US Army Reads Soldiers’ Brain Waves To Speed Up Target Identification
Where information gathering is concerned, it's hard to imagine any entity on Earth that captures more data than the US military. It's a constant digital battle to fetch new data and analyze it as quickly as possible in an effort to effectively move operations forward. But an obvious problem arises: there's often just too much data to churn through.
Fortunately, computer hardware and software continually evolves, and it can help dramatically improve this sort of work. The same applies to our neural functions, which our computers can read and interpret. A great example of this has just been revealed at the Aberdeen Proving Ground in Maryland.
Dr. Anthony Ries is a cognitive neuroscientist for the US Army and has found a way to help speed up the detection of important features in imagery captured by satellites or unmanned aerial vehicles. Generally, finding important details in imagery like this is tedious; it requires someone to manually scan piece of a map before moving onward. Ries found, though, that as soon as an interesting detail was spotted, it'd be reflected in the neural readout, and thus an idea was born.
It's not likely to impress anyone to learn that this is possible, since this kind of processing has been done before. What's interesting in this particular case, however, is just how much faster this processing can be done when a computer is sitting there waiting for a special detection to be made. Our brains can detect specific details in images much quicker than a computer, and Ries discovered that even if he displays 5 images per second to someone, that's still appropriate for a detection to be read by the neural monitor.
What this ultimately results in is a computer that rapidly shows different images to someone, and once the neural detection reads a spike, it'll automatically tag that particular photo for later inspection.
Given the way machine-learning is going, it wouldn't be at all surprising to see this kind of worked passed on to server clusters in the future. But for now, these improvements do seem to dramatically enhance the process.