Widely applicable neuromorphic computing, San

Watermelon random walk diffusion model

video: A random walk diffusion model based on data from Sandia National Laboratories algorithms running on an Intel Loihi neuromorphic platform.
seen more

Credit: Video courtesy of Sandia National Laboratories.

ALBUQUERQUE, NM – With the introduction of some mathematics, Sandia National Laboratories researchers have shown that neuromorphic computers that synthetically reproduce the logic of the brain can solve more complex problems than those presented by artificial intelligence and can even gain a place in high -efficiency. computing.

A random walk diffusion model based on data from Sandia National Laboratories algorithms operating on an Intel Loihi neuromorphic platform. Video courtesy of Sandia National Laboratories. Members of the media can download the video here (4.3 MB).

The findings, detailed in a recent article in the journal Natural Electronicsshows that neuromorphic simulations using the statistical method called random walks can track X-rays passing through bones and soft tissues, disease passing through a population, information flowing through social networks and the movements of financial markets, among other uses, said Sandia, a theoretical neuroscientist. and chief researcher James Bradley Aimone.

“Basically, we’ve shown that neuromorphic hardware can deliver computing advantages related to many applications, not just artificial intelligence, to which it’s obviously akin,” Aimone said. “Newly discovered applications range from radiation transport and molecular simulations to computer finance, biological modeling, and particle physics.”

In optimal cases, neuromorphic computers will solve problems faster and use less power than conventional computing, he said.

The bold claims should be of interest to the high-performance computing community because finding skills to solve statistical problems is a growing concern, Aimone said.

“These problems are not really good for GPUs [graphics processing units]on which future exascale systems are likely to depend, “said Aimone.” What’s exciting is that no one has really looked at neuromorphic computing for these kinds of applications before. “

Sandia engineer and paper author Brian Franke said, “The natural randomness of the processes you list will make them ineffective when directly mapped to vector processors such as GPUs on next-generation computing efforts. particle simulation, which can lead to a scalable and energy-efficient approach to solving problems of interest to us. “

Franke models photon and electron radiation to understand their effects on components.

Aimone said the team successfully applied neuromorphic computing algorithms to model random walks of gas molecules propagating through a barrier, a basic chemical problem, using the 50-million-chip Loihi platform Sandia received about a year and a half ago from Intel Corp. . “Then we showed that our algorithm can be extended to more complex diffusion processes useful in a range of applications.”

The claims are not intended to challenge the superiority of standard computing methods used to maintain services, desktops, and telephones. “There are, however, areas in which the combination of computational speed and lower energy costs can make neuromorphic computing the ultimately desirable choice,” he said.

Showing a neuromorphic advantage, both the IBM TrueNorth and Intel Loihi neuromorphic chips observed by Sandia National Laboratories researchers were significantly more energy efficient than conventional computer hardware. The graph shows that Loihi can do about 10 times more calculations per unit of power than a conventional processor. Energy is the limiting factor – more chips can be inserted to run in parallel, so faster, but the same electricity bill happens whether it’s one computer doing everything or 10,000 computers doing the work. Image courtesy of Sandia National Laboratories. Click the thumbnail for a high-resolution image.

Contrary to the difficulties presented by adding quantities to quantum computers – another interesting method of overcoming the limits of conventional computing – chips containing artificial neurons are inexpensive and easy to install, Aimone said.

It can still be costly to move data on or out of the neurochip processor. “As you collect more, it slows down the system, and eventually it won’t work at all,” said Sandia mathematician and paper author William Severa. “But we overcame this by setting up a small group of neurons that effectively computed summary statistics, and we output those summaries instead of the raw data.”

Severus wrote several of the algorithms of the experiment.

Like the brain, neuromorphic computing works by electrifying small ping-like structures, adding tiny charges emitted by surrounding sensors until a certain electrical level is reached. Then the pin, like a biological neuron, flashes a small electrical explosion, an action known as a sting. Contrary to the metronomic regularity with which information is transmitted in conventional computers, Aimone said, the artificial neurons of neuromorphic computing flash irregularly, as biologicals do in the brain, and so can take longer to transmit information. But because the process only drains energy from sensors and neurons if they contribute data, it requires less energy than formal computing, which must probe each processor whether contributing or not. The conceptually bio-based process has another advantage: Its computing and memory components exist in the same structure, while conventional computing consumes energy by remotely transferring between those two functions. The slow reaction time of the artificial neurons may initially slow down its solutions, but this factor disappears as the number of neurons increases, so more information is available in the same amount of time to be totaled, Aimone said.

The process begins using the Markov chain – a mathematical construct where, like a Monopoly game board, the next result depends only on the current state and not on the history of all previous states. That coincidence contrasts, said Sandia mathematician and paper author Darby Smith, with most related events. For example, he said, the number of days a patient has to stay in the hospital is at least partially determined by the previous stay.

Starting with the Markov random base, the researchers used Monte Carlo simulations, a fundamental computer tool, to power a series of random walks that try to cover as many routes as possible.

“Monte Carlo algorithms are a natural solution to radiation transport problems,” Franke said. “Particles are simulated in a process that mirrors the physical process.”

The energy of each walk was recorded as a single energy spike by an artificial neuron reading the result of each walk in turn. “This neural network is more energy efficient than recording every moment of every walk, as ordinary computing should do. This is partly responsible for the speed and efficiency of the neuromorphic process,” said Aimone. More chips will help the process move faster. using the same amount of energy, he said.

The next version of Loihi, said Sandia researcher Craig Vineyard, will increase its current chip from 128,000 neurons per chip to up to one million. Larger-scale systems then combine multiple chips into a board.

“It may make sense that technology like Loihi can find its way into a future high-performance computing platform,” Aimone said. “This could help make HPC much more energy efficient, climate-friendly and more affordable all around.”

The work was funded under the NNSA Advanced Simulation and Computing program and that of Sandia Laboratory Directed Research and Development program.

Sandia National Laboratories is a multi-mission laboratory operated by Sandia LLC’s National Technology and Engineering Solutions Corporation, a wholly owned subsidiary of Honeywell International Inc., for the U.S. Department of Energy’s National Nuclear Safety Administration. Sandia Labs has major research and development responsibilities in nuclear deterrence, global security, defense, energy technologies and economic competitiveness, with major facilities in Albuquerque, New Mexico, and Livermore, California.

Sandia news media contact: Neal Singer, nsinger@sandia.gov505-977-7255

Leave a Reply

Your email address will not be published.