It’s been a couple of weeks now since the terrorist attack at the Boston Marathon, and since then the media and others have been questioning whether there was an intelligence failure at the FBI. It’s not an easy question to answer, given how many suspicious persons and activities government agencies must have on their radars, and all the data there is to deal with.
It’s undeniably hard work. Semantic technology does come into play to help the government connect the dots on information related to threats and other intelligence issues. For example, Dennis Wisnosky, formerly CTO and Chief Architect, Business Mission Area, U.S. Department of Defense, discussed semantic technology’s role in that institution at SemTech in San Francisco in 2011. (Wisnosky now is spending more time with FIBO in his role providing technical strategy and operational guidance to help the The Enterprise Data Management Council finalize and implement Financial Industry Business Ontology standards – see story here.) It’s also been reported that the FBI and CIA are practicing and developing semantic processing techniques to analyse social media to improve situational awareness and identify emerging threats.
Semantic tech is being applied by military intelligence, too, in services like the Air Force and Marine Corp., which are testing or have live deployments of semantic systems developed by Modus Operandi. The Air Force is involved with analyzing email traffic on a 24/7 basis, while the Marine Corp. works with it as part of the cyber portion war games exercises. The Office of Naval Research also is utilizing it as part of its lab exercises, says Modus Operandi president Rick McNeight.
For the intelligence analyst, the vendor provides a semantic Wiki on the front-end to afford access to the massive amounts of data, that it pulls into its system from various databases, applying natural language processing and text analytics as it extracts information, then integrates it and builds relationships across the data.
“Extrapolating terms out of text doesn't do much good unless you can put terms in a model and run reasoning against those, and link that unstructured data maybe to other areas of structured data to fill in information,” says chief scientist Dr. Eric Little. The data includes full motion video, as well, which is meta-tagged and fed to its software by other sources for Modus Operandi to process, correlate and present to analysts.
When analysts set up a page in the wiki, the software does a historical search to fill appropriate existing data automatically and dynamically adds new events as they occur, when they match the criteria the analyst has specified. “The important thing is we start to interconnect relationships - Terrorist A has a relationship to Terrorist B, C or D or these relatives,” McNeight says. Analysts can use the wiki to create models -- how often one person has been with another in certain places or certain times, for instance.
“That's the semantic part of what we do. We actually search those relationships based on the model of what the analyst wants to look for, and that also can be represented graphically so that an analyst clicks on a dot and search lines as he searches through the data,” he says.
The Big Data challenge around all this is very real, the company says. “The trick is to get [triple stores] to scale to the cloud,” says McNeight. “Even triple stores that run on multi-core processors [on a single machine] can scale to some degree, but at some point they don’t hack it anymore. The next big innovation is to move them to the cloud.” Modus Operandi is working to those ends and expects to see its research come to light within the next year or so.
When it comes to the actual impact that the military has seen from gaining the abilities Modus Operandi enables, what McNeight can say is this: It is likely that there have been incidents where troops overseas were able to avoid certain roads because the dots were connected to point out that IEDs were lining their usual paths and enable them to make a diversion. Although lives may have been saved on those days, “we don’t know about it.”
Modus Operandi has in its sights the application of its technology to the commercial marketplace as a next step. "The Big Data problem is recognized in multiple commerical environments," McNeight says, and along with that will come recognition of "the potential power of using semantic technology to correlate data and provide humans some insight into making better decisions."