Researchers have reported a nano-sized neuromorphic memory device that emulates neurons and synapses simultaneously in a unit cell, another step towards achieving the goal of neuromorphic computing designed to rigorously mimic the human brain with devices semiconductor.
Neuromorphic computing aims to achieve artificial intelligence (AI) by mimicking the mechanisms of the neurons and synapses that make up the human brain. Inspired by the cognitive functions of the human brain that current computers cannot provide, neuromorphic devices have been widely studied. However, current neuromorphic circuits based on complementary metal oxide semiconductor (CMOS) simply connect artificial neurons and synapses without synergistic interactions, and concurrent implementation of neurons and synapses remains a challenge. To solve these problems, a research team led by Professor Keon Jae Lee from the Department of Materials Science and Engineering implemented the human biological working mechanisms by introducing neuron-synapse interactions into a single memory cell, rather than the conventional approach of electrically connecting artificial neural and synaptic devices.
Similar to commercial graphics cards, previously studied artificial synaptic devices are often used to accelerate parallel computations, showing clear differences with the operational mechanisms of the human brain. The research team implemented the synergistic interactions between neurons and synapses in the neuromorphic memory device, mimicking the mechanisms of the biological neural network. Moreover, the developed neuromorphic device can replace complex CMOS neural circuits with a single device, providing high scalability and cost-effectiveness.
The human brain is made up of a complex network of 100 billion neurons and 100,000 billion synapses. The functions and structures of neurons and synapses can change flexibly depending on external stimuli, adapting to the surrounding environment. The research team developed a neuromorphic device in which short- and long-term memories coexist using volatile and non-volatile memory devices that mimic characteristics of neurons and synapses, respectively. A threshold switch device is used as the volatile memory and a phase change memory is used as the non-volatile device. Two thin-film devices are integrated without intermediate electrodes, implementing the functional adaptability of neurons and synapses in neuromorphic memory.
Professor Keon Jae Lee explained, “Neurons and synapses interact with each other to establish cognitive functions such as memory and learning. Simulating both is therefore an essential element for brain-inspired artificial intelligence. The developed neuromorphic memory device also mimics the recycling effect that enables rapid learning of forgotten information by implementing a positive feedback effect between neurons and synapses.”
Source of the story:
Materials provided by Korea Advanced Institute of Science and Technology (KAIST). Note: Content may be edited for style and length.