Photonic Synapses for Ultrahigh‐Speed Neuromorphic Computing

Abstract

Neuromorphic computing based on a non‐von Neumann architecture is a promising way for the efficient implementation of artificial intelligence. A neuromorphic computing system is mainly composed of artificial synapses and neurons. Various electronic neuromorphic devices have been developed by different technologies. Neuromorphic photonics combines the efficiency of neural networks and the speed of photonics to build computing systems. Compared with their electronic counterparts, photonic neurons and synapses have the potential advantages of ultrafast operation speed, large bandwidth, and no electrical interconnect power loss, inherent to the computing by optical means. Therefore, photonic neuromorphic devices are attracting more and more attention. This work reviews the recent advances in the development of photonic synapses. Both purely photonic synapses and photoelectric synapses are described. Their structures and working mechanisms are discussed in detail.

Color‐Recognizing Si‐Based Photonic Synapse for Artificial Visual System

“…It is believed that persistent PC is caused by the capture/decapture of carriers at the interface of heterojunctions, [ 9 ] the defects of the material, [10b] or the ionized oxygen vacancies [11a] . For Si material, dangling bonds are generally believed to be the origin of the persistent PC [7b] . It is well known that there exists a significant Staebler–Wronsk effect in a‐Si film, where a large number of Si dangling bonds are generated inside the a‐Si film upon light illumination, resulting in the formation of more dangling‐bond defect states in the forbidden band.…”

“…Recently, a photonic synapse has been proposed that is triggered by light and features high‐speed and broad‐bandwidth computing. [ 7 ] The nature of the direct light response combined with the neuromorphic computing ability makes it promising for efficient artificial vision systems. So far, the synaptic functions of the photonic synapses have been demonstrated based on various materials such as perovskite quantum dots, [ 8 ] Si nanocrystals, [7b] carbon nanotubes, [ 9 ] and 2D materials, [ 10 ] as well as other material systems.…”

“…[ 7 ] The nature of the direct light response combined with the neuromorphic computing ability makes it promising for efficient artificial vision systems. So far, the synaptic functions of the photonic synapses have been demonstrated based on various materials such as perovskite quantum dots, [ 8 ] Si nanocrystals, [7b] carbon nanotubes, [ 9 ] and 2D materials, [ 10 ] as well as other material systems. [ 11 ] In addition to synaptic functions themselves, the photonic synapses have also been explored for artificial vision systems, and they have shown unique advantages for light triggering.…”

Retina‐Inspired Carbon Nitride‐Based Photonic Synapses for Selective Detection of UV Light

“…[3,[7][8][9][10][11][12][13][14][15] The dual functions of artificial synapse and photosensor would give abilities that are beyond those of a simple sensing device, i.e., photonic synapses can process detected light signals and track the history including light intensity, and the number, duration, and frequency of exposure. [14,16] Thus, these photonic synapses are applicable for smart sensors for the Internet of Things, biomedical electronics, soft robots, and neural prostheses. [3] The retinas, especially the photoreceptors, are responsible for the perception of color by detecting specific wavelength, so the combination of wavelength selectivity and signal preprocessing is a significant factor in visual perception.…”

Broadband Optoelectronic Synaptic Thin‐Film Transistors Based on Oxide Semiconductors

“…To date, plenty of synaptic devices could be used to emulate efficient operating brain such as memristors, atomic force switches, and transistors . Among all of them, transistors are widely used because of three‐terminal structures that allow signals to be transmitted along with learning process.…”