Connecting the Right Dots: Entity Resolution on Knowledge Graphs
Knowledge graphs (KGs) are as popular as ever, serving as the backbone of a variety of tasks from question answering, over recommendation systems to biomedical uses cases such as drug repurposing. The ability of KGs to store data in a semantically rich way, while still maintaining flexibility is alluring for practical application, but poses a variety of challenges in the data integration step. The integration of knowledge graphs calls for special tools, that go beyond the tried and tested approaches from the world of relational databases. This talk will introduce the dedicated challenges for entity resolution systems when dealing with knowledge graphs. Various facets of this research field are presented ranging from holistic data integration, scalability aspects all the way to modern knowledge graph embedding methods and incremental solutions.
Daniel Obraczka is a PhD candidate at ScaDS.AI investigating entity resolution on knowledge graphs with a specific focus on the use of knowledge graph embeddings to aid in this endeavour. Previous to pursuing his PhD he obtained a B.A. in Sociology, as well as a B.Sc. and M.Sc. in Computer Science.