A new thesis proposes a modular framework for reasoning over uncertainty in knowledge graphs, addressing issues like imprecise data and incomplete schemas. This approach aims to enhance semantic data integration while maintaining computational efficiency through specialized reasoning techniques.