listed pooling baseline with GMT | GCN graph classifier · BBBP | ROC-AUC: 68.25 -> 68.31 | Helped |
| Replace the strongest listed baseline with GMT | GCN graph classifier · IMDB-B | accuracy: 73.29 -> 73.48 | Helped |
| Replace the strongest listed pooling baseline with GMT | GCN graph classifier · IMDB-M | accuracy: 51.31 -> 50.66 | Hurt |
| Replace the strongest listed pooling baseline with GMT | GCN graph classifier · COLLAB | accuracy: 80.87 -> 80.74 | Hurt |
| Use graph attention and self-attention rather than mean pooling (ablation) | GMT graph classifier · Graph classification datasets | accuracy | Helped |
| Remove message-passing layers (ablation) | GMT pooling layer · Graph classification datasets | accuracy | No clear effect |
Closest papers by meaning
- Hierarchical Graph Representation Learning with Pooling-Induced Substructures (NeurIPS 2026 · similarity 0.93)
- ASAP: Adaptive Structure Aware Pooling for Learning Hierarchical Graph Representations (AAAI 2020 · similarity 0.92)
- Geometry-Aware Edge Pooling for Graph Neural Networks (NeurIPS 2025 · similarity 0.90)
- ENAHPool: The Edge-Node Attention-based Hierarchical Pooling for Graph Neural Networks (ICML 2025 · similarity 0.90)
- SkipPool: Improved Sparse Hierarchical Graph Pooling with Differentiable Exploration (AAAI 2025 · similarity 0.90)
- Graph Parsing Networks (ICLR 2024 · 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.