I am just back from ISWS 2022, and it was a great opportunity to attend it. The International Semantic Web Research Summer School took place onsite in Bertinoro, Italy, where I still haven’t got used to walking up hill and down hill all day long. It was an intense week where I learnt a lot about different things, and also met and discussed with a lot of junior and senior researchers involved in the Semantic Web Community. The directors of the school were Valentina Presutti and Harald Sack.
The main aim of the summer school was to experience the researcher’s life in all its aspects: from writing papers to presenting, going through team working, methodology, networking while also having a lot of fun.
One of the core component of the week was the Research Task Force (RTF), where we were working on a problem together. The broader topic was on Empirical Semantics, and the idea was to be paired with people you didn’t meet before but who had common interests. We were also tutored by one senior researcher, and had to come up with a paper and a presentation in a couple of days. The papers will later be compiled into a position paper that will be published on arXiv.
I was lucky to find a group easily, with researchers interested on the same topic. We were tutored by Claudia d’Amato, who I thank again for her guidance that week, and our team was composed of Agnese Chiatti from KMI, Mario Marino from the University of Bologna, and myself from the Sony CSL Paris and the Vrije Universiteit Amsterdam. We spent a very nice week together, went along well and had a great cohesion. Thank you both for this amazing week! We focused on capturing empirical semantics automatically by exploiting data evidence. More specifically, we were interested in discovering disjointness axioms from data evidence.
First things first, it was really great to have Claudia as a tutor, or even better: as a leader. She was always motivating us, guiding us not to get lost but also leaving us some freedom to explore and propose solutions, and she was always supportive. Thanks Claudia!
Most importantly, my main takeaway from this RTF is that CS is about computer of course, but also about science, and that the computer is only the tool. I knew that before of coure, but I realised that it was quite easy to forget it temporarily. We do not jump into an implementation without exactly knowing why and how we would use it for. Before that, you need to clarify and formalise the problem, have an overview of related work, design your solution. The solution you propose must be sound from the very beginning.
Lastly, I once again was reminded that research is about much more things that what I had imagined in the first place: brainstorming, proposing a solution to a problem, presenting, writing, explaining clearly, motivating your choices, and so many others.
Keynote: be inspired
I was lucky to be hired on a European Project, MUHAI, that aims to push current AI towards more meaning and understand. And it happens that Luc Steels and Frank van Harmelen are also involved in that project, so I had already had the opportunity to discuss with them and hear some of their presentations. I liked their keynote presentations because I see them as regular, good memory pills: from Luc, to know more about the generic history of AI and the different types of intelligence, and from Frank, to understand more how symbolic and sub-symbolic AI can be combined to build better AI systems.
Great keynote as well from Elena Simperl, on current research of studying KGs from the prism of socio-technical constructs. The focus was about understanding what these large KGs say about us humans. More specifically: how can analysing KG engineering empirically help understand and design best practices for building better KGs?
Tutorial: learn and understand - how
I learnt Semantic Web on the go when I started my Ph.D., since my background was more in Mathematics and Computer Science - Machine Learning and Natural Language Processing. I was glad to have a speed recap by Sebastian Rudolph and Axel Polleres on KR Querying & Reasoning for Large Knowledge Graphs, and that all the core notions from Semantic Web were now much clearer to me than a year ago. I also found very interesting the challenges underlined by Axel on the SPARQL language query, especially because I have already encountered them in my research: contextualised data, paths in the graph, scalability, querying and reasoning, sustainability.