Abstract:
Knowledge graph (KG) embedding models have recently gained increased attention.
However, most of the existing models for KG embeddings ignore the structure and
characteristics of the underlying ontology. KGs are not always representative of the
underlying configuration knowledge, they tend to capture the semantics at higher level.
However, Ontologies are much generalized semantic data models which can capture more
complex relationships between entities than KGs. This research work proposes EmEL++ embeddings – an ontology-based embedding model for theories in Description Logic EL++. EmEL++ maps the classes and relations in an ontology to an n-dimensional vector space such that the relations between classes and relations in the ontology are preserved in the vector space. We evaluate the proposed embeddings on six different datasets and show that the proposed embeddings outperform the traditional knowledge graph embeddings on the subsumption reasoning task.