Subscribe to our Feed

Thought Leaders in Big Data: Atanas Kiryakov, CEO of OntoText (Part 5)

Posted on Saturday, Aug 23rd 2014

Atanas Kiryakov: The purpose of maintaining an online process is to be able to classify and enrich content with more metadata. A typical media website contains 10 or 20 categories like specific regions in the world, which means there would typically be 10 to 30 web pages that are pretty much maintained by them. Their editors make decisions on which articles, materials, and pictures get promoted and presented on each of these web pages.

What BBC did back in 2010 with the help of our technology was make a website where they have 800 different pages. They had one page for each team, each group of teams, each player, each coach, and stadium. Basically, all sorts of objects that had something to do with the World Cup were maintained automatically. There is no editor involved in the maintenance of these pages. The quality of the metadata and the granularity of classification of all the content based on this rich metadata is so good that it allows BBC to generate automatically pages that are relevant and insightful on the specific subject.

From the time the article hits the CRM system of BBC, there should be delay of no more than a minute to the moment when it starts showing up on their website. They can’t afford a slow classification process. They can’t afford an expensive classification process. They can’t afford classic classification, because when they start showing irrelevant content on their website, they start losing audience. This is a very extreme application. We get millions of requests to their website. We have this semantic database handling all these requests in the background. This is probably the biggest and the most interesting showcase pulling all semantic technology together.

In other spaces between media and publishers, you have plenty of different applications like Euromoney, where being able to automatically process content and understand the content better is critical so that you can generate better classification and richer metadata descriptions of this content. This allows different players in this industry to take different benefits.

Sramana Mitra: If you could just answer one last question, where do you see entrepreneur opportunities in this domain?

Atanas Kiryakov: There’re plenty of entrepreneurial opportunities, because this field is quite hungry for best practices and good applications. I mentioned a few applications. We probably have another 10, but compared to the potential of this technology, the best practices are very few. I would expect that in the next two to three years, there will be an explosion of new applications that use this new technology to be available.

Currently, most of the successful users of this technology are huge enterprises. We’re speaking of Euromoney, BBC, and the biggest pharmaceutical companies. In the next three years, this will go down the stack to a bigger number of smaller organizations and to consumer web applications. That’s still not there. There are few applications that build this technology into social networks or messaging. There’re plenty of applications to start using this technology in smaller projects.

Sramana Mitra: Thank you very much for your time.

This segment is part 5 in the series : Thought Leaders in Big Data: Atanas Kiryakov, CEO of OntoText
1 2 3 4 5

Hacker News
() Comments

Featured Videos