I stepped out from the session bc of a cough from last week that doesn't want to go away, and because my batteries were running out and I have a demo to give later this afternoon
I am sitting in the biohackathon area, with a RDF4RiskAssessment poster from four people from the German BfR
@bfr, is Taras Guenther or one of the other authors also on the #fediverse
we're back from lunch (no signs, so not clear what I had), and now started the first session after lunch by Samaneh Jozashoori and Nirmal Raman Kannaiyan using KGs and LLMs on literature for the purpose of drug repurposing
the second talk in the session also shows that RDF in hospitals is there. Enough that people are now focusing on querying across databases, and no longer just aligning ontologies.
Also, it shows that writing efficient SPARQL queries is something that requires tweaking SPARQL queries. The order of instructions matter, but a SPARQL query does not have a clear linear direction
second session just started with the first talk by Sabine Österle and Jan Armida about sharing health/patient data in Switzerland, across 26 kantons and legislations and 4 national languages
third speaker is Pedro Paulo Favato Barcelos, of Health-RI about ontologies in patient history provenance
he quotes: "the opposite of an ontology is not a non-ontology, but a bad ontology" -- someone
second speaker is Ömer Durukan Kılıç from Maastricht University. He is talking about MIMIC-III, which I think is described here: https://doi.org/10.1016/j.ijmedinf.2026.106297
first speaker is Ronald Cornet (https://qlever.scholia.wiki/orcid/0000-0002-1704-5980) talking about the history of ontologies, reasoning, and more
"we have been talking about reasoning over patient data for more than 20 years, but we still have not solve it"
Egon Willighagen
Jerven
