Thought Leaders in Big Data: Atanas Kiryakov, CEO of Ontotext (Part 2)

Sramana Mitra: Let me drill down a bit. I don’t know if you’re familiar with my writings on this subject. I’ve written extensively on this topic. What I’d like to do is take three of your customers from three different domains, and do some use cases on how you’re applying AI technology, that you built and have expertise in solving specific business problems. For example, I saw that Euromoney is a customer. Tell us what problems you solve for Euromoney and how do you solve those problems.

Atanas Kiryakov: Allow me a minute to bridge the gap between the semantic web vision to the enterprise semantic technology. In 2000, we started working on implementing this semantic web vision. Towards 2007, it became clear that semantic web is not going to stop. All the standards and technologies that were developed to back the semantic web have very good applications in enterprise IT. Probably the most important piece of it is that the data management and data presentation query standards that were developed turned out to have a very good fit for a range of enterprise data management problems, specifically data integration.

There is a standard RDF that’s for representing data made in the semantic web. This standard is being designed to allow for management of data coming from different sources without centralized control. It turns out to be very good for a pharmaceutical company, particularly people within a pharmaceutical company that are busy with research development. This is a knowledge-intensive job  because while designing new drugs, they also have to take care of databases in chemistry, several types of biology, and experiments with drugs.

There was a study that showed that even four years ago, they had about 1,000 public databases that are actively used and referred to in research papers. There are massive amounts of databases out there that are immediately useful for the drug that any pharmaceutical company would want to develop. The biggest issue was that in the process of this research, they often need to make queries that span across five to ten of these databases. That’s just for one query. Then you have another query that needs to go to another five to twenty databases to answer the relevant question. They have this very challenging problem that is literally not solvable, or impractical to solve with relational database technology or other data warehousing technologies.

Pharmaceutical companies were among the leaders who started using semantic web standards for data integration of research-related databases. We have a number of clients in this field. We have a data warehouse that combines 30 public databases that we constantly monitor. The volume of information that is in there is quite extensive. All together it goes to the range of 10 to 20 billion facts that are consolidated. That’s a massively challenging problem.

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