geometric-embeddings

This thesis proposes a modular framework for scalable uncertainty reasoning in knowledge graphs, addressing imprecise attribute values, probabilistic triple existence, and incomplete schema through tailored algebraic, logical, and geometric techniques.