BriefGPT.xyz
Feb, 2022
使用随机拉普拉斯特征在双曲空间中学习
HyLa: Hyperbolic Laplacian Features For Graph Learning
HTML
PDF
Tao Yu, Christopher De Sa
TL;DR
本文介绍了一种利用超几何空间的嵌入方法提高图数据的建模效率和性能的新算法，以及如何通过Laplace算子的本征函数来逼近超几何空间中的等度不变核，从而更好地实现了超几何网络的建模。
Abstract
Due to its geometric properties,
hyperbolic space
can support high-fidelity embeddings of tree- and graph-structured data. For graph learning, points in
hyperbolic space
have been used successfully as signals in
→