Abstract
This paper focuses on the problem of unsupervised alignment of hierarchicaldata such as ontologies or lexical databases. This is a problem that appearsacross areas, from natural language processing to bioinformatics, and istypically solved by appeal to outside knowledge bases and label-textualsimilarity. In contrast, we approach the problem from a purely geometricperspective: given only a vector-space representation of the items in the twohierarchies, we seek to infer correspondences across them. Our work derivesfrom and interweaves hyperbolic-space representations for hierarchical data, onone hand, and unsupervised word-alignment methods, on the other. We firstprovide a set of negative results showing how and why Euclidean methods fail inthis hyperbolic setting. We then propose a novel approach based on optimaltransport over hyperbolic spaces, and show that it outperforms standardembedding alignment techniques in various experiments on cross-lingual WordNetalignment and ontology matching tasks.