Abstract:Abstract Photonic neural networks are brain-inspired information processing technology using photons instead of electrons to perform artificial intelligence (AI) tasks. However, existing architectures are designed for a single task but fail to multiplex different tasks in parallel within a single monolithic system due to the task competition that deteriorates the model performance. This paper proposes a novel optical multitask learning system by designing multiwavelength diffractive deep neural networks (D2NNs) with the joint optimization method. By encoding multitask inputs into multiwavelength channels, the system can increase the computing throughput and significantly alleviate the competition to perform multiple tasks in parallel with high accuracy. We design the two-task and four-task D2NNs with two and four spectral channels, respectively, for classifying different inputs from MNIST, FMNIST, KMNIST, and EMNIST databases. The numerical evaluations demonstrate that, under the same network size, multiwavelength D2NNs achieve significantly higher classification accuracies for multitask learning than single-wavelength D2NNs. Furthermore, by increasing the network size, the multiwavelength D2NNs for simultaneously performing multiple tasks achieve comparable classification accuracies with respect to the individual training of multiple single-wavelength D2NNs to perform tasks separately. Our work paves the way for developing the wavelength-division multiplexing technology to achieve high-throughput neuromorphic photonic computing and more general AI systems to perform multiple tasks in parallel.

Learning to see through multimode fibers

Multimode optical fiber transmission with a deep learning network

Learning to transmit images through optical speckle of a multimode fiber with high fidelity

Spatially Variant Deblur and Image Enhancement in a Single Multimode Fiber Imaged by Deep Learning

Image transmission through a flexible multimode fiber by deep learning

Reconstructing Images Through Multimode Fibers From the Up-Conversion Speckle Patterns via Deep Learning

Neural Network-Based Multimode Fiber Imaging and Characterization Under Thermal Perturbations

Fiber-Optic Shape Sensing Using Neural Networks Operating on Multispecklegrams

Intensity-only Mode Decomposition on Multimode Fibers using a Densely Connected Convolutional Network

Controlling spatiotemporal nonlinearities in multimode fibers with deep neural networks

Exploiting holographically encoded variance to transmit labelled images through a multimode optical fiber

Single-ended Recovery of Optical fiber Transmission Matrices using Neural Networks

Imaging through multimode fibres with physical prior

Single-shot enhanced imaging through multimode fiber based on self-learning networks

Towards optimal multimode fiber imaging by leveraging input polarization and deep learning

Light Propagation Prediction through Multimode Optical Fibers with a Deep Neural Network

Deep learning sheds new light on non-orthogonal optical multiplexing

Learning-enabled recovering scattered data from twisted light transmitted through a long standard multimode fiber

Optical multi-task learning using multi-wavelength diffractive deep neural networks

Multimode fiber image reconstruction based on parallel neural network with small training set under wide temperature variations

Self-supervised dynamic learning for long-term high-fidelity image transmission through unstabilized diffusive media