arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1902.10409cs.LGstat.ML
收藏

基于加权内积的通用相似逼近表示学习

Representation Learning with Weighted Inner Product for Universal Approximation of General Similarities

Kyoto University(京都大学);RIKEN Center for Advanced Intelligence Project(理化学研究所先进智能项目中心)

Geewook Kim, Akifumi Okuno, Kazuki Fukui, Hidetoshi Shimodaira

摘要:

我们提出了$\textit{weighted inner product similarity}$(WIPS)的神经网络为基础的图嵌入。除了神经网络的参数外,我们还通过允许正值和负值来优化内积的权重。尽管它的简单性,WIPS可以近似任意的一般相似性,包括正定,条件正定,和不定的内核。WIPS不受相似性模型选择的限制,因为它可以学习任何相似性模型,如余弦相似性,负Poincaré距离和负Wasserstein距离。我们的实验表明,该方法可以学习高质量的分布式表示的节点从真实的数据集,导致一个准确的近似相似性以及高性能的归纳任务。

英文摘要:

We propose $\textit{weighted inner product similarity}$ (WIPS) for neural network-based graph embedding. In addition to the parameters of neural networks, we optimize the weights of the inner product by allowing positive and negative values. Despite its simplicity, WIPS can approximate arbitrary general similarities including positive definite, conditionally positive definite, and indefinite kernels. WIPS is free from similarity model selection, since it can learn any similarity models such as cosine similarity, negative Poincaré distance and negative Wasserstein distance. Our experiments show that the proposed method can learn high-quality distributed representations of nodes from real datasets, leading to an accurate approximation of similarities as well as high performance in inductive 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。

↑