Tool sought to ID data links

The Office of Homeland Security is evaluating technologies that could help tie together information held by different agencies, and eventually other levels of government and the private sector, without violating legal or privacy barriers.

The office is working with research and development groups within federal agencies, industry and academia to look at technical tools that could identify the links among data held by separate sources, finding the "nonobvious relationships" and areas where potential leads need to be investigated by analysts, said Steve Cooper, senior director for information integration and chief information officer at the Office of Homeland Security.

The technology the office is looking at would find linkages solely through descriptions of what type of data is held by each source — possibly putting it into categories such as locations, people and goods.

Furthermore, the data could be in electronic, paper or any other format because not all information is held in databases at this point, Cooper said Nov. 4 at the Industry Advisory Council's Executive Leadership Conference in Hershey, Pa.

Such technology tools would not actually examine the content of the data. Instead the tools would narrow the search enough so that analysts would take the next step, only having to share information that is likely to be significant, Cooper said. That process eliminates problems of exposing classified and legally sensitive data, and it also avoids the cultural stigma of giving up control of information.

"We are exploring different approaches that allow us to integrate the information without actually seeing the content at the time of integration," Cooper said. "By analyzing what's interrelated to what, I can identify some patterns. I can see what might be worth taking some additional looks at."

Researchers in the intelligence community have been working on this technology for some time, trying to find ways to share information without revealing the source of the information or the method in which it was gathered — the traits that make information classified. Industry also is working on such tools, but for now, there aren't any commercial off-the-shelf solutions that will fit the needs of the office, Cooper said.

The Office of Homeland Security has several short-term pilot projects planned, starting with data held within the intelligence community. The office plans to use some of the funding that has been requested in the bill to create the Homeland Security Department, Cooper said, and that "should be sufficient to get us rolling."

Much more research needs to be done, Cooper said, and the Defense Advanced Research Projects Agency is helping the office with more long-range evaluations.