Intel Corp. said today it’s expanding its research efforts on future semiconductor technologies with the U.S. government’s National Nuclear Security Administration, signing a three-year deal with Sandia National Laboratories to explore the value of “neuromorphic computing” for scaled-up computational problems.

Neuromorphic computing refers to a new method of computer engineering that’s modeled on the human brain and its nervous system. The idea is to create artificial neural systems that are inspired by biological structures. Proponents believe that neuromorphic computers will one day be able to learn, retain information and make logical deductions in the same way as the human brain does.

Neuromorphic computing involves building artificial neural systems that implement “neurons,” which are the nodes that process information, and “synapses,” which connect those nodes, to transfer electrical signals using analog circuitry. This enables them to modulate the amount of electricity flowing between those nodes to mimic the varying degrees of strength that naturally occurring brain signals have.

Intel said Sandia will begin its research into neuromorphic computing using a new, 50 million-neuron Loihi-based system that was recently delivered to its facility in Albuquerque, New Mexico.

Loihi is a new kind of architecture developed especially for neuromorphic computing that co-locates memory and processing together on each individual neuron, rather than having separate areas for each task. Intel announced Loihi in September 2017, saying it was capable of incredible learning speeds thanks to its ability to mimic how the human brain functions. One of its most interesting capabilities is that it can use knowledge it already possesses to make inferences about new data, speeding up its learning process exponentially over time.

Intel said that most research on neuromorphic computing has focused on its potential for edge computing use cases. But it believes the architecture could also be applied to larger and more complex computational problems that necessitate real-time processing, problem solving, adaptation and learning. Specific use cases include scientific computing, counterproliferation, counterterrorism, energy and national security, Intel said.

Sandia has a long history of researching scientific and numerical computing, and it plans to expand on this work by using Intel’s neuromorphic system to perform scientific simulations such as modeling particle interactions in fluids, plasmas and other materials. These kinds of simulations require increasingly advanced capabilities in data science, machine learning and optimization, Sandia said.

“By applying the high-speed, high-efficiency and adaptive capabilities of neuromorphic computing architecture, Sandia National Labs will explore the acceleration of high-demand and frequently evolving workloads that are increasingly important for our national security,” said Mike Davies, director of Intel’s Neuromorphic Computing Lab.

Constellation Research Inc. analyst Holger Mueller told SiliconANGLE that neuromorphic computing was an exciting new architecture that CIOs would do well to pay attention to.

“We know that traditional CPUs aren’t able to support key use cases for the next generation applications that enterprises need to succeed in the 21st century,” Mueller said. “Finding a CPU architecture that mimics the human brain efficiently in support of new apps is one of the key efforts that’s being undertaken today.”

The three-year research effort will involve more powerful systems in the future. Sandia is set to be one of the first recipients of Intel’s next-generation neuromorphic research system, which could exceed 1 billion neurons in computational capacity.

Image: geralt/pixabay