Boolean Logic Computing Based on Neuromorphic Transistor

Abstract

General‐purpose computers usually use logic gate computing units based on complementary metal oxide semiconductors (CMOS). Due to the separate memory and computing units in Von Neumann architecture, data transmission requires great energy and time consumption. Developing novel neuromorphic devices and comprehensively investigating their logical computing mode are crucial to achieve high‐performance and low‐power neuromorphic computation. Here, a systematic summary of Boolean logic computing based on emerging neuromorphic transistors is presented. This summary encompasses logical operation modes, materials, device structures, and working mechanisms. The input mode of Boolean logic operation is classified into electrical input, optical input, and synergistic optical/electrical input. Besides, additional modulation strategies to construct programmable logic functions by electrical, optical, and thermal signals are also summarized. These strategies hold great significance as they enable dynamic reconfiguration of logic operations and provide neuromorphic devices with decision‐making capabilities. Finally, the application prospects and current challenges to Boolean logic computing based on dendritic integration are discussed from the aspects of device integration, synergistic input/modulation modes, auxiliary peripheral circuit, software/hardware system, etc. It is believed that comprehensive investigations on neuromorphic Boolean logic operations are crucial to push forward the development of future neuromorphic computing toward high efficiency and high integration density.

Bio‐inspired Two‐dimensional Nanofluidic Ionic Transistor for Neuromorphic Signal Processing

“…Information computation in the human brain relies on the signal integration function of nerve cells. Analogous to Boolean logic computing, the dendrites of a single neuron receive and process multiple input signals from hundreds of surrounding neurons and output a spike signal to the next neuron (Fig 5a) [22] . The construction of ion‐based logic operations that apply ions as signal carriers is significant in the development of neuromorphic computing.…”

Bio‐inspired Two‐dimensional Nanofluidic Ionic Transistor for Neuromorphic Signal Processing

“…Analogous to Boolean logic computing, the dendrites of a single neuron receive and process multiple input signals from hundreds of surrounding neurons and output a spike signal to the next neuron (Fig 5a). [22] The construction of ion-based logic operations that apply ions as signal carriers is significant in the development of neuromorphic computing. After thoroughly characterizing the voltage-gating performance of MXene-based ionic transistor, the high on-off ratio of a single device was leveraged to integrate an ion-based logic gate circuit for neuromorphic signal integration.…”

Piezotronic neuromorphic devices: principle, manufacture, and applications

“…Based on the regulation of synaptic weights, neuromorphic devices perform various tasks by sensing external stimuli, storing relevant information, and performing computations, ultimately enhancing information processing efficiency and intelligence. The simulation of information transmission/processing in neurons promises to construct efficient neuromorphic engineering, which has a wide application in adaptive and interactive AI systems [6,7].…”