Northrop Grumman is developing new technology to allow intelligence systems to learn patterns and predict, rather than detect, threats, said Patrick Antkowiak, vice president of the company’s advanced concepts and technologies division. “Some of the best chess players in the world now are computers. What’s going on there is that this idea of machine learning,” or enabling a computer to predict an opponent’s move with greater accuracy based on past data, explained Antkowiak in a briefing Wednesday in Washington D.C. “Think of what happens when we combine that ability with our sensing technology” on battlefield surveillance platforms such as JSTARS or Global Hawk, he added. “We get sensor architectures that can learn…and don’t just forensically give us a view of where we’ve been, but start to predict where we need to be,” noted Antkowiak. “We go from response, to anticipatory sort of systems.” Though the concept is still in the research and development stage, it could theoretically be retrofitted to many existing systems, he said.
With key members of Congress wavering on the possibility of a $350 billion defense reconciliation bill, defense experts told Air & Space Forces Magazine that the Pentagon is likely drawing up budget backup plans—but such plans would face hard choices between high-end weapons and low-cost drones and other programs in…