|
DIFFRACTIVE OPTICS, OPTICAL TECHNOLOGIES
On chip optical neural networks based on MMI microring resonators for image classification
T. T. Buia, D. T. Leb, T. H. Nguyenc, T. T. Leb a FPT University Hoa Lac High Tech Park, 100000, Hanoi, Viet Nam
b International School (VNU-IS), Vietnam National University (VNU)-Hanoi, 100000, Hanoi, Viet Nam, Xuan Thuy St. 144
c Hanoi University of Natural Resources and Environment, 100000, Hanoi, Viet Nam, Phu Dien St. 41A
Abstract:
We propose a new on-chip optical neural network (OONN) based on multimode interference-microring resonators (MMI-RRs). The suggested structure eliminates the need for wave-length division multiplexers (WDM) to create an optical neuron on a single chip. New microring resonator structure based on $4\times4$ MMI coupler with a size of $24~\mu m\times2900~\mu m$ is used for the basic elements of the computation matrix, as a result a higher bandwidth and free spectral range (FSR) can be achieved. The Si3N4 platform along with the graphene sheet is designed to modulate the signals and weights of the neural networks at a very high speed. The Si3N4 can provide wide range of operating wavelengths and can work directly with the wavelengths of color images. The structure's benefits include rapid computing speed, little loss, and the ability to handle both positive and negative values. The OONN has been applied to the MNIST dataset with a speed faster than 2.8 to 14x times compared with the conventional GPU methods.
Keywords:
all-optical dot product, image processing, multimode interference coupler, optical convolutional neural networks, optical signal processing, microring resonators, silicon photonics
Received: 24.08.2022 Accepted: 30.03.2023
Citation:
T. T. Bui, D. T. Le, T. H. Nguyen, T. T. Le, “On chip optical neural networks based on MMI microring resonators for image classification”, Computer Optics, 47:4 (2023), 588–595
Linking options:
https://www.mathnet.ru/eng/co1159 https://www.mathnet.ru/eng/co/v47/i4/p588
|
Statistics & downloads: |
Abstract page: | 8 | Full-text PDF : | 4 | References: | 2 |
|