ICTP Workshop 2007/BlogNotes on Semantic Web

From Bjoern Hassler's website
Jump to: navigation, search


{{#titlehere:}} Please note that the present pages on the ICTP workshop were obtained from a mediawiki used during the workshop. There may be a number of links that go to missing pages, and other inconsistencies. Much of it is also very informal, and should be seen as workshop notes!From: User:Mathe/BlogEntry: 2007 December 07 08:13:49 UTC

Marco is back on the presentation floor. He is continuing his presentation by tackling the issue of how one can find their recorded material with ease. In essence, he is trying to provide a solution of finding the needle in the hay. Incidentally the application they use for this task is called 'needle'.

The needle profiles the videos and associates them with slides and audio from the recording. To achieve this, an audio extraction has to be performed among other things and a voice recognition system is required.When needle has done what it must do to aid a user in finding that which they need, a user can just type in their search. All the hits will be returned with roughly the following columns: description, video clip, slides, audio track.


Is this system any good? The statistics show an above 60% success but this is easily explained by the dependence on speech recognition to perform the profiling. Ideally if each teacher before the recording were subjected to go through the voice recognition system for tuning then the results would improve.But this has time implications and as we know time is money!


The whole principle of how the above system is built takes us to an interesting concept known as the Semantic web i.e syntactic web. A quick intro on semantic web can be found at http://en.wikipedia.org/wiki/Semantic_Web. A lot of this is of course being explained by Marco so after an hour or so, the lecture can be reviewed at http://sdu.ictp.it/eya/richmedia07.php.

First step for understanding the semantic web requires understanding of XML. The next step requires understanding a language called resource description framework (RDF). RDF uses XML. to better understand RDF creating a flow chart be helpful since the actual syntax is not necessarily more readable to us as it is meant to be understood by the machines. The flowchart can be understood better by viewing it in terms of classes but for all the mathematicians present you are urged to view the flowchart in terms of sets and relations.

The third step is to understand how RDF relates to ontology. In a line ontology can be defined as a specification of a conceptualization. Ontology can be explained from 3 disciplines: philosophy, linguistics and computer science. In computer science this explained as a "formal, explicit specification of shared conceptualisation". Structure of an ontology has two distinct parts: name and properties. In the simple case ontology can be explained as a taxanomy.

Important note: logic is important to ontology. Thus common sense must be really common to the structure of ontology.


How can we include RDF to html pages? This is the basic problem faced by ontologists. Without an efficient solution the semantic web will not exist in the magnitude that we would want. In the afternoon we will use these concepts for our hands on experience. Can't wait!