New Software Uses Artificial Intelligence To Sift Through Data
By: Valerie Insinna
February 1, 2015
As the military and U.S. intelligence agencies struggle to digest mountains of information captured from social media and sensors all over the world, companies continue to create software that can more quickly and accurately help analysts isolate relevant data.
One new platform under development, called Via, is able to independently scan data streams and alert users if patterns of interest are found, said Eric Little, vice president and chief scientist of Modus Operandi, a Melbourne, Florida-based software company.
Via features a “semantic reasoner,” a type of artificial intelligence that applies human-like logic to find correlations among pieces of information, he said.
“A lot of intel analysis is about pattern analysis and pattern recognition. Analysts have lots of patterns in their heads that they understand,” he said. Instead of having to manually search through data to build a case, they can codify what they are looking for into the Via software, and it will continuously run that algorithm and look for information that is similar.
“Think of it as very complex if-then statements. If someone is a member of this group, and they have some friends who are weapons dealers and they have some friends who are financiers, then they are probably linked to terrorist activities,” he said.
Most legacy big data programs can perform statistical analysis, but cannot use logic to make more sophisticated connections among multiple pieces of information, Little said. For instance, older software can map hot zones of improvised explosive device activity based on past incidents. Using Via, a user can fuse that with other intelligence relating to IED attacks, such as the probability of such an event happening on a day where there is inclement weather or whether the bombs found in a certain area comprise similar parts.
Another feature of the program is that its interfaces are modeled to look like existing social media platforms such as Facebook, Wikipedia and Yelp, he said. Users can upload photos, add comments to information, or even rate other analysts’ work.
“We have customers in the intel space who are normally fairly young, fairly inexperienced, with limited education backgrounds,” he said. “We borrow a lot from commonly used types of applications such as Facebook or Wikipedia or something like that. When you open the software, it already has somewhat of a familiar look and feel to it.”
“We are currently working with a couple of our military customers right now,” Little said. The product has not undergone field unit testing yet, but is being employed in some of Modus Operandi’s existing small business innovation research grants.Source: http://www.nationaldefensemagazine.org/articles/2015/1/31/2015february-new-software-uses-artificial-intelligence-to-sift-through-data