Computer technology is constantly advancing, but there are some ways in which it’s come to a surprising halt. The integrated circuits that perform the computation process have gotten about as good as they can get with current technology. To move toward the future, researchers are looking at new ways to make computation processes faster and more efficient.
One such way to do that is to explore neural computing, a model inspired by the way human brains function. Neural computing is often seen as “fuzzy,” something that we don’t have a solid grasp on because we don’t understand the human brain well enough to simply model computers on it. That said, there are teams working on applying what we do know about the human mind to computers.
“We’re taking a stab at the scope of what neural algorithms can do,” says Braid Aimone, a computational neuroscientist at Sandia National Laboratories. “We’re not trying to be exhaustive, but rather we’re trying to highlight the kind of application over which algorithms may be impactful.”
Take learning, for example. The human brain never stops learning, and is constantly responding to stimuli while doing other things. Machines, on the other hand, learn something, test it, and then they’re done. If we could get computers to learn continually, without having to tell them to do so, that would be a huge jump forward in computer power.
Even within the limits of computing power, technology has advanced at a breathtaking pace in the last few decades alone, with advancements begetting more advancements, faster and faster. If the development of neural computing leads to better ways for computers to do their jobs, it would be possible to develop even better ways to make our lives easier, help solve complex problems through computer modeling, and use that new technology to better run businesses and understand the global economy.