arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2103.12491cs.LG
收藏

扩展语义知识的Zero-shot图嵌入算法

Expanding Semantic Knowledge for Zero-shot Graph Embedding

University of Macau(澳门大学);University of Science and Technology Beijing(北京科技大学);Kwai Inc.(快手公司);Tsinghua University(清华大学)

Zheng Wang, Ruihang Shao, Changping Wang, Changjun Hu, Chaokun Wang, Zhiguo Gong

摘要:

Zero-shot图嵌入是有监督图学习的一个主要挑战。虽然最近的一种方法RECT已经显示出很好的性能,但其工作机制尚不清楚,仍然需要大量的训练数据。在本文中,我们深入了解RECT,并解决其基本限制。我们表明,其核心部分是GNN原型模型,其中类原型由其平均特征向量描述。因此,RECT将节点从原始输入特征空间映射到中间级别的语义空间,该语义空间将原始输入特征连接到可见和不可见的类。这一机制使区域教育和培训方案对可见和不可见的类别都能很好地发挥作用,但也减少了歧视。为了实现其全部潜力,我们提出了两个标签扩展策略。具体来说,除了扩展可见类的标记节点集之外,我们还可以扩展未见过类的标记节点集。在真实数据集上的实验验证了本文方法的优越性。

英文摘要:

Zero-shot graph embedding is a major challenge for supervised graph learning. Although a recent method RECT has shown promising performance, its working mechanisms are not clear and still needs lots of training data. In this paper, we give deep insights into RECT, and address its fundamental limits. We show that its core part is a GNN prototypical model in which a class prototype is described by its mean feature vector. As such, RECT maps nodes from the raw-input feature space into an intermediate-level semantic space that connects the raw-input features to both seen and unseen classes. This mechanism makes RECT work well on both seen and unseen classes, which however also reduces the discrimination. To realize its full potentials, we propose two label expansion strategies. Specifically, besides expanding the labeled node set of seen classes, we can also expand that of unseen classes. Experiments on real-world datasets validate the superiority of our methods.

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。

↑