Researchers Develop Highly Scalable Neuromorphic Hardware
Researchers develop a brain-inspired highly scalable neuromorphic hardware by co-integrating single transistor neurons and synapses.
Neuromorphic hardware has gained much research attention as they can implement artificial intelligence(AI) functions, and consume very little power of less than 20 watts. The neuromorphic hardware uses specialized computing architectures consisting of dedicated processing units that emulate the behavior of neurons directly in hardware, and a network of physical interconnections to facilitate the rapid exchange of information.
The neurons and synapses (that remembers the connection between two neurons) are realized by digital and analog circuits, and therefore, there is a limit in terms of hardware efficiency and costs. It is necessary to improve the hardware cost in order to apply it to mobile and IoT devices.
Researchers at the Korea Advanced Institute of Science and Technology (KAIST) fabricated a highly scalable neuromorphic hardware by co-integrating single transistor neurons and synapses. The hardware is implemented using complementary metal-oxide-semiconductor (CMOS) technology, reducing chip cost and simplifying fabrication procedures.
Professor Yang-Kyu Choi, who led the research, said that this work can dramatically reduce the hardware cost by replacing the neurons and synapses that were based on complex digital and analog circuits with a single transistor. “We have demonstrated that neurons and synapses can be implemented using a single transistor,” said Joon-Kyu Han, the first author. “By co-integrating single transistor neurons and synapses on the same wafer using a standard CMOS process, the hardware cost of the neuromorphic hardware has been improved, which will accelerate the commercialization of neuromorphic hardware.”