
摘要
低维节点嵌入在大型图中的多种预测任务中已被证明极其有用,从内容推荐到识别蛋白质功能。然而,大多数现有的方法要求在训练嵌入时所有节点都必须存在于图中;这些先前的方法本质上是半监督的,无法自然地推广到未见过的节点。本文介绍了一种名为GraphSAGE的一般归纳框架,该框架利用节点特征信息(例如,文本属性)高效生成之前未见过的数据的节点嵌入。与为每个节点单独训练嵌入不同,我们学习了一个函数,通过采样和聚合节点局部邻域的特征来生成嵌入。我们的算法在三个归纳节点分类基准测试中优于强大的基线方法:我们在基于引用和Reddit帖子数据的演化信息图中对未见过的节点进行分类,并展示了我们的算法可以使用多图数据集中的蛋白质-蛋白质相互作用数据推广到完全未见过的图。
代码仓库
weiyinwei/mmgcn
pytorch
GitHub 中提及
Arvidsmile/InductiveDialogueGraph
tf
GitHub 中提及
silent-code/deep-trace
tf
GitHub 中提及
pmlg/pytorch_GCN
pytorch
GitHub 中提及
weiyinwei/huign
pytorch
GitHub 中提及
twjiang/graphSAGE-pytorch
pytorch
chatterjeeayan/upna
pytorch
GitHub 中提及
erfmah/answering_graph_queries
pytorch
GitHub 中提及
zxhhh97/ABot
pytorch
GitHub 中提及
arangoml/fastgraphml
pytorch
GitHub 中提及
stellargraph/stellargraph
tf
GitHub 中提及
qema/orca-py
pytorch
GitHub 中提及
hmcreamer/hackRice19
pytorch
GitHub 中提及
williamleif/GraphSAGE
官方
tf
GitHub 中提及
isotlaboratory/ml4vrp
pytorch
GitHub 中提及
IllinoisGraphBenchmark/IGB-Datasets
pytorch
GitHub 中提及
massquantity/LibRecommender
tf
GitHub 中提及
基准测试
| 基准 | 方法 | 指标 |
|---|---|---|
| graph-classification-on-cifar10-100k | GraphSage | Accuracy (%): 66.08 |
| graph-regression-on-zinc-500k | GraphSage | MAE: 0.398 |
| graph-regression-on-zinc-full | GraphSAGE | Test MAE: 0.126±0.003 |
| link-property-prediction-on-ogbl-citation2 | NeighborSampling (SAGE aggr) | Ext. data: No Number of params: 460289 Test MRR: 0.8044 ± 0.0010 Validation MRR: 0.8054 ± 0.0009 |
| link-property-prediction-on-ogbl-citation2 | Full-batch GraphSAGE | Ext. data: No Number of params: 460289 Test MRR: 0.8260 ± 0.0036 Validation MRR: 0.8263 ± 0.0033 |
| link-property-prediction-on-ogbl-collab | GraphSAGE (val as input) | Ext. data: No Number of params: 460289 Test Hits@50: 0.5463 ± 0.0112 Validation Hits@50: 0.5688 ± 0.0077 |
| link-property-prediction-on-ogbl-collab | GraphSAGE | Ext. data: No Number of params: 460289 Test Hits@50: 0.4810 ± 0.0081 Validation Hits@50: 0.5688 ± 0.0077 |
| link-property-prediction-on-ogbl-collab | GraphSAGE (val as input) | Number of params: 460289 |
| link-property-prediction-on-ogbl-ddi | GraphSAGE | Ext. data: No Number of params: 1421057 Test Hits@20: 0.5390 ± 0.0474 Validation Hits@20: 0.6262 ± 0.0037 |
| link-property-prediction-on-ogbl-ppa | GraphSAGE | Ext. data: No Number of params: 424449 Test Hits@100: 0.1655 ± 0.0240 Validation Hits@100: 0.1724 ± 0.0264 |
| node-classification-on-brazil-air-traffic | GraphSAGE (Hamilton et al., [2017a]) | Accuracy: 0.404 |
| node-classification-on-chameleon-60-20-20 | GraphSAGE | 1:1 Accuracy: 62.15 ± 0.42 |
| node-classification-on-citeseer-05 | GraphSAGE | Accuracy: 33.8% |
| node-classification-on-citeseer-1 | GraphSAGE | Accuracy: 51.0% |
| node-classification-on-citeseer-60-20-20 | GraphSAGE | 1:1 Accuracy: 78.24 ± 0.30 |
| node-classification-on-citeseer-full | GraphSAGE | Accuracy: 71.40% |
| node-classification-on-citeseer-with-public | GraphSAGE | Accuracy: 67.2 |
| node-classification-on-cora-05 | GraphSAGE | Accuracy: 37.5% |
| node-classification-on-cora-1 | GraphSAGE | Accuracy: 49.0% |
| node-classification-on-cora-3 | GraphSAGE | Accuracy: 64.2% |
| node-classification-on-cora-60-20-20-random | GraphSAGE | 1:1 Accuracy: 86.58 ± 0.26 |
| node-classification-on-cora-full-supervised | GraphSAGE | Accuracy: 82.2% |
| node-classification-on-cora-with-public-split | GraphSAGE | Accuracy: 74.5% |
| node-classification-on-cornell-60-20-20 | GraphSAGE | 1:1 Accuracy: 71.41 ± 1.24 |
| node-classification-on-europe-air-traffic | GraphSAGE (Hamilton et al., [2017a]) | Accuracy: 27.2 |
| node-classification-on-facebook | GraphSAGE (Hamilton et al., [2017a]) | Accuracy: 38.9 |
| node-classification-on-film-60-20-20-random | GraphSAGE | 1:1 Accuracy: 36.37 ± 0.21 |
| node-classification-on-flickr | GraphSAGE (Hamilton et al., [2017a]) | Accuracy: 0.641 |
| node-classification-on-non-homophilic | GraphSAGE | 1:1 Accuracy: 71.41 ± 1.24 |
| node-classification-on-non-homophilic-1 | GraphSAGE | 1:1 Accuracy: 64.85 ± 5.14 |
| node-classification-on-non-homophilic-2 | GraphSAGE | 1:1 Accuracy: 79.03 ± 1.20 |
| node-classification-on-non-homophilic-4 | GraphSAGE | 1:1 Accuracy: 62.15 ± 0.42 |
| node-classification-on-pattern-100k | GraphSage | Accuracy (%): 50.516 |
| node-classification-on-ppi | GraphSAGE | F1: 61.2 |
| node-classification-on-pubmed-003 | GraphSAGE | Accuracy: 45.4% |
| node-classification-on-pubmed-005 | GraphSAGE | Accuracy: 53.0% |
| node-classification-on-pubmed-01 | GraphSAGE | Accuracy: 65.4% |
| node-classification-on-pubmed-60-20-20-random | GraphSAGE | 1:1 Accuracy: 86.85 ± 0.11 |
| node-classification-on-pubmed-full-supervised | GraphSAGE | Accuracy: 87.1% |
| node-classification-on-pubmed-with-public | GraphSAGE | Accuracy: 76.8% |
| node-classification-on-reddit | GraphSAGE | Accuracy: 94.32% |
| node-classification-on-squirrel-60-20-20 | GraphSAGE | 1:1 Accuracy: 41.26 ± 0.26 |
| node-classification-on-texas-60-20-20-random | GraphSAGE | 1:1 Accuracy: 79.03 ± 1.20 |
| node-classification-on-usa-air-traffic | GraphSAGE (Hamilton et al., [2017a]) | Accuracy: 31.6 |
| node-classification-on-wiki-vote | GraphSAGE (Hamilton et al., [2017a]) | Accuracy: 24.5 |
| node-classification-on-wisconsin-60-20-20 | GraphSAGE | 1:1 Accuracy: 64.85 ± 5.14 |