A multi-institutional collaboration including the US Department of Energy’s Argonne National Laboratory has come up with a novel material that can be used to make computer chips. It uses neuromorphic circuitry and computer architecture to mimic brain functions. Shriram Ramanathan, a professor from Purdue University, led the team.
Subramanian Sankaranarayanan, a paper coauthor with a joint appointment at Argonne and the University of Illinois, Chicago, said that the human brain can actually alter as a result of learning new things. He further said that they had developed a tool for machines in order to rearrange their circuits like a brain.
This capability will make AI based computers execute difficult tasks, such as analyzing complex medical images, more promptly and precisely using even a lot less energy. We can also expect autonomous vehicles and robots in space with the capability to rewire their circuits depending on environment in the future.
Perovskite nickelate (NdNiO3) is the key material in the novel device and consists of neodymium, nickel, and oxygen. The team mixed the aforementioned material with hydrogen and attached electrodes to it which let electrical pulses be applied at various voltages.
This new material is having a multi-layered personality. It can turn on and block electrical current and can store and release electricity as well. The novel thing about it is the addition of two functions akin to the separate behavior of synapses and neurons in the brain.
The experimental results show that simply changing the voltage manipulates the motion of hydrogen ions within the nickelate. A certain voltage converges hydrogen at the nickelate centre, creating neuron-like behavior. That hydrogen is shuttled out of centre by a different voltage; this results in a synapse like behavior. At various voltages, the resulting locations and concentration of hydrogen produce the on-off currents of computer chips.
With this nickelate device, scientists will work on creating a system of artificial neurons and synapses that have the capability to learn and modify from experience. This system of neurons will grow or shrink subject to the arrival of new information and thus will have the ability to work with extreme energy efficiency. This energy efficiency translates into low operational costs.
This device hopefully has a bright future. This is because the device can be created at room temperature.
DOE Office of Basic Energy Sciences, the Air Force Office of Scientific Research, and National Science Foundation had funded the Argonne related work.
Subscribe to our newsletter to get our newest articles instantly!