Sifting Through The Data Stacks

With more information coming into the US Air Force’s Distributed Common Ground System (DCGS) than ever before—from space, cyber, and air sensors—the service needs to find a way to embrace new concepts such as machine learning, metadata tagging, and open source integration, Lt. Gen. Robert Otto said Sept. 14 at ASC1?5. Speaking at a panel with representatives from Microsoft and a Silicon Valley artificial intelligence startup, Otto said USAF needs to recognize the “tremendous” value the private sector can bring to the ISR mission, specifically in these realms. James Crawford, CEO of Silicon Valley imagery analysis firm Orbital Insight, noted commercial satellite exploitation is now so developed users can track trends ranging from the global oil supply to the pace of construction projects in China, but making sense of the data is where the future of machine learning is going. First you compare “pixels to numbers” on any subject, then get “numbers to insight,” Crawford said. Aside from machine science and data, humans will continue to be very important. “What we are trying to do is sift out the haystack into a very small pile of hay,” he said, in order to give humans “the time to look at the smaller pile of data.”