Abstract: Methods, systems, apparatuses, and computer program products are provided that are configured to perform localization of position data, specifically using a trained localization neural network. In the context of an apparatus, the apparatus is caused to receive observed feature representation data. The apparatus is further configured to transform the observed feature representation data into standardized feature representation data utilizing a trained localization neural network. The apparatus is further configured to compare the standardized feature representation data and the map feature representation data and identify local position data based on the comparison.

Abstract: Embodiments of the present invention provide systems, methods, and computer storage media for detecting and classifying an exposure defect in an image using neural networks trained via a limited amount of labeled training images. An image may be applied to a first neural network to determine whether the images includes an exposure defect. Detected defective image may be applied to a second neural network to determine an exposure defect classification for the image. The exposure defect classification can includes severe underexposure, medium underexposure, mild underexposure, mild overexposure, medium overexposure, severe overexposure, and/or the like. The image may be presented to a user along with the exposure defect classification.

Abstract: A computer-implemented method for training a random matrix network is presented. The method includes initializing a random matrix, inputting a plurality of first vectors into the random matrix, and outputting a plurality of second vectors from the random matrix to be fed back into the random matrix for training. The random matrix can include a plurality of two-terminal devices or a plurality of three-terminal devices or a film-based device.

Abstract: A system for monitoring a signal on an optical fiber includes a fiber optic connector having a housing couplable to a receptacle. An optical fiber that transmits a first optical signal has first fiber core at least partially surrounded by a cladding and has a first end terminating proximate the housing. The first optical signal is transmitted along the first fiber core. An optical tap has a first tap waveguide arranged and is configured to receive at least part of the first optical signal as a first tap signal. The first tap waveguide comprises an output port for the first tap signal for directing the tap signal to a detector unit. In other embodiments, a detector unit detects light from the optical signal that is propagating along the fiber cladding.

Abstract: Provided are a method for detecting display screen quality, an apparatus, an electronic device and a storage medium. The method includes: receiving a quality detection request sent by a console deployed on a display screen production line, the quality detection request including a display screen image collected by an image collecting device on the display screen production line; inputting the display screen image into a defect detection model to obtain a defect detection result, the defect detection model being obtained by training historical defective display screen images using a structure of deep convolutional neural networks and an object detection algorithm; and determining, according to the defect detection result, a defect on a display screen corresponding to the display screen image, a defect category corresponding to the defect, and a position corresponding to the defect.

Abstract: A method for jointly training a classification model and a confidence model. The method includes receiving a training data set including a plurality of training data subsets. From two or more training data subsets in the training data set, the method includes selecting a support set of training examples and a query set of training examples. The method includes determining, using the classification model, a centroid value for each respective class. For each training example in the query set of training examples, the method includes generating, using the classification model, a query encoding, determining a class distance measure, determining a ground-truth distance, and updating parameters of the classification model. For each training example in the query set of training examples identified as being misclassified, the method further includes generating a standard deviation value, sampling a new query, and updating parameters of the confidence model based on the new query encoding.

Abstract: A method for a calibration of at least one laser diode, in particular at least one laser diode of a laser projection device. The at least one laser diode is calibrated on the basis of a comparison of at least one currently acquired characteristic value of the at least one laser diode with at least one characteristic value, stored in at least one database, of a model laser diode that is at least substantially identical in construction to the at least one laser diode.

Abstract: Techniques for training neural networks are provided. According to one set of embodiments, a first array is processed in a spreading component to produce a second array, where a first dimension of the first array corresponds to at least one sequence of approximately orthogonal numeric vectors representing tokens, and where the spreading component combines values along the first dimension. The second array is processed in a transformer neural network to determine correlations between the sequence, which produces a third array. One or more batches of the third array are processed in a de-spreading component to produce a fourth array.

Abstract: A method and hardware system for mapping an input map of a convolutional neural network layer to an output map are disclosed. An array of processing elements are interconnected to support unidirectional dataflows through the array along at least three different spatial directions. Each processing element is adapted to combine values of dataflows along different spatial directions into a new value for at least one of the supported dataflows. For each data entry in the output map, a plurality of products from pairs of weights of a selected convolution kernel and selected data entries in the input map is provided and arranged into a plurality of associated partial sums. Products associated with a same partial sum are accumulated on the array and accumulated on the array into at least one data entry in the output map.

Abstract: A system for using free-space optics to interconnect a plurality of computing nodes can include a plurality of optical transceivers that facilitate free-space optical communications among the plurality of computing nodes. The system may ensure a line of sight between the plurality of computing nodes and the optical transceivers to facilitate the free-space optical communications. The line of sight may be preserved by the position or placement of the computing nodes in the system. The position or placement of the computing nodes may be achieved by using different shaped enclosures for holding the computing nodes.

Abstract: A digital mobile fronthaul (MFH) network includes a baseband processing unit (BBU) having a digitization interface configured to digitize, using delta-sigma digitization, at least one wireless service for at least one radio access technology. The network further includes a transport medium in operable communication with the BBU. The transport medium is configured to transmit a delta-sigma digitized wireless service from the BBU. The network further includes a remote radio head (RRH) configured to operably receive the delta-sigma digitized wireless service from the BBU over the transport medium.

Abstract: The techniques disclosed herein improve the efficiency of a system by providing intelligent agents for managing data associated with objects that are displayed within mixed-reality and virtual-reality collaboration environments. Individual agents are configured to collect, analyze, and store data associated with individual objects in a shared view. The agents can identify real-world objects and virtual objects discussed in a meeting, collect information about each object and store the collected information in an associated database for access across multiple collaboration environments or communication sessions. The data can be shared between different communication sessions without requiring users to manually store and present a collection of content for each object. The intelligent agents and their associated databases can also persist through different communication sessions to enhance user engagement and improve productivity.

Abstract: The present disclosure discloses a system and a method. In an example implantation, the system and the method can receive an image at a first deep neural network, estimate a distance between an object depicted in the image and a vehicle, wherein the first deep neural network estimates the distance, determine whether the estimated distance is greater than a predetermined distance threshold, and generate an alert when the estimated distance is not greater than the predetermined distance threshold.

Abstract: Techniques for transmitting an optical signal through optical fiber with an improved cost effective stimulated Brillouin scattering (SBS) suppression include externally modulating a light beam emitted from a light source with a high frequency signal. The light beam is also modulated externally with an RF information-carrying signal. The high frequency signals are at least twice a highest frequency of the RF signal. The high frequency signals modulating the light source can be gain and phase adjusted by the first set of gain and phase control circuit to achieve a targeted spectrum shape. The adjusted high frequency signals then are split, providing a portion of the split signals to modulate the light source and another portion of the split signals to the second set of phase and gain control circuit for adjusting a phase/gain.

Abstract: Techniques are disclosed relating to determining whether document objects included in an image correspond to known document types. In some embodiments, a computing system maintains information specifying a set of known document types. In some embodiments, the computing system receives an image that includes objects. In some embodiments, the computing system analyzes, using a first neural network, the image to identify a document object and location information specifying a location of the document object within the image. In some embodiments, the computing system determines, using a second neural network, whether the document object within the image corresponds to a document type specified in the set of known document types, where the determining is performed based on the location information of the document object. In some embodiments, disclosed techniques may assist in automatically extracting information from documents, which in turn may advantageously decrease processing time for onboarding new customers.

Abstract: A video scene detection machine learning model is provided. A computer device receives feature vectors corresponding to audio and video components of a video. The computing device provides the feature vectors as input to a trained neural network. The computing device receives from the trained neural network, a plurality of output feature vectors that correspond to shots of the video. The computing device applies optimal sequence grouping to the output feature vectors. The computing device further trains the trained neural network based, at least in part, on the applied optimal sequence grouping.

Abstract: A programmable data storage device configured to process images via an embedded processor is disclosed. The processor identifies luminance and chrominance data of a received image, and retrieves a first machine learning model stored in the storage device. The first model is applied for making a first prediction about the image based on luminance data, and a first determination is made in regards to a criterion. In response to making the first determination, a first label associated with the first prediction is returned. A second determination is also made in regards to the criterion. In response to making the second determination, a second machine learning model stored in the storage device is retrieved. The second machine learning model is applied for making a second prediction about the image based on the color data associated with the image, and a second label associated with the second prediction is determined.

Abstract: An artificial intelligence system for analyzing imagery, the system comprising a computing device, the computing device designed and configured to receive a plurality of photographs related to a human subject; analyze the plurality of photographs to identify a conditional indicator contained within the plurality of photographs; generate a classification algorithm utilizing the conditional indicator, wherein the classification algorithm utilizes the conditional indicator as an input and outputs a conditional profile; and determine a conditional status of the human subject utilizing the conditional profile.

Abstract: The disclosure notably relates to a computer-implemented method of machine-learning. The method includes obtaining a dataset including 3D modeled objects which each represent a respective mechanical part. The dataset has one or more sub-datasets. Each sub-dataset forms at least a part of the dataset. The method further includes, for each respective sub-dataset, determining a base template and learning a neural network configured for inference of deformations of the base template each into a respective 3D modeled object. The base template is a 3D modeled object which represents a centroid of the 3D modeled objects of the sub-dataset. The learning includes a training based on the sub-dataset. This constitutes an improved method of machine-learning with a dataset including 3D modeled objects which each represent a respective mechanical part.

Abstract: A method for performing diagnostics of a structure subject to loads, in particular an aircraft structure, is described, said method being implemented by means of an arrangement of sensors located at relevant points of the structure and corresponding neural networks, and comprising: training the neural network in order to establish an associative relationship between the local displacement of the structure in a subset of relevant points and the local displacement of the structure in at least one residual relevant point; detecting the local displacement of the structure in a plurality of relevant points under operating conditions; estimating the local displacement of the structure in at least one residual relevant point by means of the associated neural network on the basis of the pre-established associated relationship; and comparing the local displacement of the estimated structure with the detected local displacement at the residual relevant point.