KNoWS at ISWC 2025
Our Knowledge on Web-Scale team is at the
International Semantic Web Conference 2025 this week!
We have papers on link prioritization, time functions in SPARQL, low-code ODRL access management,
a new concept we call Trustflows, and queryable verifiable credentials;
insights on client-side query federation; and tutorials on RDF & SPARQL 1.2 and RDF-Connect pipelines.
Our contributions
Ruben Eschauzier and Ruben Taelman will present a maintrack paper on
"Revisiting Link Prioritization for Efficient Traversal in Structured Decentralized Environments".
This paper evaluates link prioritization for Link Traversal Query Processing in structured environments
like #projectSolid.
It defines and extends the R3 metric, reimplements algorithms modularly in Comunica, and
finds that existing methods offer no performance gain over FIFO traversal.
Ruben Taelman also takes the lead in
the tutorial on RDF and SPARQL 1.2.
This tutorial gives a hands-on introduction to RDF 1.2 and SPARQL 1.2,
updating the core Semantic Web standards after a decade.
It explains motivations, new triple-term and reification features, and key specification changes,
enabling attendees to apply RDF-star concepts in future work.
Ruben Taelman’s PhD student Jonni Hanski will also present a paper in the wikidata workshop on
"Observations on automated client-side query federation over Wikidata SPARQL endpoints".
It explores automated client-side query federation across
the two new SPARQL endpoints after the graph split.
It evaluates source assignment strategies,
showing that automated federation can replace manual SERVICE clauses for some queries,
though scalability challenges remain.
Julián Rojas will be presenting
our poster on Trustflows,
a new operational model for the read–write Web of Data.
By combining CQRS and Event Sourcing with Linked Data, it decouples reads and writes,
supports evolving data models, and adds explicit trust contexts,
validated in the PACSOI health data use case.
Julián, together with Arthur Vercruysse and Ieben Smessaert,
ran the RDF-Connect tutorial.
This tutorial introduces RDF-Connect, a framework for provenance-aware,
multilingual streaming data pipelines.
Through hands-on exercises, participants build interoperable workflows across languages,
track provenance with RDF/PROV-O, and create a weather forecast knowledge graph for Nara, Japan.
You can do the tutorial as well at home.
The same Ieben and Julián also have a contribution with
a demo paper on time functions in SPARQL.
This paper proposes a set of SPARQL extension time functions to treat partial
(e.g. year, month) or floating (no timezone) temporal literals as intervals,
enabling meaningful cross-datatype comparisons, filtering, and sorting.
Implemented in Comunica, the approach helps address limitations in real-world knowledge graphs and
suggests improvements for future standards.
Wout Slabbinck and Beatriz Esteves also have
a demo paper on LOAMA,
a low-code application for enforcing access via ODRL policies.
It allows non-technical users to specify usage rules visually, integrates with semantic policies, and
supports enforcement in data environments.
Last but not least, Gertjan De Mulder has
a poster paper presenting a method to make Verifiable Credentials (VCs) queryable via SPARQL in OID4VP authorization flows.
Credentials are represented in RDF and stored in named graphs;
SPARQL queries select claims to be disclosed.
Those claims are then transformed back into JSONPath pointers to build
a selective Verifiable Presentations that is privacy-preserving and semantically enriched.