Research progress in optical neural networks: theory, applications and developments

Abstract

With the advent of the era of big data, artificial intelligence has attracted continuous attention from all walks of life, and has been widely used in medical image analysis, molecular and material science, language recognition and other fields. As the basis of artificial intelligence, the research results of neural network are remarkable. However, due to the inherent defect that electrical signal is easily interfered and the processing speed is proportional to the energy loss, researchers have turned their attention to light, trying to build neural networks in the field of optics, making full use of the parallel processing ability of light to solve the problems of electronic neural networks. After continuous research and development, optical neural network has become the forefront of the world. Here, we mainly introduce the development of this field, summarize and compare some classical researches and algorithm theories, and look forward to the future of optical neural network.

High-Speed Multi-Layer Convolutional Neural Network Based on Free-Space Optics

“…+T camera + T transf er_data (15) where, 2L-Latency represents latency of the two optical layers. T source represents modulation delay of the input images; and so, considering SLMs with 1 kHz switching frequency [7], T source equals 1ms.…”

“…To date, photonic platform and optical neural networks (ONNs) are an appropriate solution to overcome various drawbacks of electrical implementation [9], [10]. Ultrabroad bandwidth, high interconnection and inherent parallelism capability are the key advantages of optical processing technology [11]- [14] which offers task execution at the speed of light [15]. Recently, some new interests have been activated in developing ONN, and new ONNs have been proposed for implementing photonic multilayer perceptrons (MLPs) [10], [16]- [18] as well as photonic CNNs [6], [7], [19]- [21].…”

Metasurface-enabled on-chip multiplexed diffractive neural networks in the visible

“…Recently, optical neural networks (ONNs) 14 – 20 have attracted much attention due to their high speed, high parallelism, and low energy consumption compared with neural networks running by electrons. As a kind of ONNs, the all-optical diffractive neural networks have been proposed and experimentally demonstrated by constructing 3D printing diffractive surfaces to form a physical network 21 at terahertz wavelengths and achieve specific functions 22 – 26 .…”

Curved Photodetectors Based on Perovskite Microwire Arrays via In Situ Conformal Nanoimprinting

“…As the human body, various objects, and environment contain diverse complex curved surfaces, curved devices can be efficiently incorporated into the human body and environment to expand their capabilities. [5][6][7][8][9][10] In the medical field, the implantation of curved optoelectronic devices on complex surfaces of the skin or artificial organs is important for biological information compared with those of conventional materials. [37][38][39][40][41][42][43][44] For example, Zhang et al fabricated top-down etched microdisks and high-performance micro-lasers based on the CH 3 NH 3 PbBr 3 (MAPbBr 3 ) perovskite.…”

“…As the human body, various objects, and environment contain diverse complex curved surfaces, curved devices can be efficiently incorporated into the human body and environment to expand their capabilities. [ 5–10 ] In the medical field, the implantation of curved optoelectronic devices on complex surfaces of the skin or artificial organs is important for biological information monitoring. [ 11–13 ] In aerospace engineering, conformal antenna arrays and sensors attached to the irregular surface of an aircraft can effectively reduce the space required for antenna installation, weight of the aircraft, air resistance, and radar scattering area.…”