Researchers at DCS Corporation and the US Army Research Lab presented a paper at the annual Intelligent User Interface Conference in Cyprus in March, after putting datasets of human brainwaves into an artificial intelligence neural network.
Subsequently, the network learned to recognize a human being when searching for a target. The findings were the result of a years-long research program named Cognition and Neuroegonomics Collaborative Technology Alliance.
Cats at #IUI2017 how do animals react to intelligent user interfaces? pic.twitter.com/lORk9IDSXN
— Daniel Garijo (@dgarijov) March 16, 2017
Target selection in warfare is crucial to the effectiveness of a military, but also to how that military's country of origin seems worldwide. Masses of civilian fatalities, or "collateral damage," as the US army calls it, seriously undermines and erodes at the international reputation of that country.
Current technology is able to see across greater distances than the human eye, and electronic circuits can shoot faster than human reflexes and muscles can pull a trigger, however knowing what target to select is still the province of humans. The Cognition and Neuroegonomics Collaborative Technology Alliance plans to shrink that gap.
Up until now machine learning has been heavily dependent upon structured data that is fed to it in the form of software. But identifying a target with ease — as Arnold Schwarzenegger effortlessly does in the Terminator movies — is still very difficult in the real world for robots and machines. The human memory on the other hand, has that recognition innately built in thanks, in part, to memory.
"We know that there are signals in the brain that show up when you perceive something that's salient," one of the research paper's authors, Matthew Jaswa, is quoted as saying.
These reflections are called P300 responses, which consist of bursts of electric activity that the parietal lobe of the brain emits in reaction to simulation. In other words, the P300 response is the brain's answer to a quick-decision task, such as judging whether something that suddenly moves is a target.
The goals of the researchers is to create a neural net that can instantly learn, in real-time, accurate target selection by following the eye movements and brainwaves of soldiers in battle.
While the report is clear that the time has not yet come when robots are able to outgun humans, the neural net may speed that process up.
