Yanzeng Li's paper "VGStore: A Multimodal Extension to SPARQL for Querying RDF Scene Graph" has been accepted by ISWC 2022.
This paper aims to solve the problem that traditional RDF systems and SPARQL engines cannot understand and support multi-modal (such as visual) semantic queries, and the limitation of semantic parsing-based multi-modal KBQA application. 
This paper proposes a SPARQL extension based on virtual predicates. To verify the method, this paper proposes to organize the structured scene graph dataset Visual Genome in the form of RDF, and the multi-modal question answering is realized through the proposed SPARQL extension.