Markus Krötzsch ISWC 2022 keynote
A brilliant keynote by Markus Krötzsch for this year's ISWC.
"The era of standard semantics has ended"
Yes, yes! 100%! That idea was in the air for a long time, but Markus really captured it in clear and precise language.
This talk is a great birthday present for Wikidata's ten year anniversary tomorrow. The Wikidata community had over the last years defined numerous little pockets of semantics for various use cases, shared SPARQL queries to capture some of those, identified constraints and reasoning patterns and shared those. And Wikidata connecting to thousands of external knowledge bases and authorities, each with their own constraints - only feasible since we can, in a much more fine grained way, use the semantics we need for a given context. The same's true for the billions of Schema.org triples out there, and how they can be brought together.
The middle part of the talk goes into theory, but make sure to listen to the passionate summary at 59:40, where he emphasises shared understanding, that knowledge is human, and the importance of community.
"Why have people ever started to share ontologies? What made people collaborate in this way?" Because knowledge is human. Because knowledge is often more valuable when it is shared. The data available on the Web of linked data, including Wikidata, Data Commons, Schema.org, can be used in many, many ways. It provides a common foundation of knowledge that enables many things. We are far away from using it to its potential.
A remark on triples, because I am still thinking too much about them: yes to Markus's comments: "The world is not triples, but we make it triples. We break down the world into triples, but we don't know how to rebuild it. What people model should follow the technical format is wrong, it should be the other way around" (rough quotes)
At 1:17:56, Markus calls back our discussions of the Wikidata data model in 2012. I remember how he was strongly advocating for more standard semantics (as he says), and I was pushing for more flexible knowledge representations. It's great to see the synthesis in this talk.
| Previous entry: Karl-Heinz Witzko | Next entry: Ten years of Wikidata |
