Self-supervised Graph Representation Learning via Bootstrapping
Self-supervised Graph Representation Learning via Bootstrapping
复制标题
DOI：
10.1016/j.neucom.2021.03.123
复制
发表时间：
2020-11
期刊：
影响因子：
6
通讯作者：
Feihu Che;Guohua Yang;Dawei Zhang;J. Tao;Pengpeng Shao;Tong Liu
中科院分区：
文献类型：
--
作者：
Feihu Che;Guohua Yang;Dawei Zhang;J. Tao;Pengpeng Shao;Tong Liu
图神经网络（GNN）将深度学习技术应用于图结构数据，并在图表示学习方面取得了令人满意的性能。然而，现有的GNN严重依赖于标记数据或精心设计的负样本。为了解决这些问题，我们提出了一种新的自监督图表示方法：深度图自举（DGB）。DGB由两个神经网络组成：在线网络和目标网络，它们的输入是初始图的不同增广视图。在线网络被训练来预测目标网络，而目标网络用在线网络的缓慢移动平均值更新，这意味着在线网络和目标网络可以相互学习。因此，所提出的DGB可以在无监督的方式下学习图表示而无需负示例。此外，我们总结了三种图结构数据的增强方法，并将其应用于DGB。在基准数据集上的实验表明，DGB的性能优于当前最先进的方法，以及增强方法如何影响性能。
Graph neural networks (GNNs) apply deep learning techniques to graph-structured data and have achieved promising performance in graph representation learning. However, existing GNNs rely heavily on labeled data or well-designed negative samples. To address these issues, we propose a new self-supervised graph representation method: deep graph bootstrapping (DGB). DGB consists of two neural networks: online and target networks, and the input of them are different augmented views of the initial graph. The online network is trained to predict the target network while the target network is updated with a slow-moving average of the online network, which means the online and target networks can learn from each other. As a result, the proposed DGB can learn graph representation without negative examples in an unsupervised manner. In addition, we summarize three kinds of augmentation methods for graph-structured data and apply them to the DGB. Experiments on the benchmark datasets show the DGB performs better than the current state-of-the-art methods and how the augmentation methods affect the performances.