RDF slicing

[keme686]

The current slicing approaches use simple SPARQL graph patterns to select triples that match the query pattern. They dont have expressive power to include depth of the graph structure . For example:
CONSTRUCT WHERE {
?s rdf:type foaf:Person.
?s ?p ?o.
}
This kind of queries only extracts one level of the path, i.e., every node connected to the person. For some applications, e.g., question answering, this might not be enough. The sould be more connected graph around the resources to get the result. For example:

Question: “Which premier league team does David Beckham played?”

For this kind of questions that needs to go second level of the graph to get the team from premier league.