论文阅读 [TPAMI-2022] Learning on Attribute-Missing Graphs

论文阅读 [TPAMI-2022] Learning on Attribute-Missing Graphs

论文搜索(studyai.com)

搜索论文: Learning on Attribute-Missing Graphs

搜索论文: http://www.studyai.com/search/whole-site/?q=Learning+on+Attribute-Missing+Graphs

关键字(Keywords)

Task analysis; Gallium nitride; Generative adversarial networks; Convolution; Recurrent neural networks; Linear programming; graph neural network; attribute-missing graphs; distribution matching; node attribute completion; node classification; link prediction

机器学习; 机器视觉; 运筹与优化

生成对抗; 递归神经网络; 图网络; 图卷积网络; SAT问题

摘要(Abstract)

Graphs with complete node attributes have been widely explored recently.

最近，具有完整节点属性的图得到了广泛的研究。.

While in practice, there is a graph where attributes of only partial nodes could be available and those of the others might be entirely missing.

在实践中，有一个图，其中只有部分节点的属性可用，而其他节点的属性可能完全缺失。.

This attribute-missing graph is related to numerous real-world applications and there are limited studies investigating the corresponding learning problems.

该属性缺失图与许多实际应用有关，并且研究相应学习问题的研究有限。.

Existing graph learning methods including the popular GNN cannot provide satisfied learning performance since they are not specified for attribute-missing graphs.

现有的图学习方法（包括流行的GNN）无法提供令人满意的学习性能，因为它们没有针对属性缺失图进行指定。.

Thereby, designing a new GNN for these graphs is a burning issue to the graph learning community.

因此，为这些图形设计一个新的GNN是图形学习社区面临的一个紧迫问题。.

In this article, we make a shared-latent space assumption on graphs and develop a novel distribution matching-based GNN called structure-attribute transformer (SAT) for attribute-missing graphs.

在本文中，我们对图进行了共享潜在空间假设，并针对属性缺失图开发了一种新的基于分布匹配的GNN，称为结构-属性变换器（SAT）。.

SAT leverages structures and attributes in a decoupled scheme and achieves the joint distribution modeling of structures and attributes by distribution matching techniques.

SAT在解耦方案中利用结构和属性，并通过分布匹配技术实现结构和属性的联合分布建模。.

It could not only perform the link prediction task but also the newly introduced node attribute completion task.

它不仅可以执行链路预测任务，还可以执行新引入的节点属性完成任务。.

Furthermore, practical measures are introduced to quantify the performance of node attribute completion.

此外，还介绍了量化节点属性完成性能的实用方法。.

Extensive experiments on seven real-world datasets indicate SAT shows better performance than other methods on both link prediction and node attribute completion tasks…

在七个真实数据集上进行的大量实验表明，SAT在链路预测和节点属性完成任务方面都比其他方法表现出更好的性能。。.

作者(Authors)

[‘Xu Chen’, ‘Siheng Chen’, ‘Jiangchao Yao’, ‘Huangjie Zheng’, ‘Ya Zhang’, ‘Ivor W. Tsang’]