Structural Summarization of Semantic Graphs

RDF graphs comprise highly complex data, both from a structural and from a semantic perspective. This makes them hard to discover and learn, and hinders their usability. An elegant basis for summarizing graphs is provided by the graph quotient formalism. In a nutshell, a graph quotient specifies a way to view some graph nodes as equivalent to each other, and represents a graph through its equivalence classes based on this equivalence. I will present work carried in my last team over the last few years, on quotient summarization of semantic-rich RDF graph. In particular, I will introduce a set of summaries particularly suited for the heterogeneous structure of RDF graphs, and discuss novel results at the interplay of summarization and saturation with RDF Schema rules