Taxonomy Completion

Taxonomy completion focuses on automatically enriching existing hierarchical knowledge structures by adding new concepts in the correct positions. Current research emphasizes methods that leverage pre-trained language models and geometric embeddings to effectively capture semantic and structural relationships between concepts, addressing challenges like asymmetric relations and the efficient handling of both insertion and attachment operations. These advancements improve the accuracy and efficiency of taxonomy building across diverse domains, impacting knowledge representation, information retrieval, and other applications relying on well-organized knowledge bases.