Can ontology, an esoteric study of the nature of existence, and computer science combine to help manage military data? An April 18 workshop at the University at Buffalo (UB) will explore this big data conundrum, as well as related topics.
The one-day "Ontologies for Information Integration" event will include presentations from experts in ontology and military intelligence, according to workshop co-organizer Barry Smith, a UB professor and director of the National Center for Ontological Research, which provides ontological services to a variety of organizations, including the U.S. Department of Defense.
For ontology researchers, military intelligence is an intriguing challenge. The U.S. Army, for instance, uses a cloud-based system called the Distributed Common Ground System-Army (DCGS-A) to collect, process and distribute large amounts of data from myriad sources.
[ The military intelligence cloud is about to get bigger. Read Military Plans Multi-Exabyte Storage Cloud. ]
"It's a gigantic, big data monster … which they try to put all their data into, particularly (information) pertaining to things like terrorist movements in Afghanistan," Smith told InformationWeek in a phone interview. As a data management platform, however, DCGS-A is far from perfect.
"As you can imagine, this data is very heterogeneous," said Smith. "It includes things like weather and disease data, and (information) about meetings, weapons and so forth. And it comes from many different sources, which means that it's practically impossible to search, in any sensible way, by the usual retrieval procedures."
Government officials have tested various strategies to make this data easier to manage, but more work is needed. "I'm part of an attempt to demonstrate that ontology can help to retrieve data from this cloud store," Smith said.
The DCGS-A challenge involves structured data, but one where the "structures are very different from one case to the next," Smith explained.
One database, for instance, might include information about people with certain skills. "In that database, the heading for people would be something like 'P,'" said Smith. "And then you have another database, which is about people and their addresses. And in that database, the heading for people might be something like 'Person.'"
And then a third database might include data about people's organizations. Its heading for people might be "human beings."
Not surprisingly, this lack of consistency spells trouble.
"You and I know that a person is a human being, but a computer doesn't know that," Smith said. "Ontology … gives you a smaller set of labels so that you can tag those data headings using common labels, and thereby merge the data in ways that prove useful for retrieval and analysis." This approach shares common ground with the proposed semantic Web, a framework that would extend Web principles from documents to data.
"In essence, all you're doing is tagging data and giving it a description of what it is," Cambridge Semantics chief technical officer and semantic Web guru Sean Martin told InformationWeek in November.
According to Smith, researchers are developing ontologies to address the information integration needs of military and other complex government projects. One such undertaking is the development of a next-generation air traffic control system, a massive project that requires input from government and commercial entities, including the U.S. Air Force, weather agencies and international airlines.
"The idea is to get rid of the control towers, and the air traffic control will take place inside the cockpit," said Smith. Air traffic control would be managed via "a gigantic network of computers inside airplanes, rather than what we have at the moment … a point-to-point messaging system between people in towers and human beings in cockpits," he said. However, the project's global scope presents many challenges, including the fact that airlines outside the U.S. have different ways of classifying weather phenomena and other data.
Companies want more than they're getting today from big data analytics. But small and big vendors are working to solve the key problems. Also in the new, all-digital Analytics Wish List issue of InformationWeek: Jay Parikh, the Facebook's infrastructure VP, discusses the company's big data plans. (Free registration required.)