Related references
Note: Only part of the references are listed.
Article
Optics
George Sarantoglou et al.
Summary: In this study, experimental results show that a quantum-dot laser can be tuned to operate as either a leaky integrate and fire or resonant and fire neuron by adjusting the bias. The multiband emission of quantum-dot devices enhances neurocomputational capabilities, leading to increased spike firing rate and suppressed neural spike duration. These new operation regimes, combined with thermal insensitivity and silicon cointegration, make these neuromorphic nodes a promising platform for large-scale photonic spiking neural networks.
PHOTONICS RESEARCH
(2021)
Article
Optics
Siming Zheng et al.
Summary: The study introduces a plug-and-play method using deep-learning denoisers as regularization priors for spectral compressive imaging, demonstrating efficiency and flexibility. The method shows excellent reconstruction results under various compressive coding mechanisms and can be applied to multiple spectral SCI systems.
PHOTONICS RESEARCH
(2021)
Article
Optics
Chengshuai Yang et al.
Summary: A new hybrid algorithm (AL + DL) was developed for high-fidelity image reconstruction in Compressed Ultrafast Photography (CUP), improving accuracy compared to conventional methods. This study validates superior performance of the AL + DL algorithm in image fidelity over traditional algorithms, promoting applications of CUP in related fields.
PHOTONICS RESEARCH
(2021)
Article
Optics
Jianhui Ma et al.
Summary: This work aims to develop an automatic scatter estimation framework for efficient photon distribution estimation, utilizing GPU-based MC simulations and an over-relaxation algorithm combined with deep reinforcement learning. The proposed method can rapidly and accurately estimate scatter signals with a large range of photon numbers, achieving superior performance in terms of structural similarity, peak signal-to-noise ratio, and relative absolute error.
PHOTONICS RESEARCH
(2021)
Article
Optics
Yunqi Luo et al.
Summary: This study proposed a deep learning-empowered adaptive optical focusing framework, achieved real-time light focusing and refocusing through Timely-Focusing-Optical-Transformation-Net (TFOTNet), effectively handling spatiotemporal non-stationarity of the medium.
PHOTONICS RESEARCH
(2021)
Article
Optics
Xianxin Guo et al.
Summary: The study presents a practical scheme for end-to-end optical backpropagation in neural networks, using saturable absorption for nonlinear units and approximating backward-propagating gradients in a pump-probe scheme. The approach achieves equivalent performance to computational networks on image classification benchmarks, even in deep networks with multiple sequential gradient approximation, providing a feasible path toward truly all-optical neural networks.
PHOTONICS RESEARCH
(2021)
Article
Optics
Zhenyu Li et al.
Summary: The smart sensor achieves label-free multicomponent chemical analysis with a high prediction accuracy and simple detection strategy using a label-free ring resonator and neural network model, promising great potential for applications in various fields.
PHOTONICS RESEARCH
(2021)
Article
Optics
Qiang Cai et al.
Summary: The study introduces a simple approach based on photonic reservoir computing for modulation format identification in optical fiber communications. This technique can efficiently identify different modulation formats with an accuracy of over 95% in various transmission situations by optimizing control parameters of the P-RC layer. By utilizing simple devices, this method offers a resource-efficient alternative for MFI.
PHOTONICS RESEARCH
(2021)
Article
Optics
Xingyu Wang et al.
Summary: This paper presents an adaptive compensation method using deep learning technology to improve the performance of OAM-encoded QKD schemes and designs a QKD scheme with real-time phase correction capability to ensure information security.
PHOTONICS RESEARCH
(2021)
Review
Optics
Yihao Xu et al.
Summary: Photonics has greatly influenced research and applications, with artificial neural networks providing a potential solution to simplify complex design processes and enhance efficiency and accuracy.
PHOTONICS RESEARCH
(2021)
Article
Optics
Yanan Han et al.
Summary: A modified supervised learning algorithm for optical spiking neural networks is proposed in this study, which introduces synaptic time-delay plasticity to improve performance. The proposed algorithm outperforms traditional weight-based methods in tasks of spike sequence learning and classification.
PHOTONICS RESEARCH
(2021)
Article
Optics
Sunae So et al.
Summary: This study utilized deep learning to design spectrally sensitive multiband absorbers in the visible range, proposing a five-layered metal-insulator-metal grating structure designed with an ANN and validating the optical properties of the design. The research demonstrates that the trained ANN can learn physical knowledge and suggests a method to reduce the size of the ANN.
PHOTONICS RESEARCH
(2021)
Article
Optics
Shuo Zhu et al.
Summary: This paper proposes a physics-informed learning method for imaging through unknown scattering media, achieving high reconstruction fidelity for sparse objects and solving the limited generalization capability issue of deep learning methods.
PHOTONICS RESEARCH
(2021)
Article
Optics
Xiao Wang et al.
Summary: A deep-learning method is proposed to recognize OAM modes under misalignment using an alignment-free fractal multipoint interferometer. The experiment achieved 98.35% recognizing accuracy, showing a stable solution for OAM mode recognition in free-space optical communication.
PHOTONICS RESEARCH
(2021)
Article
Optics
Che Liu et al.
Summary: Intelligent coding metasurface is capable of manipulating electromagnetic waves and associating digital information simultaneously, showing promising applications. The proposed physics-driven variational auto-encoder cGAN, based on unsupervised conditional generative adversarial networks, offers an efficient approach to design intelligent coding metasurface holograms.
PHOTONICS RESEARCH
(2021)
Article
Optics
Albert Ryou et al.
Summary: This study introduces a free-space optical ANN with linear weight summation and nonlinear activation enabled by saturable absorption of thermal atoms, achieving successful image classification of handwritten digits with a significant improvement in classification accuracy compared to linear models.
PHOTONICS RESEARCH
(2021)
Article
Optics
Peng Dai et al.
Summary: This study demonstrates the reverse design of F-P cavity structure using deep learning, which enables the production of a wide color space coverage over 215% of sRGB with high accuracy. Compared to evolutionary algorithms, neural networks show significant improvement in design efficiency, highlighting the potential of deep learning in achieving extremely accurate design of nanostructured color filters.
PHOTONICS RESEARCH
(2021)
Review
Optics
Peter R. Wiecha et al.
Summary: Deep learning in nano-photonics is mostly discussed in terms of its potential for inverse design of photonic devices or nano-structures. While specific, drawbacks of the approaches require critical evaluation. Additionally, deep learning has evolved as a tool for a variety of applications such as accelerating photonics simulations, data reconstruction, imaging, and knowledge discovery.
PHOTONICS RESEARCH
(2021)
Article
Optics
Ronghui Lin et al.
Summary: This study demonstrates a neural network capable of designing multiple symmetry-protected bound states in freeform structures with predefined symmetry, where the latent representation allows for continuous tuning of geometry and high design accuracy for controlling multiple BIC frequencies. Rich band inversion and accidental degeneracy in these freeform structures are shown by direct interaction with the latent representation.
PHOTONICS RESEARCH
(2021)
Article
Optics
Zafran Hussain Shah et al.
Summary: SR-SIM enhances spatial resolution of fluorescently labeled samples, while deep-learning based denoising using RED-Net improves image quality and robustness. The combination of computational reconstruction and denoising shows strong performance even with changes in microscope settings.
PHOTONICS RESEARCH
(2021)
Article
Optics
Jinran Qie et al.
Summary: The connection between Maxwell's equations and artificial neural networks has revolutionized nanophotonic design, providing efficiency and convenience. However, transitioning from conventional design methods to machine-learning tools can be challenging due to a steep learning curve. The web-based design tool with a trained deep neural network offers a user-friendly and efficient alternative to traditional full-wave simulations.
PHOTONICS RESEARCH
(2021)
Article
Optics
Linh Nguyen et al.
Summary: A new approach to optical fiber sensing using deep neural network models allows for specific measurement in the presence of strong noise without the need for additional fiber processing steps. This technique is highly generalizable, capable of identifying any measurand of interest and enabling sensing in noisy environments.
PHOTONICS RESEARCH
(2021)
Article
Optics
Zheng Zhen et al.
Summary: Being invisible has been a dream for centuries, and recent breakthroughs in technology have led to the development of a transmitted metasurface cloak, opening up new possibilities for transparent invisibility.
PHOTONICS RESEARCH
(2021)
Article
Optics
Yahui Zhang et al.
Summary: In this study, all-optical binary convolution using a photonic spiking VCSEL neuron was proposed and experimentally demonstrated, showcasing successful results, robustness with high-resolution images, and advantages such as ultrafast speed and simple hardware implementation. This system shows the potential of spiking photonic VCSEL neurons for high-speed neuromorphic image processing systems and future photonic spiking convolutional neural networks.
PHOTONICS RESEARCH
(2021)
Article
Optics
Shanshan Zheng et al.
Summary: This study proposes a method for single-shot incoherent imaging through highly nonstatic and optically thick turbid media using a deep neural network, achieving high fidelity reconstruction of object images in severe environments with high turbidity.
PHOTONICS RESEARCH
(2021)
Article
Optics
Baurzhan Muminov et al.
Summary: The study demonstrates generalizable image reconstruction using simple neural networks, showing faster speed and robustness, which are favorable for future real-time low power vision systems.
PHOTONICS RESEARCH
(2021)
Article
Optics
Yangming Ren et al.
Summary: This study introduces an efficient method for designing photonic devices, which combines deep neural networks with genetic algorithm, requiring less training data compared to previous methods. Several ultra-compact silicon photonics devices with challenging properties were designed using this method, generally complying with silicon photonics fabrication design rules.
PHOTONICS RESEARCH
(2021)
