Semantic databases are poised to revolutionize the way that financial institutions conduct risk assessments, according...
to David Saul, the chief scientist at Boston, Mass.-based financial services giant State Street Corp.
Saul -- a senior vice president at State Street who took on the added role of chief scientist about a year ago -- believes that speedier risk assessments represent just one of many benefits that semantic technology can offer financial services organizations.
The phrase “semantic database” refers to a data model that allows users to evaluate and analyze information from disparate sources. Semantic databases measure the relationships between data to help users identify and analyze “what if” scenarios. The ability of semantic databases to measure relationships makes them particularly well suited to the financial services industry, according to Saul.
SearchDataManagement.com spoke with Saul recently to learn more about his ongoing research into semantic databases. He talked about the definition and benefits of semantic technology and offered a glimpse into State Street’s own semantic database pilot program. He also had some advice for organizations considering a similar project. Here are some excerpts from that conversation:
How do you describe the concept of semantic database technology?
David Saul: The first exposure that I had to this at a conceptual level was a long time ago -- probably longer than I’d like to remember. In the very early days as the Internet was getting started, people were talking about hyperlinks and how you would connect one site to another. Today, we don’t even think about that. We go to one site. We see a link and go someplace else. The reason that we’re able to [achieve that connectivity] is that there is a standard that exists. That standard, which is HTTP, is really the same conceptual basis for semantic databases. [My view] of semantic technology is that it is going to provide us with the same kind of connectivity for data that we have today in the World Wide Web between locations.
For more on data management in the financial services industry
Read the Whatis.com definition of the Dodd-Frank Act
Find out why the Dodd-Frank Act could mean data management problems for some
Read why reference data management is causing headaches in the financial sector
Semantic technology has been around for quite a long time, but it seems that we have been hearing more about it lately. Why is that?
Saul: One of the reasons is that, like a lot of technologies, it has matured. But we’re also getting to the point where hardware costs [for storage and processors] have gone down and we can now start making some of the tradeoffs which may have been too expensive before.
Could you give me an example of how semantic databases can prove useful at financial services firms?
Saul: There are lots of examples if you use your imagination. One has to do with risk assessments. Today we’ve got a lot of internal risk assessments that we do. We’re looking at risk information that might be coming from multiple transactional systems; it might be coming into us from external sources; and it might even be included in an email communication. Now [suppose] we want to look and [find out] our total risk exposure to a particular geographic environment. [One way] to do that is to pull reports from all of these systems with all of the risk information and then somebody sits down and does it manually. Or maybe they take all of these things and they build a spreadsheet that has information from all of them. That is a slow process and it could be an inaccurate process. For a larger requirement, what we might do is take extracts of risk data from multiple systems, from external sources and then we would build a risk repository, which is yet another database, and from that we would develop a series of reports that takes even longer.
How would semantic database technology speed up this process?
Saul: What a semantic database would do in that case is leave the data where it is. It’s not extracting it and creating yet another database. It has a complete map of exactly what that data is and the semantic tool allows you to take a view into all of those pieces, put it together and produce the answer to that risk question you were asking. It holds out the potential for making those risk assessments automated in a way that we’ve never had before. [It offers a] capability to do this much more quickly and adapt to changing requirements much more quickly than in the past.
Is State Street doing this right now?
Saul: We are doing this in a pilot phase and we are doing it with a view very much to doing the same thing with our customers, because customers have exactly the same kinds of issues. The next step is to go into larger scale production and also to go as soon as we possibly can to work jointly with some of our customers on integrating some of our data with their data. [We see semantic databases] as a very fertile ground for collaboration with our customers, where we’re applying the semantic technology to problems that they’re [experiencing].
In summary, what do you think are the key benefits of semantic databases?
Saul: Time is one benefit that I would put at the top of the list [because semantic technology offers users] the ability to integrate a number of data sources very quickly. I see that as the primary benefit. The second is accuracy. We’re going to be able to certainly do these integrations a lot better than any manual process. We’re solving a problem that everyone in the financial services industry has. We’ve got lots and lots of data and if we become better able to integrate and correlate that data and turn it into really useful information, then we’re going to be able to make better decisions more quickly.
What advice would you have for somebody considering a semantic database project?
Saul: My advice would be to get started soon because it’s one of those things where you can start small. The other piece of advice would be to think of this not just as an IT project, because most of the benefits are going to come from the business users, the consumers of this data. Get them involved early on because they’re the ones who best understand the semantics of the data in the first place. They own the data and they’re going to do the best job of describing the data.