![\"\"](\"http:\/\/cdn.nextgov.com\/media\/img\/upload\/2017\/05\/08\/050817brainwavesNG\/nextgov-medium.jpg\" "\\"\\"")
Modern sensors can see farther than humans. Electronic circuits can shoot faster than nerves and muscles can pull a trigger. Humans still outperform armed robots in knowing what to shoot at \u2014 but new research funded in part by the Army may soon narrow that\u00a0gap.<\/p>\n
Researchers from\u00a0DCS\u00a0Corp and the Army Research Lab fed datasets of human brain waves into a neural network \u2014 a type of artificial intelligence \u2014 which learned to recognize when a human is making a targeting decision. They presented their\u00a0paper<\/a>\u00a0on it at the annual\u00a0Intelligent User Interface<\/a>\u00a0conference in Cyprus in\u00a0March.<\/p>\n Why is this a big deal? Machine learning relies on highly structured data, numbers in rows that software can read. But identifying a target in the chaotic real world is incredibly difficult for computers. The human brain does it easily, structuring data in the form of memories, but not in a language machines can understand. It\u2019s a problem that the military has been grappling with for\u00a0years.<\/p>\n \u201cWe often talk about deep learning. The challenge there for the military is that that involves huge datasets and a well-defined problem,\u201d Thomas Russell, the chief scientist for the Army, said at a recent National Defense Industrial Association\u00a0event<\/a>. \u201cLike Google just solved the Go game\u00a0problem.\u201d<\/p>\n Last year, Google\u2019s DeepMind lab\u00a0showed<\/a>\u00a0that an\u00a0AI\u00a0could beat the world\u2019s top player in the game of Go, a game considered exponentially harder than chess. \u201cYou can train the system to do deep learning in a [highly structured] environment but if the Go game board changed dynamically over time, the\u00a0AI\u00a0would never be able to solve that problem. You have to figure out\u2026in that dynamic environment we have in the military world, how do we retrain this learning process from a systems perspective? Right now, I don\u2019t think there\u2019s any way to do that without having the humans train those\u00a0systems.\u201d<\/p>\n Their research branched out of a multi-year, multi-pronged program called the\u00a0Cognition and Neuroergonomics Collaborative Technology\u00a0Alliance.<\/a><\/p>\n \u201cWe know that there are signals in the brain that show up when you perceive something that\u2019s salient,\u201d said researcher Matthew Jaswa, one of the authors on the paper.<\/a>\u00a0These are called\u00a0P300 responses,<\/a>\u00a0bursts of electric activity that the parietal lobe of the brain emits in response to stimuli.\u00a0Discovered in the 1960s<\/a>, the P300 response is basically the brain\u2019s answer to a quick-decision task, such as whether an object that appears suddenly is a\u00a0target.<\/p>\n The researchers hope their new neural net will enable experiments in which a computer can easily understand when a soldier is evaluating targets in a virtual scenario, as opposed to having to spend lots of time teaching the system to understand how to structure different individuals\u2019 data, eye movements, their P300 responses, etc. The goal, one day, is a neural net that can learn instantaneously, continuously, and in real-time, by observing the brainwaves and eye movement of highly trained soldiers doing their\u00a0jobs.<\/p>\n \u201cIf you can improve this to the point where you can put it on guys in the field, you can get to the point where they\u2019re just looking at things and doing their normal tasks,\u201d Jaswa said. \u201cAll their years of experience that feed into that normal situational awareness. We\u2019re peeking into what their brains are doing. If you can have enough guys in a squad looking at similar things, then we can say, \u2018Three or four guys looked at this thing. It\u2019s probably\u00a0important.\u201d<\/p>\n The research does not mean that robots can now outshoot humans. There\u2019s a lot more work to do. But the neural net could make that study go a lot\u00a0faster.<\/p>\n<\/div>\n