The researchers reported a nanometer-sized neuromorphic memory device that emulates neurons and synapses simultaneously in a unit cell, another step toward completing the goal of neuromorphic computing designed to rigorously simulate the human brain using semiconductor devices.
Neuromorphic computations aim to implement artificial intelligence (AI) by mimicking the mechanisms of neurons and synapses that make up the human brain. Inspired by the cognitive functions of the human brain that modern computers cannot provide, neuromorphic devices have been extensively studied. However, modern neuromorphic schemes based on complementary metal oxide semiconductors (CMOS) simply connect artificial neurons and synapses without synergistic interaction, and the simultaneous implementation of neurons and synapses remains a challenge. To solve these problems, a research team led by Professor Keon Je Lee from the Department of Materials Science and Technology introduced the biological mechanisms of human work by introducing the interaction of neurons and synapses in a single memory cell rather than the traditional electrical connection approach. artificial neural and synaptic devices.
Like commercial video cards, artificial synaptic devices, previously studied, are often used to speed up parallel computing, showing obvious differences from the operating mechanisms of the human brain. The research team implemented a synergistic interaction between neurons and synapses in a neuromorphic memory device, emulating the mechanisms of a biological neural network. In addition, the developed neuromorphic device can replace complex CMOS circuits of neurons with a single device, providing high scalability and economy.
The human brain is made up of a complex network of 100 billion neurons and 100 trillion synapses. The functions and structures of neurons and synapses can vary flexibly depending on external stimuli, adapting to the environment. The research team developed a neuromorphic device in which short-term and long-term memories coexist using non-volatile and non-volatile memory devices that mimic the characteristics of neurons and synapses, respectively. A threshold switch device is used as a non-volatile memory, and a phase switching memory is used as a non-volatile memory. The two thin-film devices are integrated without intermediate electrodes, realizing the functional adaptability of neurons and synapses in neuromorphic memory.
Professor Keon Zhe Li explained: “Neurons and synapses interact with each other to establish cognitive functions such as memory and learning, so modeling both is an important element for brain-inspired artificial intelligence. The developed neuromorphic memory device also mimics which allows the rapid assimilation of forgotten information by realizing the effect of positive feedback between neurons and synapses ”.
Source of history:
Materials provided Korean Advanced Institute of Science and Technology (KAIST). Note: Content can be edited by style and length.