IT Home reported on December 23 that scientists have made another major breakthrough in the field of AI, and an AI synaptic transistor that mimics the human brain has recently come out.
This AI synaptic transistor behaves like human cognition, capable of processing and storing information at the same time, marking a significant shift from traditional machine-Xi tasks to performing associative Xi.
图源：Xiaodong Yan / Northwestern University
More importantly, this AI synaptic transistor can operate effectively at room temperature, and has the advantages of fast operation, low energy consumption, and information retention without power supply, and has a wide range of applications.
In a press release on the study, which was jointly led by researchers from Northwestern University, Boston College and the Massachusetts Institute of Technology (MIT), Professor Mark Hersam said:
The architecture of the brain is fundamentally different from that of a digital computer.
In a digital computer, data needs to move back and forth between a microprocessor and memory, which inevitably consumes a lot of energy and creates bottlenecks when trying to perform multiple tasks at the same time.
In the human brain, memory and information processing are co-located and fully integrated, resulting in orders of magnitude more energy efficiency. Our synaptic transistors also enable concurrent memory and information processing functions that more faithfully mimic the brain.
The Hessam team developed synaptic transistors, primarily using the Moiré pattern.
Note: A Mohr pattern is a geometric pattern that appears when two patterns are superimposed on top of each other. When 2D materials are stacked, new properties emerge that are not present in a single-layer structure. When these hierarchies are twisted to form Mohr patterns, it is possible to harvest unprecedented electronic properties.
In this new synaptic transistor, the researchers combined two different types of atomic-thin materials, bilayer graphene and hexagonal boron nitride. Graphene and hexagonal boron nitride are structurally very similar, but they are also different enough to produce an unusually strong Moiré effect, which enables the manipulation of the electronic properties of the graphene layer.
This manipulation can be performed at room temperature, producing synaptic transistors with enhanced neuromorphic functions.
In the Lenovo Xi test, the AI transistor can compare "101" and "111" to distinguish and identify the latter as more similar to "000".
This ability to process complex and imperfect inputs has significant implications for real-world AI applications.
The reference address of the paper is attached
-
Yan, X., Zheng, Z., Sangwan, V.K. et al. Moiré synaptic transistor with room-temperature neuromorphic functionality. Nature 624, 551–556 (2023).
https://doi.org/10.1038/s41586-023-06791-1