
摘要
图神经网络(GNNs)是一种有效的图表示学习框架。GNNs 采用邻域聚合方案,通过递归地聚合和转换节点的邻域节点表示向量来计算该节点的表示向量。许多 GNN 变体已被提出,并在节点分类和图分类任务中取得了最先进的结果。然而,尽管 GNNs 在图表示学习领域带来了革命性的变化,但对其表示性质和局限性的理解仍然有限。本文提出了一种理论框架,用于分析 GNNs 捕捉不同图结构的表达能力。我们的研究结果表征了流行的 GNN 变体(如图卷积网络和 GraphSAGE)的判别能力,并表明它们无法学会区分某些简单的图结构。随后,我们开发了一种简单架构,该架构在 GNN 类中具有最高的表达能力,并且其威力等同于 Weisfeiler-Lehman 图同构测试。我们在多个图分类基准数据集上对我们的理论发现进行了实证验证,并展示了我们的模型达到了最先进的性能。
代码仓库
PurdueMINDS/RelationalPooling
pytorch
GitHub 中提及
weihua916/powerful-gnns
官方
pytorch
GitHub 中提及
igorsterner/commute-opt-gnn
pytorch
GitHub 中提及
zhliping/Deep-Learning
pytorch
GitHub 中提及
gmum/umwpl2021
GitHub 中提及
guillaumejaume/tuto-dl-on-graphs
pytorch
GitHub 中提及
gospodima/extended-simgnn
pytorch
GitHub 中提及
willy-b/tiny-GIN-for-ogbg-molhiv
pytorch
GitHub 中提及
mfmceneaney/Lambda-GNNs
pytorch
GitHub 中提及
k4my4r/ECG-Classification
pytorch
GitHub 中提及
yuwvandy/g2gnn
pytorch
GitHub 中提及
egyptdj/graph-neural-mapping
pytorch
GitHub 中提及
pfnet-research/treewidth-prediction
GitHub 中提及
lukecavabarrett/pna
pytorch
calciver/Graph-Isomorphism-Networks
tf
GitHub 中提及
karolismart/dropgnn
pytorch
GitHub 中提及
基准测试
| 基准 | 方法 | 指标 |
|---|---|---|
| graph-classification-on-bp-fmri-97 | GIN | Accuracy: 45.4% F1: 42.3% |
| graph-classification-on-cifar10-100k | GIN | Accuracy (%): 53.28 |
| graph-classification-on-collab | GIN-0 | Accuracy: 80.2% |
| graph-classification-on-cox2 | GIN-0 | Accuracy(10-fold): 81.13 |
| graph-classification-on-dd | GIN | Accuracy: 77.311±2.223 |
| graph-classification-on-enzymes | GIN | Accuracy: 68.303±4.170 |
| graph-classification-on-hiv-dti-77 | GIN | Accuracy: 55.1% F1: 53.6% |
| graph-classification-on-hiv-fmri-77 | GIN | Accuracy: 52.5% F1: 35.6% |
| graph-classification-on-imdb-b | GIN-0 | Accuracy: 75.1% |
| graph-classification-on-imdb-b | GIN | Accuracy: 81.250±3.775 |
| graph-classification-on-imdb-m | GIN-0 | Accuracy: 52.3% |
| graph-classification-on-mutag | GIN-0 | Accuracy: 89.4% |
| graph-classification-on-nci1 | GIN | Accuracy: 84.818±0.936 |
| graph-classification-on-nci1 | GIN-0 | Accuracy: 82.7% |
| graph-classification-on-nci109 | GIN | Accuracy: 84.155±0.812 |
| graph-classification-on-peptides-func | GIN | AP: 0.6043±0.0216 |
| graph-classification-on-proteins | GIN | Accuracy: 75.536±1.851 |
| graph-classification-on-proteins | GIN-0 | Accuracy: 76,2% |
| graph-classification-on-ptc | GIN-0 | Accuracy: 64.40% |
| graph-classification-on-re-m5k | GIN-0 | Accuracy: 57.5% |
| graph-classification-on-reddit-b | GIN-0 | Accuracy: 92.4 |
| graph-property-prediction-on-ogbg-code2 | GIN+virtual node | Ext. data: No Number of params: 13841815 Test F1 score: 0.1581 ± 0.0026 Validation F1 score: 0.1439 ± 0.0020 |
| graph-property-prediction-on-ogbg-code2 | GIN | Ext. data: No Number of params: 12390715 Test F1 score: 0.1495 ± 0.0023 Validation F1 score: 0.1376 ± 0.0016 |
| graph-property-prediction-on-ogbg-molhiv | GIN+virtual node | Ext. data: No Number of params: 3336306 Test ROC-AUC: 0.7707 ± 0.0149 Validation ROC-AUC: 0.8479 ± 0.0068 |
| graph-property-prediction-on-ogbg-molhiv | GIN | Ext. data: No Number of params: 1885206 Test ROC-AUC: 0.7558 ± 0.0140 Validation ROC-AUC: 0.8232 ± 0.0090 |
| graph-property-prediction-on-ogbg-molpcba | GIN+virtual node | Ext. data: No Number of params: 3374533 Test AP: 0.2703 ± 0.0023 Validation AP: 0.2798 ± 0.0025 |
| graph-property-prediction-on-ogbg-molpcba | GIN | Ext. data: No Number of params: 1923433 Test AP: 0.2266 ± 0.0028 Validation AP: 0.2305 ± 0.0027 |
| graph-property-prediction-on-ogbg-ppa | GIN+virtual node | Ext. data: No Number of params: 3288042 Test Accuracy: 0.7037 ± 0.0107 Validation Accuracy: 0.6678 ± 0.0105 |
| graph-property-prediction-on-ogbg-ppa | GIN | Ext. data: No Number of params: 1836942 Test Accuracy: 0.6892 ± 0.0100 Validation Accuracy: 0.6562 ± 0.0107 |
| graph-regression-on-esr2 | GIN | R2: 0.668±0.000 RMSE: 0.509±0.668 |
| graph-regression-on-f2 | GIN | R2: 0.887±0.000 RMSE: 0.342±0.887 |
| graph-regression-on-kit | GIN | R2: 0.833±0.000 RMSE: 0.444±0.833 |
| graph-regression-on-lipophilicity | GIN | R2: 0.819±0.007 RMSE: 0.537±0.010 |
| graph-regression-on-parp1 | GIN | R2: 0.922±0.000 RMSE: 0.349±0.922 |
| graph-regression-on-pcqm4mv2-lsc | GIN | Test MAE: 0.1218 Validation MAE: 0.1195 |
| graph-regression-on-pgr | GIN | R2: 0.696±0.000 RMSE: 0.532±0.696 |
| graph-regression-on-zinc-500k | GIN | MAE: 0.526 |
| graph-regression-on-zinc-full | GIN | Test MAE: 0.068±0.004 |
| molecular-property-prediction-on-esol | GIN | R2: 0.938±0.011 RMSE: 0.509±0.044 |
| molecular-property-prediction-on-freesolv | GIN | R2: 0.964±0.008 RMSE: 0.744±0.083 |
| node-classification-on-pattern-100k | GIN | Accuracy (%): 85.590 |