Papers › Self-Attention Graph Pooling
Self-Attention Graph Pooling
Junhyun Lee, Inyeop Lee, Jaewoo Kang
Advanced methods of applying deep learning to structured data such as graphs have been proposed in recent years. In particular, studies have focused on generalizing convolutional neural networks to graph data, which includes redefining the convolution and the downsampling (pooling) operations for graphs. The method of generalizing the convolution operation to graphs has been proven to improve performance and is widely used. However, the method of applying downsampling to graphs is still difficult to perform and has room for improvement. In this paper, we propose a graph pooling method based on self-attention. Self-attention using graph convolution allows our pooling method to consider both node features and graph topology. To ensure a fair comparison, the same training procedures and model architectures were used for the existing pooling methods and our method. The experimental results demonstrate that our method achieves superior graph classification performance on the benchmark datasets using a reasonable number of parameters.
In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.
Code
Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.
Code Syntology ran Syntology
Not run by Syntology. Nothing on this page verifies that the listed code works.
Tasks
Results from the paper archive 2025-07-28
| Task | Dataset | Model | Metric | Value | Rank at snapshot | Leaderboard | Report |
|---|---|---|---|---|---|---|---|
| Graph Classification | D&D | SAGPool_h | Accuracy | 76.45% | #37 of 53 | Archive leaderboard | report |
| Graph Classification | D&D | SAGPool_g | Accuracy | 76.19% | #40 of 53 | Archive leaderboard | report |
| Graph Classification | FRANKENSTEIN | SAGPool_g | Accuracy | 62.57 | #5 of 6 | Archive leaderboard | report |
| Graph Classification | FRANKENSTEIN | SAGPool_h | Accuracy | 61.73 | #6 of 6 | Archive leaderboard | report |
| Graph Classification | NCI1 | SAGPool_g | Accuracy | 74.06% | #54 of 69 | Archive leaderboard | report |
| Graph Classification | NCI1 | SAGPool_h | Accuracy | 67.45% | #66 of 69 | Archive leaderboard | report |
| Graph Classification | NCI109 | SAGPool_g | Accuracy | 74.06 | #31 of 38 | Archive leaderboard | report |
| Graph Classification | NCI109 | SAGPool_h | Accuracy | 67.86 | #37 of 38 | Archive leaderboard | report |
| Graph Classification | PROTEINS | SAGPool_h | Accuracy | 71.86% | #97 of 103 | Archive leaderboard | report |
| Graph Classification | PROTEINS | SAGPool_g | Accuracy | 70.04% | #100 of 103 | Archive leaderboard | report |
Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.
Methods
Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections