scholarly journals Digital optical processing of optical communications: towards an Optical Turing Machine

Nanophotonics ◽  
2017 ◽  
Vol 6 (3) ◽  
pp. 507-530 ◽  
Author(s):  
Joe Touch ◽  
Yinwen Cao ◽  
Morteza Ziyadi ◽  
Ahmed Almaiman ◽  
Amirhossein Mohajerin-Ariaei ◽  
...  
Keyword(s):  
Computational Models ◽  
Optical Computing ◽  
Data Representation ◽  
Optical Processing ◽  
Hop Count ◽  
Optical Wave ◽  
Optical Computation ◽  
Combine Phase ◽  
Permanent Storage ◽  
Power Restoration

AbstractOptical computing is needed to support Tb/s in-network processing in a way that unifies communication and computation using a single data representation that supports in-transit network packet processing, security, and big data filtering. Support for optical computation of this sort requires leveraging the native properties of optical wave mixing to enable computation and switching for programmability. As a consequence, data must be encoded digitally as phase (M-PSK), semantics-preserving regeneration is the key to high-order computation, and data processing at Tb/s rates requires mixing. Experiments have demonstrated viable approaches to phase squeezing and power restoration. This work led our team to develop the first serial, optical Internet hop-count decrement, and to design and simulate optical circuits for calculating the Internet checksum and multiplexing Internet packets. The current exploration focuses on limited-lookback computational models to reduce the need for permanent storage and hybrid nanophotonic circuits that combine phase-aligned comb sources, non-linear mixing, and switching on the same substrate to avoid the macroscopic effects that hamper benchtop prototypes.

Acousto-optic data representation method of cosmic systems telemetrical information for coherent optical processing

1994 ◽  
Author(s):  
J. Bykovsky ◽  
Elbrus N. Eloev ◽  
K. Kuhorenko ◽  
Alexander M. Panin ◽  
N. Solodovnikov
Keyword(s):  
Data Representation ◽  
Optical Processing ◽  
Representation Method ◽  
Coherent Optical Processing

scholarly journals Optical Image Encryption Using a Nonlinear Joint Transform Correlator and the Collins Diffraction Transform

Photonics ◽  
2019 ◽  
Vol 6 (4) ◽  
pp. 115 ◽  
Author(s):  
Juan M. Vilardy O. ◽  
Ronal A. Perez ◽  
Cesar O. Torres M.
Keyword(s):  
Fourier Transforms ◽  
Lateral Displacement ◽  
Random Phase ◽  
Design Parameters ◽  
Invariance Property ◽  
Joint Transform Correlator ◽  
Optical Processing ◽  
Optical Wave ◽  
Special Cases ◽  
Nonlinear Joint

The Collins diffraction transform (CDT) describes the optical wave diffraction from the generic paraxial optical system. The CDT has as special cases the diffraction domains given by the Fourier, Fresnel and fractional Fourier transforms. In this paper, we propose to describe the optical double random phase encoding (DRPE) using a nonlinear joint transform correlator (JTC) and the CDT. This new description of the nonlinear JTC-based encryption system using the CDT covers several optical processing domains, such as Fourier, Fresnel, fractional Fourier, extended fractional Fourier and Gyrator domains, among others. The maximum number of independent design parameters or new security keys of the proposed encryption system using the CDT increases three times in comparison with the same encryption system that uses the Fourier transform. The proposed encryption system using the CDT preserves the shift-invariance property of the JTC-based encryption system in the Fourier domain, with respect to the lateral displacement of both the key random mask in the decryption process and the retrieval of the primary image. The viability of this encryption system is verified and analysed by numerical simulations.

scholarly journals AI-assisted on-chip nanophotonic convolver based on silicon metasurface

Nanophotonics ◽  
2020 ◽  
Vol 9 (10) ◽  
pp. 3315-3322 ◽  
Author(s):  
Kun Liao ◽  
Tianyi Gan ◽  
Xiaoyong Hu ◽  
Qihuang Gong
Keyword(s):  
Optical Computing ◽  
Processing System ◽  
Characteristic Size ◽  
Optical Signal ◽  
Mathematical Operation ◽  
Optical Processing ◽  
Convolution Operation ◽  
All Optical ◽  
On Chip ◽  
Further Development

AbstractConvolution operation is of great significance in on-chip all-optical signal processing, especially in signal analysis and image processing. It is a basic and important mathematical operation in the realization of all-optical computing. Here, we propose and experimentally implement a dispersionless metalens for dual wavelengths, a 4f optical processing system, and then demonstrate the on-chip nanophotonic convolver based on silicon metasurface with the optimization assistance of inverse design. The characteristic size of the dispersionless metalens device is 8 × 9.4 μm, and the focusing efficiency is up to 79% and 85% at wavelengths of 1000 and 1550 nm, respectively. The feature size of the convolver is 24 × 9.4 μm, and the proposed convolver allows spatial convolution operation on any desired function at dual wavelengths simultaneously. This work provides a potential scheme for the further development of on-chip all-optical computing.

scholarly journals VoroCNN: Deep convolutional neural network built on 3D Voronoi tessellation of protein structures

2021 ◽  
Author(s):  
Ilia Igashov ◽  
liment Olechnovič ◽  
Maria Kadukova ◽  
Česlovas Venclovas ◽  
Sergei Grudinin
Keyword(s):  
Neural Network ◽  
Convolutional Neural Network ◽  
Computational Models ◽  
Voronoi Tessellation ◽  
Protein Structures ◽  
Data Representation ◽  
Deep Convolutional Neural Network ◽  
Molecular Structures ◽  
Supplementary Information ◽  
Evolutionary Information

Abstract Motivation Effective use of evolutionary information has recently led to tremendous progress in computational prediction of three-dimensional (3D) structures of proteins and their complexes. Despite the progress, the accuracy of predicted structures tends to vary considerably from case to case. Since the utility of computational models depends on their accuracy, reliable estimates of deviation between predicted and native structures are of utmost importance. Results For the first time, we present a deep convolutional neural network (CNN) constructed on a Voronoi tessellation of 3D molecular structures. Despite the irregular data domain, our data representation allows us to efficiently introduce both convolution and pooling operations and train the network in an end-to-end fashion without precomputed descriptors. The resultant model, VoroCNN, predicts local qualities of 3D protein folds. The prediction results are competitive to state of the art and superior to the previous 3D CNN architectures built for the same task. We also discuss practical applications of VoroCNN, for example, in recognition of protein binding interfaces. Availability The model, data, and evaluation tests are available at https://team.inria.fr/nano-d/software/vorocnn/. Supplementary information Supplementary data are available at Bioinformatics online.

scholarly journals Deep Learning Machine using Hierarchical Cluster Features

2019 ◽  
Vol 29 (3) ◽  
pp. 82
Author(s):  
Sara Salman ◽  
Jamila H. Soud
Keyword(s):  
Deep Learning ◽  
Computational Models ◽  
Hierarchical Cluster ◽  
Data Representation ◽  
Cluster Method ◽  
Classification Problems ◽  
Levels Of Abstraction ◽  
Image Translation ◽  
Multiple Levels ◽  
Learning Machine

Deep learning of multi-layer computational models allowed processing to recognize data representation at multiple levels of abstraction. These techniques have greatly improved the latest ear recognition technology. PNN is a type of radiative basis for classification problems and is based on the Bayes decision-making base, which reduces the expected error of classification. In this paper, strong features of images are used to give a good result, therefore, SIFT method using these features after adding improvements and developments. This method was one of the powerful algorithms in matching that needed to find energy pixels. This method gives stronger feature on features and gives a large number of a strong pixel, which is considered a center and neglected the remainder of it in our work. Each pixel of which is constant for image translation, scaling, rotation, and embedded lighting changes in lighting or 3D projection. Therefore, the interpretation is developed by using a hierarchical cluster method; to assign a set of properties (find the approximation between pixels) were classified into one.

scholarly journals Optical differentiator based on a trilayer metal-dielectric structure

2021 ◽  
Vol 45 (3) ◽  
pp. 356-363
Author(s):  
A.I. Kashapov ◽  
L.L. Doskolovich ◽  
D.A. Bykov ◽  
E.A. Bezus ◽  
D.V. Nesterenko
Keyword(s):  
Metal Layer ◽  
Optical Computing ◽  
Metal Substrate ◽  
Dielectric Structure ◽  
Multiple Wave ◽  
First Derivative ◽  
Wave Interference ◽  
Optical Computation ◽  
Optical Differentiator ◽  
Simulation Results

Optical properties of a resonant trilayer metal-dielectric-metal (MDM) structure that consists of an upper metal layer, a dielectric layer, and a metal substrate are investigated. Using a multiple wave interference model, we prove that the reflection coefficient of the MDM structure may strictly vanish. The existence of a reflectance zero makes it possible to use the MDM structure as an optical differentiator. The numerical simulation results presented demonstrate the possibility of optical computation of the first derivative with respect to either time or spatial variable. The obtained results may find application in novel analog optical computing and optical information processing systems.

Applying Fluid Dynamics Simulations to Improve Processing and Remediation of Nuclear Waste

2011 ◽  
Author(s):  
Kelly J. Knight ◽  
Jon Berkoe ◽  
Brigette Rosendall ◽  
Joel Peltier ◽  
Chris Kennedy
Keyword(s):  
Fluid Dynamics ◽  
Nuclear Waste ◽  
Computational Models ◽  
Fluid Dynamic ◽  
Engineering Analysis ◽  
Complex Processes ◽  
Performance Requirements ◽  
Permanent Storage ◽  
Dynamics Simulations ◽  
Liquid Flows

Transport and processing of nuclear waste for treatment and storage can involve unique and complex thermal and fluid dynamic conditions that pose potential for safety risk and/or design uncertainty and also are likely to be subjected to more precise performance requirements than in other industries. From an engineering analysis perspective, certainty of outcome is essential. Advanced robust methods for engineering analysis and simulation of critical processes can help reduce risk of design uncertainty and help mitigate or reduce the amount of expensive full-scale demonstration testing. This paper will discuss experience gained in applying computational fluid dynamics models to key processes for mixing, transporting, and thermal treatment of nuclear waste as part of designing a massive vitrification process plant that will convert high and low level nuclear waste into glass for permanent storage. Examples from industrial scale simulations will be presented. The computational models have shown promise in replicating several complex physical processes such as solid-liquid flows in suspension, blending of slurries, and cooling of materials at extremely high temperature. Knowledge gained from applying simulation has provided detailed insight into determining the most critical aspects of these complex processes that can ultimately be used to help guide the optimum design of waste handling equipment based on credible calculations while ensuring risk of design uncertainty is minimized.

scholarly journals VoroCNN: Deep convolutional neural network built on 3D Voronoi tessellation of protein structures

2020 ◽  
Author(s):  
Ilia Igashov ◽  
Kliment Olechnovic ◽  
Maria Kadukova ◽  
Česlovas Venclovas ◽  
Sergei Grudinin
Keyword(s):  
Neural Network ◽  
Convolutional Neural Network ◽  
Computational Models ◽  
Voronoi Tessellation ◽  
Protein Structures ◽  
Computational Prediction ◽  
Data Representation ◽  
Deep Convolutional Neural Network ◽  
Molecular Structures ◽  
Evolutionary Information

MotivationEffective use of evolutionary information has recently led to tremendous progress in computational prediction of three-dimensional (3D) structures of proteins and their complexes. Despite the progress, the accuracy of predicted structures tends to vary considerably from case to case. Since the utility of computational models depends on their accuracy, reliable estimates of deviation between predicted and native structures are of utmost importance.ResultsFor the first time we present a deep convolutional neural network (CNN) constructed on a Voronoi tessellation of 3D molecular structures. Despite the irregular data domain, our data representation allows to efficiently introduce both convolution and pooling operations of the network. We trained our model, called VoroCNN, to predict local qualities of 3D protein folds. The prediction results are competitive to the state of the art and superior to the previous 3D CNN architectures built for the same task. We also discuss practical applications of VoroCNN, for example, in the recognition of protein binding interfaces.AvailabilityThe model, data, and evaluation tests are available at https://team.inria.fr/nano-d/software/vorocnn/[email protected], [email protected]

The Theoretical Contrast in Scanning and Conventional High Resolution Microscopes

1969 ◽  
Vol 27 ◽  
pp. 170-171
Author(s):  
E. Zeitler ◽  
M. G. R. Thomson
Keyword(s):  
Small Volume ◽  
Volume Element ◽  
Optical Processing ◽  
Image Construction ◽  
Object Interaction ◽  
Transmission Electron ◽  
Scanning Transmission ◽  
Optical Treatment ◽  
Electron Microscopes ◽  
Intensity Contrast

In the formation of an image each small volume element of the object is correlated to an areal element in the image. The structure or detail of the object is represented by changes in intensity from element to element, and this variation of intensity (contrast) is determined by the interaction of the electrons with the specimen, and by the optical processing of the information-carrying electrons. Both conventional and scanning transmission electron microscopes form images which may be considered in this way, but the mechanism of image construction is very different in the two cases. Although the electron-object interaction is the same, the optical treatment differs.

Sign in / Sign up

Export Citation Format

Share Document