arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1902.09192cs.LGcs.AIstat.ML
收藏

图卷积网络的批量虚拟对抗训练

Batch Virtual Adversarial Training for Graph Convolutional Networks

Tsinghua University(清华大学)

Zhijie Deng, Yinpeng Dong, Jun Zhu

摘要:

提出了一种新的图卷积网络(GCN)正则化方法--批量虚拟对抗训练(BVAT)。BVAT解决了GCN不考虑模型输出分布对输入周围局部扰动的平滑性的缺点。我们提出了两种算法,基于样本的BVAT和基于优化的BVAT,这是适合于促进图结构数据的模型的平滑性,无论是找到虚拟对抗扰动的一个子集的节点远离对方或生成虚拟对抗扰动的所有节点的优化过程。在Cora、Citeseer和Pubmed三个引文网络数据集和一个知识图谱数据集Nell上的实验验证了该方法的有效性,在半监督节点分类任务中取得了最佳的结果.

英文摘要:

We present batch virtual adversarial training (BVAT), a novel regularization method for graph convolutional networks (GCNs). BVAT addresses the shortcoming of GCNs that do not consider the smoothness of the model's output distribution against local perturbations around the input. We propose two algorithms, sample-based BVAT and optimization-based BVAT, which are suitable to promote the smoothness of the model for graph-structured data by either finding virtual adversarial perturbations for a subset of nodes far from each other or generating virtual adversarial perturbations for all nodes with an optimization process. Extensive experiments on three citation network datasets Cora, Citeseer and Pubmed and a knowledge graph dataset Nell validate the effectiveness of the proposed method, which establishes state-of-the-art results in the semi-supervised node classification tasks.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑