A partial Top-Down approach to the Semantic Web

New, Improved *Semantic* Web!Image by dullhunk via Flickr

Here is a cool, down-to-earth Semantic web post, in another blog (Francisco Antonio Cerón García’s «Meta-internet» blog), with lots of useful information, delicious for… info-gluttonous Semantic geeks (like myself-hehe):

Top-Down: A New Approach to the Semantic Web

March 13, 2008 by identityandconsulting

(This cool blog is now in my «Web 3.0 blog-roll» – of course)

Now, as regards Mr. Garcia’s core-question:

«But what if it is not even necessary to build the first generation of semantic tools? What if instead of trying to teach computers natural language, we hard-wired into computers the concepts of everyday things like books, music, movies, restaurants, stocks and even people. Would that help us be more productive and find things faster?»

The answer is -most probably- that if we did «hard-wire into computers the concepts of everyday things« (etc), then we’d effectively end up with a (meta-)Ontology-of-everything, akin to the famous CYC ontology (now open-source, as «OpenCYC») , with an incredible number of semantic and logical definitions already built-into it.

Moreover, if we don’t really need such a huge amount of semantic and logical information, all we need to do is extract what we need from OpenCYC, Wordnet, and many other sources, adding our own authoring, to arrive at a problem-specific ontology.

  • In any case, a universal global ontology-of-everything has been proved inadequate;
  • which is why the CYC project became open-source, at some point, abandoning most of its original (business-) ambitions.

Nevertheless, an «ontology-of-everything» may be feasible, iff it is also able to learn by itself, through its own inference engine and NLP module. In fact, there is (at least) one excellent researcher who tries to do precisely this:

  • The creator of the «Texai.org Blog« is an ex-employee of CYC Corporation, now a retired (but hard-working) freelancer, devoted whole-heartedly to his own project, which is open-source:
  • The creation of a «truly intelligent» Automatic Learning System.

So, Mr. Garcia’s proposal doesn’t seem to be wrong or unfeasible; on the contrary, it’s a logical step to take.

Moreover, you don’t really have to use a huge-size Semantic Knowledge Base, for a good top-down project like this; all you need to do is identify the problem-specific ontologies and the semantic information that is relevant to the problem, as well as develop some special tools, i.e. natural-language (NLP) understanding tools with a limited scope and limited parsing ability.

  • At THAT stage, you can process all those dumb web-pages (without explicit Semantic content) at the bottom-end, using the wisdom of your top-end application(s).

Besides, what Mr. Garcia says is not very different from what certain projects based on NLP are already doing (e.g. POWERSET).

As for myself, I collaborate with a company who need semantic applications A.S.A.P. that can chew-up existing web-pages; pages without RDF or other bottom-up semantic stuff, so I understand what Mr. Garcia is saying: -This is what we already do, given the lack of something better, as regards the web-pages we’re interested in. It’s a bit one-sided, of course, since the web-pages themselves are not semantic; nor likely to become semantic through their own authors’ initiative (in the near future). However, all this can be combined with Semantic information elsewhere, which is freely available, as well as our own Semantic Knowledge-base (with completely different items and topics than those particular web-pages analysed semantically).

In any case, after more than two decades in NLP and Prolog I feel that the ultimate best solution will be the (future) use of inferences and NLP at all levels, both top-down and bottom-up. (It will take time, but… in the meantime we can already achieve a lot, with what we already have).

Related articles

Other posts in this blog about the Semantic Web:


Εισάγετε τα παρακάτω στοιχεία ή επιλέξτε ένα εικονίδιο για να συνδεθείτε:

Λογότυπο WordPress.com

Σχολιάζετε χρησιμοποιώντας τον λογαριασμό WordPress.com. Αποσύνδεση /  Αλλαγή )

Φωτογραφία Google

Σχολιάζετε χρησιμοποιώντας τον λογαριασμό Google. Αποσύνδεση /  Αλλαγή )

Φωτογραφία Twitter

Σχολιάζετε χρησιμοποιώντας τον λογαριασμό Twitter. Αποσύνδεση /  Αλλαγή )

Φωτογραφία Facebook

Σχολιάζετε χρησιμοποιώντας τον λογαριασμό Facebook. Αποσύνδεση /  Αλλαγή )

Σύνδεση με %s