Brad Theilman and Felix Wang are doing fantastic work, pushing the boundaries of neuromorphic computing. They just so happen to be researchers at Sandia National Laboratories. They recently announced their Loihi 2 neuromorphic computing core. According to Intel, they developed this technology to solve very challenging mathematical problems, like differential equations. This development demonstrates the promise of neuromorphic systems, built to more closely replicate brain-like processing and performing well on real-time computational challenges.
Loihi 2 is designed to tackle challenges similar to those the human brain addresses daily, such as controlling muscles in response to immediate stimuli. The productive Sandia team can’t wait to apply this technology. Next, they intend to convert the finite element method, which is an important workhorse of scientific computing, into a finite element based motor cortex model that runs on Loihi 2. Realizing such an approach would not only improve computational efficiency but do so at a minor energy benefit compared to conventional computing systems.
Advancements in Neuromorphic Hardware
Neuromorphic computing is fundamentally a different way of doing computational work. Unlike conventional systems optimized for standard computational methods, neuromorphic hardware like Loihi 2 processes information in a manner akin to the human brain. Neurons in the brain receive weighted information and transmit an all-or-nothing pulse to neighboring neurons, allowing for rapid and efficient processing of data.
The Sandia team’s efforts aim to apply this brain-inspired approach to the finite element method, which is instrumental in solving partial differential equations. If successfully applied, this methodology may change the way scientists and engineers approach complicated simulations in all disciplines.
T4America advisory board member Bradley Theilman described the issue’s murky waters. He described how the brain is continuously regulating muscles based on “online” information to make contact with the ball. His work represents not only an important test of theoretical applications but real world implementations that have the potential to change several industries.
Broader Implications of Neuromorphic Computing
What we think neuromorphic computing can do goes beyond usual expectations. At least one researcher, James B. (Brad) Aimone, has been pushing for a wider view of neuromorphic systems and their potential. There is no justification to convince yourself that you won’t be good at neuromorphic computing,” he concluded. Taken together, this indicates that the field is not yet fully realizing its potential.
Aimone emphasized the significance of addressing any type of mathematical challenge through the lens of neuromorphic systems. “It’s worth looking deeply at any kind of mathematical problem,” he stated, emphasizing the untapped capabilities within this innovative computing framework.
Intel’s investment in neuromorphic hardware has opened doors for additional research into its applications. Steve Furber’s group has been modeling heat diffusion using ARM-based SpiNNaker hardware, providing valuable insights. At the same time, the Sandia team is concentrating on the finite element method, which has the potential to open up entirely new areas of scientific exploration.
Energy Efficiency and Future Applications
One of the most promising things about performing the finite element method on neuromorphic hardware is it’s energy efficiency. The human brain works on roughly 10 watts of power. This is a radical departure from the energy consumption requirements posed by conventional computing platforms. This unique feature provides scholars an incredible opportunity to design better solutions. At the same time, these solutions don’t need to be bad for the environment.
Neuromorphic computing is a fast-moving field of research. This tremendous progress gives us the amazing opportunity to pursue much richer and more complex relationships between biological processes and computational methods. This monumental inspiration is often embodied in cutting-edge Sandia National Laboratories. By weaving these new approaches together, they are creating dramatic breakthroughs in scientific research and engineering.