AI Article Synopsis
Article Abstract
In this study, we propose low power consumption, programmable on-chip optical nonlinear units (ONUs) for all-optical neural networks (all-ONNs). The proposed units were constructed using a III-V semiconductor membrane laser, and the nonlinearity of the laser was used as the activation function of a rectified linear unit (ReLU). By measuring the relationship of the output power and input light, we succeeded in obtaining the response as an activation function of the ReLU with low power consumption. With its low-power operation and high compatibility with silicon photonics, we believe that this is a very promising device for realizing the ReLU function in optical circuits.
Download full-text PDF |
Source |
|---|---|
| http://dx.doi.org/10.1364/OL.471603 | DOI Listing |
Publication Analysis
Top Keywords
Similar Publications
All dielectric metasurface based diffractive neural networks for 1-bit adder.
Nanophotonics
April 2024
Beijing Key Laboratory of Metamaterials and Devices, Key Laboratory of Terahertz Optoelectronics, Ministry of Education, Beijing Advanced Innovation Center for Imaging Theory and Technology, Department of Physics, Capital Normal University, Beijing, 100048, China.
Diffractive deep neural networks ( ) have brought significant changes in many fields, motivating the development of diverse optical computing components. However, a crucial downside in the optical computing components is employing diffractive optical elements (DOEs) which were fabricated using commercial 3D printers. DOEs simultaneously suffer from the challenges posed by high-order diffraction and low spatial utilization since the size of individual neuron is comparable to the wavelength scale.
View Article and Find Full Text PDFLeveraging multiplexed metasurfaces for multi-task learning with all-optical diffractive processors.
Nanophotonics
November 2024
Department of Electrical and Computer Engineering, North Carolina State University, Raleigh, 27695, USA.
Diffractive Neural Networks (DNNs) leverage the power of light to enhance computational performance in machine learning, offering a pathway to high-speed, low-energy, and large-scale neural information processing. However, most existing DNN architectures are optimized for single tasks and thus lack the flexibility required for the simultaneous execution of multiple tasks within a unified artificial intelligence platform. In this work, we utilize the polarization and wavelength degrees of freedom of light to achieve optical multi-task identification using the MNIST, FMNIST, and KMNIST datasets.
View Article and Find Full Text PDFAnti-Interference All-Optical Logic Computing Based on a 2D Polarization-Sensitive Photodiode.
Nano Lett
January 2025
State Key Laboratory of Superlattices and Microstructures, Institute of Semiconductors, Chinese Academy of Sciences, Beijing 100083, China.
Optical logic operation is promising for ultrafast information processing and optical computing due to the high computation speed and low power consumption. However, conventional optical logic devices require either a complex structure and circuit design or a constant voltage supply, which impedes the development of high-density integrated circuits. Here, all-optical logic devices are designed using a self-powered polarization-sensitive photodiode of the GeSe homojunction, which is attributed to an anisotropic band structure and built-in electric field.
View Article and Find Full Text PDFOpto-intelligence spectrometer using diffractive neural networks.
Nanophotonics
August 2024
Department of Electronic Engineering, Tsinghua University, Beijing 100084, China.
Spectral reconstruction, critical for understanding sample composition, is extensively applied in fields like remote sensing, geology, and medical imaging. However, existing spectral reconstruction methods require bulky equipment or complex electronic reconstruction algorithms, which limit the system's performance and applications. This paper presents a novel flexible all-optical opto-intelligence spectrometer, termed OIS, using a diffractive neural network for high-precision spectral reconstruction, featuring low energy consumption and light-speed processing.
View Article and Find Full Text PDFMetasurface-Based Image Classification Using Diffractive Deep Neural Network.
Nanomaterials (Basel)
November 2024
International School of Microelectronics, Dongguan University of Technology, Dongguan 523808, China.
The computer-assisted inverse design of photonic computing, especially by leveraging artificial intelligence algorithms, offers great convenience to accelerate the speed of development and improve calculation accuracy. However, traditional thickness-based modulation methods are hindered by large volume and difficult fabrication process, making it hard to meet the data-driven requirements of flexible light modulation. Here, we propose a diffractive deep neural network (DNN) framework based on a three-layer all-dielectric phased transmitarray as hidden layers, which can perform the classification of handwritten digits.
View Article and Find Full Text PDFWant AI Summaries of new PubMed Abstracts delivered to your In-box?
Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!