Making unexpected acquaintances - you need a physical event for that to happen

June 01, 2021 by Arjen Santema

SEMANTiCS: You are already several years one of the Dutch thought leaders in knowledge systems and semantic web technologies and a regular participant in Semantics. How do you look forward to Semantics 2021, which is going to be a hybrid conference with both physical and virtual attention?

Jan Voskuil: SEMANTiCS has always been a great conference with different communities behind it, both from academia as well as from the world of practitioners. It is about open-mindedness, serendipity, and of course about exchanging ideas and experiences across communities. Sharing knowledge through an online meeting works fine, but meeting people on the spot, making unexpected acquaintances, looking each other in the eye: you need a physical event for that to happen. This is why I am looking forward so much to SEMANTiCS 2021. For those who cannot travel, for whatever reasons, it is good to have a plan B. But the idea of meeting people in the flesh is exciting!

SEMANTiCS: You are track leader of the local topic ‘asset management’. Can you tell something more about this topic and why it is so ‘hot’ today? 

Jan Voskuil: Assets are complex things. There are many assets, and many different parties supply information about each of these. Asset owners must always have an up-to-date and complete view. This makes managing asset information about, say, a bridge as difficult as constructing one. There is wide consensus that RDF has an enormous potential in this domain, and this consensus has been there for ages. In the Netherlands, we are a little bit ahead in this regard, I guess. Here, the consensus has led to several standardization initiatives. The number of projects that apply RDF was small but significant, and because of these standards, that number is growing fast now. The promise of RDF is better to exchange of information and, secondly, better knowledge management. Not a little bit better, but disruptively better. To make that happen, though, there are of course hurdles to overcome. So, the track will focus on these aspects: what are these standards and where are they heading, what have we learned implementing them, were the expected benefits realized, how do we set up useful object type libraries, how do we put them to use, what are the main challenges in practice? In a sponsored talk in the main conference room, I will try to present a synthesis.

Want to learn more about Semantic Technologies in Dutch industry? Register for SEMANTiCS 2021!

SEMANTiCS: Are there other specific topics that have your special interest?

Jan Voskuil: I basically started doing semantic web technologies back in 2012, so I have seen a broad range of real-world projects in many different domains. Apart from asset management and related domains, my current focus is on government and publication metadata. I think that where those two domains meet, interesting things are bound to happen. That may take some time though, as both domains each deal with a lot of complex issues.

SEMANTiCS: One last personal question. Your LinkedIn profile says you are in origin a linguist. What does someone with a background in a rich domain such as linguistics make feel at home in the highly formalized world of computer sciences? 

Jan Voskuil: The two worlds have more in common than you think. My PhD-research was about the way verbs are classified, not by people, but indirectly, by linguistic phenomena like syntax and morphology. I was able to show that similar taxonomies arise independently in wildly different languages, including Indonesian and Germanic ones. To understand such patterns, one must look at the intersection of language, human cognition, and formal logic. Now, when we exchange information using computers, this involves two ingredients: data, and agreement on how to interpret these data. These agreements are formalized in models. It was Giancarlo Guizzardi who in 2005 wrote this very important book on such models. His observation was that most models used in computer-based information exchange are rife with logical inconsistencies. His method to create models that are solid, and sound, is now a central tenet in the FAIR data movement. It says that to create great models, you must look at, well, that very same intersection where language as used in a domain meets cognition and logic. As an ontologist, I am busy with at least some of the same things as the ones I was looking at during my time at the university. I am quite happy that things turned out this way.