Smartlogic Chops at the Gordian Knot of the Semantic Web

September 5, 2015

This semantic Web thing just won’t take a nap. The cheerleaders for the Big Data and analytics revolutions are probably as annoyed as I am. Let’s face it. Semantic was a good buzzword years ago. The problem remains that anything to do with indexing, taxonomies, ontologies, and linguistics lacks sizzle.

If you want analytics, you definitely want predictive analytics. (I agree.Who wants those tired Statistics 101 methods when Kolmogorov-Arnold methods are available. Not me, that’s one of my relatives. I am the dumb Arnold.)

If you want data, you want Big Data. The notion of having large volumes of zeros and ones to process in real time is more exciting than extracting a subset which meets requirements for validity and then doing historical analyses. The real time thing is where it is at.

I read “The Promise of the Semantic Web, Truth of Fiction?” hoping for an epiphany. Failing that high water mark of intellectual insight, I would have been satisfied with a fresh spin on an old idea. No joy.

I read:

Semantic technologies have the capacity to extract meaning from unstructured information found within an enterprise and make them available for processing. Our new Semaphore 4 platform combines the power of semantic technologies with our ontology management, auto classification, and semantic enhancement server to help organizations identify, classify and tag their content in order to use the intelligence within it to manage their business.

The information strikes me as a bit of the old rah rah for a specific product. The system is proprietary. The licensee must perform some work to allow the “platform” to deliver optimum outputs.

What about the answer to the question of a promise as truth or fiction?

The answer is to license a proprietary product. I am okay with that, but when the title of the write up purports to tackle an issue of substance and deflects substantive analysis with a sales pitch, I realize that I am out of step with the modern methods.

Here’s my take on the question about the semantic Web.

Folks, the semantic Web thing is a reality. A number of outfits have been employing semantic methods for years. The semantics, however, are plumbing and out of site. The companies pitching RDF, Owl, and other conventions are following a wave which built, formed, and crashed on the shore years ago.

At this time, next generation information access vendors incorporate linguistic and semantic methods in their plumbing. The particular pipe and joint are not elevated to be the solution. The subsystems and their components are well understood, readily available methods.

As a result, one gets semantics with systems from Diffeo, Recorded Future, and other innovators.

The danger with asking a tough question and then answering it with somewhat stale information is that someone may come along and say, “There are vendors who are advancing the state of the art with innovative solutions.”

That is the main reason that MIT and Google have funded these NGIA (next generation information access) outfits. Innovation is more than asking a question, not answering it, and delivering a sales pitch for a component. Not too useful to me, gentle reader.

I would suggest that the Gordian knot is in the mind of the semantic solution marketer, not the mind of a prospect with a real time content problem for which modern technology enables effective solutions.

Stephen E Arnold, September 5, 2015

Comments

Comments are closed.

  • Archives

  • Recent Posts

  • Meta