2D Memory Transistors for Neuromorphic Computing

The boom in artificial intelligence has driven in-memory computing for high speed and power efficiency. Motivated by the need, neuromorphic computing (NC)—mimicking biological spike neural networks—ies a promising alternative that requires artificial neurons and synapses to construct NC circuits. 2D memory transistors have received intensive interest in recent research for NC applications as both artificial neurons and synapses with additional advantages for high device density and energy efficiency. This review intends to highlight recent progress made in the exploration of 2D memory transistors and their implementation into electrical and optoelectronic NC circuits. Discussions of remaining challenges and future perspectives are also provided.