SiMHOMer: Siamese Models for Health Ontologies Merging and Validation Through Large Language Models

Ontologies play a key role in representing and structuring domain knowledge. In the biomedical domain, the need for this type of representation is crucial for structuring, coding, and retrieving data. However, available ontologies do not encompass all the relevant concepts and relationships. In this paper, we propose the framework SiMHOMer (Siamese Modela for Health Ontologies Merging), to semantically merge and integrate the most relevant ontologies in the healthcare domain, including diseases, symptoms, drugs, and adverse events. We propose to rely on the siamese neural models we developed and trained on biomedical data, BioSTransformers, to identify new relevant relations between different concepts and to create new semantic relations, the objective being to build a new consistent merging ontology that specialists could use as a new resource for various health-related use cases. To validate the new relations, we have leveraged existing relations in the UMLS Metathesaurus and the Semantic Network. To evaluate our findings, a large language model is also used. Our first results show promising improvements for future research.