The Computer Of The Future Will Be An Electronic Human Brain

By admin,

  Filed under: AI, Biomimicry, Computing, Electronics, Neuromorphic Hardware, Neuroscience, Qualcomm, Zeroth
  Comments: Comments Off on The Computer Of The Future Will Be An Electronic Human Brain

[Image via Shutterstock]
Qualcomm is perfecting new software that can better process information from huge amounts of sensors and deal with uncertainty–just like the powerful computer already in our heads.Earlier this year, President Obama announced what is perhaps the biggest collaborative scientific undertaking in the U.S. since the Human Genome Project. The BRAIN Initiative is a $300 million project to map activity in every one of the brain’s billions of neurons. There are countless scientific institutions working on the project. And then there is Qualcomm, a company best known for making wireless telecommunications chips.

Not only is Qualcomm an active participant in The BRAIN Initiative, it’s also working on an entirely new class of processor, called the Neural Processor Unit (branded as the Qualcomm Zeroth), that could have applications in the medical world. But what’s a wireless company doing in the BRAIN world? What is one example we see in biology that’s good at processing sensory information efficiently? That’s the brain.

For years–long before the BRAIN Initiative was announced–Qualcomm has been working on tools for large-scale brain simulation through its partner Brain Corp, a startup with a staff of neuroscientists. “In the business of wireless technology and mobile devices, we’re seeing more devices with all types of sensors. In order to process and derive meaning from sensor information, you can do a lot of processing and throw horsepower at it, or say what is one example we see in biology that’s good at processing sensory information efficiently? That’s the brain,” explains Samir Kumar, director of business development at Qualcomm.

ORIGINAL: FastCo Exist

Think about how much the brain takes in through the senses–every millisecond, new sounds, sights, touches, and tastes are instantaneously processed and fused together into information about the world. Qualcomm is trying to replicate that ability so a smartphone could, say, fuse together data about the environment for a comprehensible look at immediate surroundings.

Qualcomm’s research also has applications in the medical world. “The brain is filled with electrical activity, and activity patterns–consciousness, behavior, those kinds of activities–people have never been able to observe before, or develop treatments for when those patterns [indicate disease],” says Tony Lewis, senior director, product management, corporate R&D at Qualcomm. With help from Qualcomm’s research, he says, “people will look at complex [brain] patterns and be able to identify them, to come up with treatments and restore functionality.

The wireless company’s brain-related work goes beyond research into brain simulation. Qualcomm is also working on a real piece of hardware, the Zeroth. “If you look at a normal CPU, it’s built to do things like balance a checkbook, but if you want to deal with uncertain information, probabilistic information, having 64 bit precision is not the best use of hardware resources,” says Lewis. “We can design better processors to handle this info. We think that we can handle uncertain, fuzzy data much more efficiently.

Qualcomm doesn’t anticipate replacing today’s processors with the Zeroth–instead, it would act as a co-processor. The Zeroth could be good at things like learning and then anticipating user preferences on a smartphone, for example. In the car, the Zeroth could help vehicles become aware of the surrounding environment and protect drivers from unsafe situations. Is it a little eerie? Sure. But as we inch closer towards technology like autonomous vehicles, we’re already ceding much of our control over to computers.

These days, Qualcomm is refining the Zeroth design, but it doesn’t have a firm production date yet. But Lewis says there is “a lot of interest” in the processor from academics and big companies alike.

Comments are closed for this post.