Researchers at Sandia National Laboratories and Stanford University are pursing neuromorphic computing. The exceptional ability of brain to remember, learn, and process huge amount of data is the inspiration behind the study.
The scientist earlier developed a segment of computer, a device, who’s one part acted as an artificial synapse. This structure is similar to the way neurons exchange information in the brain.
The journal Science published the report, on 25th April. The team of scientists reported that the prototype of nine such devices operated and performed better than the expected. Additionally, the factors like energy efficiency, processing speed, durability, and reproducibility are the criteria to measure and rate the performance.
Traditional Electronics to be Integrated with Artificial Synapse
The researchers are looking to blend traditional electronics with artificial synapse. This combination is likely to emerge as a stepping stone contributing in artificial intelligent learning in small gadgets.
The artificial synapse is like a battery allowing scientists can direct the up or down flow of energy between two terminals. This flow of electricity depicts how the brain functions and processes the information. The design is particularly efficient since the memory storage and data processing take place in one go. However, the data is first processed and then moved further for the storage, in conventional computers.
Scott Knee, a graduate student from Alberto Salleo lab, and lecturer of materials science and engineering at Stanford co-authored study. If a memory system learns with speed and energy efficiency, it can be implemented in laptops and phones. Moreover, this would provide various opportunities to build own networks and assist in problem solving. The performance of these devices calls for monitoring to program various artificial synapse.