How brainwaves could teach software to ID threats

The Army can collect a lot of imagery on the battlefield from drones and other sensors, amassing a trove of data that can be used to identify potential threats. The problem is that the best way to actually ID a threat is still for a human to look at the image, which isn’t always practical when the images number in the millions.

Software, at the moment, can only go so far in analyzing those images and making a threat-or-no-threat decision. So the Army Research Laboratory has started a research effort aimed at using human brainwaves to one day “train” computers to recognize what threats look like. Jean Vettel, a neuroscientist with ARL at Aberdeen Proving Ground, Md., talked about the research at the Defense Department’s recent Lab Day, according to an Army release.

In order to teach a software algorithm to identify threats, ARL first has to collect a wide variety of images depicting both benign and threatening situations. Vettel and ARL’s neuroscience branch then will equip experienced combat soldiers with EEG, or electroencephalography, headgear that can sense and record their brainwave activity. Researchers would show a series of images on a screen, and the soldier could tag the image using only brainwaves—a faster and more efficient method than asking them to give a verbal cue or press a button, Vettel said.

"And then when we have images labeled, we can take those images and give it to a machine learning algorithm that can learn to distinguish between threatening or non-threatening images,” she said. “We'll have used soldier expertise to train the algorithm."

Eventually, the system also could be used to train less-experienced soldiers in threat detection, by having them learn from soldiers who have been deployed.

Vettel's branch also wants to get this technology out into the field, having developed a combat helmet with EEG sensors on the inside that could be connected to a smartphone or tablet. Vettel said experienced soldiers have told her that they can have a gut instinct about threatening situations that can sometimes be hard to put into words. With an EEG system, their heightened awareness or anxieties could be communicated to others in their squad, raising awareness for everyone.

"Not only is this good for units where you have novice soldiers," Vettel said, "but also for soldiers who are sleep deprived."

Soldiers won’t be wearing EEG-equipped headgear anytime soon, but Vettel said the technology could be available in about 25 years.

"We are a future capability," she said. "A lot of our research now focuses on reliably recording these signals in complex settings. It's, overall, letting the technology adapt to the soldier, rather than having the soldier adapt to the technology."