https://github.com/acbull/GPT-GNN

Code for KDD'20 "Generative Pre-Training of Graph Neural Networks"
https://github.com/acbull/GPT-GNN

graph-neural-networks graph-representation-learning pre-training self-supervised-learning

Last synced: 3 months ago
JSON representation

Code for KDD'20 "Generative Pre-Training of Graph Neural Networks"

• Host: GitHub
• URL: https://github.com/acbull/GPT-GNN
• Owner: acbull
• License: mit
• Created: 2020-06-26T21:15:27.000Z (over 4 years ago)
• Default Branch: master
• Last Pushed: 2024-07-25T08:07:50.000Z (4 months ago)
• Last Synced: 2024-07-25T09:32:01.145Z (4 months ago)
• Topics: graph-neural-networks, graph-representation-learning, pre-training, self-supervised-learning
• Language: Python
• Homepage:
• Size: 9.51 MB
• Stars: 478
• Watchers: 12
• Forks: 90
• Open Issues: 23
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

• StarryDivineSky - acbull/GPT-GNN - Training of Graph Neural Networks 图神经网络的生成式预训练。在预处理阶段，算法会首先随机地遮盖掉图中的一些边和点，利用生成模型来生成（预测）这些边的存在和节点的属性。模型的损失函数会使得预测的结果尽量接近真实的网络结构。这样的话，在GPT-GNN训练完成后，其内部的图神经网络层就可以被拿出来进行调优。 (图预训练 Pre-Training of Graph / 网络服务_其他)