SmILE Semantic Information Systems and Language Engineering Group

Well, let's start with an apology to both myself and potential readers, because I kind of don't like these shameless self-promoting shouts... However, it seems that sometimes it's not that bad idea to share our happiness and surprise with the rest of the world (when some unnamed people I trust and respect claim it's good for me and our team, for SmILE and for DERI, too, there must be a point in doing so ;) ).

So, why are we happy, what's this surprise about? On Thursday, August 7, we were told that we are semi-finalists of the Elsevier Grand Challenge after the positive evaluation of our Teaching Machines to Teach Us project proposal by the panel of judges. According to Elsevier, the challenge is aimed at general knowledge enhancement in life sciences. Its main purpose is to improve the way scientific information is communicated and used, presumably by means of research advances in computer science and bio-informatics. It happened that about 70 contestants submitted abstracts of their tentative projects addressing the challenge topic. As the challenge has been advertised at places like a Nature blog, the New Scientist publisher's blog, and the Elsevier's Artificial Intelligence or Gene journals, it soon became obvious that the competition will be very likely quite tough. In the end, there were 10 submissions out of 70 accepted for the next phase of the challenge. I think the fact that our project has survived this rather brutal cut-off explains both our happiness and surprise (a very pleasant one, though)!

In the submitted abstract, we tried to propose a concise and pragmatic plan of how to integrate my research in an empirical KR&R for emergent knowledge (see this and this preliminary technical report for details) with Tudor's and Siggi's SALT and Konnex frameworks. The application suite that will come out from the project is supposed to exploit the asserted (i.e., provided by human agents) and emergent (i.e., automatically extracted) knowledge associated with the life science publications in order to better integrate, search and retrieve their content automatically. Should you be curious about any details, feel free to go to the project web page. You can also ask me or other team members to learn more.

What's going to happen now? A bit of another embarrassing self-promotion with Elsevier, as they want to inform the emerging challenge community about the semi-finalists. Then we'll get access to a substantial amount of data in their life science journals. I've already requested as much content as possible from about 50 premier journals related to cancer research and oncology, since this is the domain I'd prefer in the scope of our research application (for some irrelevant naive, dreamery (as usual :) ) and personal reasons). And then, most importantly, a lot of tough, but intriguing work lies ahead of us. By December 1, 2008, we have to submit a paper to Elsevier, describing the progress and results of our project to date in detail. On December 15, 2008, I'll present the outcomes in Cambridge, Ma., USA to the panel of judges and Elsevier representatives. By that time, we'll try hard to deliver as good results as possible. So far, even being one of the semi-finalists is quite good and potentially relevant for future opportunities. However, it'd be even greater if we managed to be among the four finalists to be selected after the December 15 meeting - those will be the runner-ups for the overall victory in the challenge...

Until then, keep your fingers crossed and stay tuned - further updates will be posted to this blog and also to the project web page.