Egon Willighagen (@egonw.social.edu.nl.ap.brid.gy)

SWAT4HCLS 2026 A bit over a week ago, SWAT4HCLS 2026 took place, with the matching biohackathon on Thursday (see this post. I attempted a bit of live coverage on mastodon: day 1 and day 2. But it seems the semantic web community interested in SWAT4HCLS has not found the fediverse yet. So, make sure to check this full list of abstracts. The meeting consisted of four keynotes, each one was quite interesting. Cornet gave a nice historic perspective of the venue and of the semantic web field, which is a great way to welcome the participants to your institute. The talk also touches on the main theme of the meeting: clinical data. It is a long standing (and important) research field, but progress is slow. Cornet comments along the lines that _we have been talking about reasoning over patient data for more than twenty years, but we still have not solve it_. The problem is really not only privacy, but simple also lack of a common language. As Sabine Österle explains about sharing health/patient data in Switzerland, across 26 kantons and legislations and 4 national languages. Another issue is more technical, running SPARQL across hospitals involves more than just aligning ontologies, but also requires (too much) fiddling with SPARQL queries. There was plenty of other content too, however. For example, I was pleasantly surprised by the RDF4RiskAssessment work, the RO-Crates for BioImaging, and FDPcrawleR. All these projects have direct links to research ongoing in our TGX team. Hanna Bast gave the second keynote of the first day, about QLever (doi:10.1145/3132847.3132921). She talked about some of the recent improvements, something we really needed for Scholia. She showed a technical approach to make federated queries faster, tho it currently only works between endpoints that both run QLever. One thing I am looking forward to, is playing with the notion of materialized views, but the biohackathon was too short to get around to that during the Thursday. The second day kicked off with a keynote by Janna Hastings, whose work I greatly admire. I was not disappointed today, and she showed the Behaviour Change Intervention Ontology and Chebifier (doi:10.1039/D3DD00238A). The last talk I want to mention in the blog is by two researcher working with Michel Dumontier. They presented a study about deduplication in/of knowledge graphs. This is something I want to read in more detail. https://github.com/egonw/blog2/commits/new/_posts/2026-04-04-swat4hcls-2026.markdown