Disha Purchit from the CAIMed group led by Prof. Dr. Maria-Esther Vidal presented her research on an innovative neuro-symbolic system for medical knowledge discovery at the recently held WSDM Conference. The WSDM Conference, as one of the leading venues for experts in the field of web search and data management, provided the ideal platform to showcase the interdisciplinary insights and technological advancements from the research group to an international audience.

As part of her presentation, Purchit demonstrated how the novel approach—by combining symbolic learning from medical ontologies with Knowledge Graph Embeddings (KGE)—helps address the challenges of incomplete medical knowledge graphs. It was particularly emphasized how this method preserves semantic consistency while enabling more accurate predictions. The experiments presented at the conference, which focused on applications in lung cancer care, showed that the approach not only outperformed existing models but also opened up new perspectives for personalized medicine.

The WSDM Conference not only provided an opportunity to share groundbreaking results but also to promote scientific exchange with leading experts from around the world. The conference took place until March 14, 2025, at the Hannover Congress Centrum (HCC) and was chaired by Prof. Dr. Wolfgang Nejdl and Prof. Dr. Sören Auer.