Unifying Knowledge: Entity Linking in Scholarly Knowledge Graphs

The poster presents a study of challenges with linking and disambiguating author entities in Scholarly knowledge graphs. Author entities are the central part of scholarly data. The varied representation of author names due to different cultural norms and disambiguation challenges due to common names, makes it a challenging and important problem. We also compare the representation of author labels between two scholarly knowledge graphs DBLP and SemOpenAlex knowledge graphs, to highlight inconsistencies. Entity linking is an important challenge to solve to make knowledge graphs accessible through Knowledge Graph Question Answering. We used of DBLP-QuAD and Scholarly-QALD KGQA datasets for this study.