ISWC 2014, some impressions

I have just arrived from Riva del Garda, Italy, where I participated in the 13rd International Semantic Web Conference (ISWC’14). As we typically do in our group (OEG) I will write a short report of my impressions about what happened there, although through this blog for the first time. Instead of writing a long boring report I will do it in the form of short “snippets”…

Riva del Garda

OVERALL. This was a very good and high level conference as usual! It was a pleasure to meet such many old colleagues and friends (from KMi, INRIA, UNIZAR, previous projects and conferences, etc.).

opening presentation

KEYOWORDS. “Data” was the dominant keyword in all tracks (in-use, industrial, etc.) but in the research one (where “semantic”, “ontology”, and “sparql” dominated). Is this indicative of a lack of alignment between theory and practise in the area? (I´m asking to myself).

SOME AWARDS. The Protégé OWL Plugin, which had an unquestionable impact during the last decade, won the 10-years SWSA award. There were also awards for the best paper reviewers! A very cool idea to stimulate such a (frequently) difficult and time consuming task.

KEYNOTES. The keynotes were good, although maybe not as exceptional as the ones I remembered from my last ISWC (2011 in Bonn). I particularly enjoyed the one by Nigel Shadbolt (ODI), which illustrated many cool uses and visualizations of open data, and the one by Prabhakar Raghavan (Google) that was more a master class on information retrieval than a usual keynote.  Further, Yolanda Gil cited my colleagues Daniel Garijo and Oscar Corcho (from OEG) in her keynote! 🙂

NLIWoD. I attended the workshop on Natural Language Interfaces for the Web of Data. I really liked the invited talk by Andre Freitas (Insight), about combining distributional semantics + structured data to solve schema-agnostic queries, which I think is the right approach to follow.

DEVELOPERS. The Developers track was an innovation this year. I attended part of the sessions and discovered some new refreshing ideas that I liked a lot. For instance the proposal for hosting linked data for free on GitHub, Google Code and similar services. I also knew about Linked Data fragments and triple pattern fragments for the first time.

Tutorial on Building the Multilingual Semantic Web

LIDER’s TUTORIAL. The tutorial we organised on Building the Multilingual Semantic Web went pretty well. It was not a crowded one (we counted 31 people attending “actively”) but the participants followed it with interest and gave useful feedback as well.

MINUTE MADNESS. This time, the number of posters was really high, but the minute madness session (a series of quick 35 sec. introduction to their work by the poster’s authors) gave me the chance of being more selective and making a better use of my time.

POSTERS. The poster session itself was a log one but it was worth spending my time on it. I talked to a lot of people and knew a lot of new stuff (I am adding a few pointers at the bottom of this post). For the first time, I took pictures of the posters that I found more interesting (for my records), which turned out to be a good idea, given my bad memory for names!

DEMOS. The well-deserved best demo award was for a demonstrator of the triple pattern fragments. They encapsulated the whole DBpedia in a Raspberry-Pi and run relatively fast queries on it. I wonder how much expressive this “triple fragments” stuff is but the approach could help in the adoption of SW techniques, specially for those data providers that cannot deploy/maintain highly available SPARQL endpoints.

RESEARCH TRACK. The research track was also very interesting (many different stuff to chose among as usual). Although this track would deserve a separate post, let me highlight AGDISTIS, a system for named entity disambiguation developed at University of Leipzig, which turned out to be winner of the best paper award.

SW CHALLENGE. From the “Semantic Web challenge” my favourite presentation was the one about the Graph of Things (second prize in the challenge), for their cool visualizations of a very large network of dynamic + static data.

LINKS. Finally, a few pointers to some cool stuff I want to explore a little more…