AI Paper Atlas

Research map / Graph learning / GNN architectures

Self-Attention Graph Pooling

Junhyun Lee, Inyeop Lee, Jaewoo Kang

ICML 2019 · Graph learning › GNN architectures · approach: Graph structure learning · 370 citations (OpenAlex)

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 changedWhereMeasuredEffect
Use self-attention graph pooling instead of gPoolHierarchical graph classification · D&Daccuracy: 75.01 ± 0.86 -> 76.45 ± 0.97Improved
Use self-attention graph pooling instead of gPoolHierarchical graph classification · PROTEINSaccuracy: 71.10 ± 0.90 -> 71.86 ± 0.97Improved
Use self-attention graph pooling instead of DiffPoolHierarchical graph classification · NCI1accuracy: 62.32 ± 1.90 -> 67.45 ± 1.11Improved
Use self-attention graph pooling instead of DiffPoolHierarchical graph classification · NCI109accuracy: 61.98 ± 1.98 -> 67.86 ± 1.41Improved
Use self-attention graph pooling instead of gPoolHierarchical graph classification · FRANKENSTEINaccuracy: 61.46 ± 0.84 -> 61.73 ± 0.76No clear effect
Use self-attention graph pooling instead of SortPoolGlobal graph classification · D&Daccuracy: 72.53 ± 1.19 -> 76.19 ± 0.94Improved
Use self-attention graph pooling instead of SortPoolGlobal graph classification · PROTEINSaccuracy: 66.72 ± 3.56 -> 70.04 ± 1.47Improved
Use self-attention graph pooling instead of SortPoolGlobal graph classification · NCI1accuracy: 73.82 ± 0.96 -> 74.18 ± 1.20No clear effect
Use self-attention graph pooling instead of SortPoolGlobal graph classification · NCI109accuracy: 74.02 ± 1.18 -> 74.06 ± 0.78No clear effect
Use self-attention graph pooling instead of SortPoolGlobal graph classification · FRANKENSTEINaccuracy: 60.61 ± 0.77 -> 62.57 ± 0.60Improved
Add two-hop edges using adjacency augmentation (ablation)SAGPool_h · D&Daccuracy: 76.45 ± 0.97 -> 77.07 ± 0.82Improved
Add two-hop edges using adjacency augmentation (ablation)SAGPool_h · PROTEINSaccuracy: 71.86 ± 0.97 -> 71.82 ± 0.81No clear effect

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy