A Platform for Semantic Web Studies
to be presented at Web Science 2010
At the very beginning of Watson, there was this idea that not only it should be a gateway to the Semantic Web by providing services for applications to exploit it, but also because it would represent a new platform for us, researchers, to better understand it. Indeed, Watson creates a huge collection of ontologies and indexes it so that processing its content can be done in an efficient way. Therefore, back in 2007, we published a paper at the EON workshop entitled characterizing knowledge on the semantic web with Watson, which presented interesting findings, in particular on the concrete use of semantic technologies at the time.
Following this, I still very often present Watson as two things: the well-known platform for developers allowing them to build next generation Semantic Web applications, and a platform for researchers allowing them to study the Semantic Web. However, in general, not a lot has happened since the EON paper in 2007 concerning this second aspect... in appearance.
In reality, in particular as part of Carlo's PhD, a lot has happened on monitoring and analyzing one very important aspect of ontologies online: their relationships. So here is a whole blog post just to say that we finally decided to write an overview of this work on using Watson as a platform for Semantic Web studies, in a paper for the Web Science 2010 conference. The idea here is to show how Watson can be used as a foundation for such a platform, and to describe some of the studies we have been conducting using it, to answer questions such as "How do ontologies evolve?", "How does a modular ontology typically look like?" and many others.
So, if you happen to be in Raleigh, North Carolina on the 26-27 April 2010 and that's the kind of questions you have been asking yourself, come and talk to me, I might know a way to find the answers.
ORES - A Workshop on Ontology Repositories
I have just sent around the first CFP for a workshop collocated with ESWC 2010 (and so located in a very nice part of the world...) about ontology repositories and editors for the Semantic Web (see the ORES 2010 website). There are so many interesting issues to be discussed in this area that I am surprised it is the first of the sort.
One of the most important of these issues concerns the interoperability and federation of ontology repositories, reflecting the many discussions we already had with people from the Open Ontology Repository initiative, the NCBO BioPortal, and many others. It is nice to see all these people getting involved in this event.
Another aspect quite important to me concerns the role of ontology repositories in the ontology lifecycle, including integrations with ontology editors (c.f. the Watson plugin), ontology modularization, ontology versioning (c.f. IWOD), ontology metadata, collaboration, etc.
In sum, this promises to be a quite exciting event with many connections to broader projects. If you want to be part of it, here are the dates to put in your calendar: Deadline for paper submissions - 1st March 2010, workshop - 30th or 31st May 2010.
A couple of papers accepted recently
I didn't have much time lately to update this blog with the latest news, but we recently had a few submissions accepted that are related to Watson and are worth mentioning here already.
At K-CAP 2009, on agreement and disagreement in ontologies: This is one I really like, mainly because I wrote it, but also because this is something I wanted to work on for quite some times, and I think the result is really interesting. Basically, the paper describes measures of agreement and disagreement of an ontology with a statement (e.g. Person subClassOf Human). The measures provide more than just a yes/no test like a coherence check would do, but provide granular measures of both agreement and disagreement. No point going into the details here (read the paper!), but the part I really find nice is that is it leads to the possibility to measure consensus and controversy on statements in a collection of ontologies (and I am talking about Watson here!) Experiments detailed in the paper show really interesting possibilities with this. Also, it is fascinating to look at agreements and disagreements between ontologies in a particular domain, to see that different "camps" exist, with groups of agreeing ontologies commonly disagreeing with other groups of ontologies.
We also, got a demo of Cupboard accepted at K-CAP 09: submitted with the guys from INRIA and Karlsruhe. Cupboard is going really well with more and more users joining. If you are around at K-CAP, come talk to us at the demo session!
A paper on the ontology of relations between ontologies at KEOD 2009: As part of Carlo's PhD, he needed an ontology to formalize and reason upon relations between ontologies (inclusion, version, compatibility, etc.) This paper describes the design of this ontology (called DOOR, descriptive ontology of ontology relations). This is actually a quite complicated design, but also a quite exciting outcome as the ontology is going to be at the basis of Carlo's system (Kannel) for detecting and managing relations between ontologies in large ontology repositories (talking about Watson again here).