Low-Dimensional Hyperbolic Knowledge Graph Embeddings (ACL 2020)

We propose a family of methods to learn low-dimensional knowledge graph embeddings that simultaneously capture hierarchical relations (e.g. father of) and logical patterns (e.g. symmetry or anti-symmetry).
Paper: arxiv.org/abs/2005.00545
PyTorch Code: github.com/HazyResearch/KGEmb
TF Code: github.com/tensorflow/neural-...

[@julessci2716]

Nice ideas