Multi-label Learning with a Cone-Based Geometric Model

Abstract

Recent approaches for knowledge-graph embeddings aim at connecting quantitative data structures used in machine learning to the qualitative structures of logics. Such embeddings are of a hybrid nature, they are data models that also exhibit conceptual structures inherent to logics. One motivation to investigate embeddings is to design conceptually adequate machine learning (ML) algorithms. This paper investigates a new approach to embedding ontologies into geometric models that interpret concepts by closed convex cones. As a proof of concept this cone-based embedding was implemented in a ML algorithm for weak supervised multi-label learning. The system was tested with the gene ontology and showed a performance similar to comparable approaches, but with the advantage of exhibiting the conceptual structure underlying the data.