We are occassionally asked “What is so special about illumin8?” Here’s an example of some of the magic operating behind the scenes of a typical illumin8 search. Suppose a researcher was investigating the problem of soil erosion. Entering the simple phrase soil erosion in our search box generates a powerful result set with summarized, organized information on relevant products, companies, organizations, authors, and approaches. Quite a bit can be gleaned from just this initial summary overview. And selecting one of the solutions, such as “grass”, lets us see a set of records discussing the use of grass to solve soil erosion. Seems simple, but there is more to this than meets the eye.

First, the user has only entered the phrase soil erosion in the search box (the word “grass” hasn’t been entered). Second, illumin8 automatically understands that soil erosion is a problem to be solved. It searches its database to find all instances where the phrase occurs in the correct problem context. It then finds entities connected with the problem, for example companies or organizations investigating solutions to soil erosion, approaches to solving soil erosion, etc. Finally it combines the results and places them in the proper category.

Let’s look at a record included in the grass solution.

This is a pretty straightforward semantic match with the query. What is even more interesting though, is that the system automatically finds synonyms to “grass” and fuses the records containing them under the “grass” solution. For example, records containing words like pasture, forage, and forage plants are all automatically discovered and included as part of the “grass” solution set. Here is an example:

I hope you can see the power of this technology. Typical search products would not pinpoint the search term in a specific, meaningful role, nor would they extract entities which connect to the term in an intelligent way. But that is what illumin8 does. And it goes further by summarizing, adding synonyms, etc., The benefit is a richer search which finds the right answers in the least amount of time.

I’m a little behind with this post, but hopefully it will be worth it. More than 1,000 participants from around the globe converged on the San Jose Fairmont toting laptops, notebooks and, well, ontologies.Too bad I couldn’t get Occam’s razor through security (bad joke I know). Here are some quick highlights from what was a really interesting and even, at times, exciting show:

• The big buzz was around a “social interest” RDF enabled site, backed by Paul Allen, Vulcan Ventures, and currently in beta-invite only mode. It is called “twine”; the other milder buzz was around PowerSet (natural language search over Freebase and Wikipedia), which just launched.

• I could find no one offering a commercial product based on natural language processing (NLP) at internet scale. Nor could I find anyone offering semantic search with integrated premium scientific content. So illumin8 is currently unique in the industry, and our ability to apply our NLP engine over this vast content remains a very compelling differentiator.

• Google played the role of spoiler in the semantic party. It is skeptical of the semantic web, and very skeptical about the potential for semantic search on a global consumer scale. Yahoo, on the other hand, is a strong believer and they were pushing their Search Monkey platform very hard.

All in all, a very worthwhile show; I hope to be back next year.

to a browser near you, something that will reveal insights and solutions from billions of documents