REDES NEURONALES

VolverVolver

Resultados 86 resultados LastUpdate Última actualización 28/11/2022 [20:55:00] pdf PDF xls XLS

Solicitudes publicadas en los últimos 30 días / Applications published in the last 30 days



Página1 de 4 nextPage   por página


DATA PACKET CLASSIFICATION METHOD AND SYSTEM BASED ON CONVOLUTIONAL NEURAL NETWORK

NºPublicación: US2022374733A1 24/11/2022

Solicitante:

INSTITUTE OF COMPUTING TECH CHINESE ACADEMY OF SCIENCES [CN]

WO_2021088234_A1

Resumen de: US2022374733A1

The disclosure provides a data packet classification method and system based on a convolutional neural network including merging each rule set in a training rule set to form a plurality of merging schemes, and determining an optimal merging scheme for each rule set in the training rule set on the basis of performance evaluation; converting a prefix combination distribution of each rule set in the training rule set and a target rule set into an image, and training a convolutional neural network model by taking the image and the corresponding optimal merging scheme as features; and classifying the target rule set on the basis of image similarity, and constructing a corresponding hash table for data packet classification.

traducir

SYSTEMS AND METHODS FOR ADAPTIVE TRAINING NEURAL NETWORKS

NºPublicación: US2022374715A1 24/11/2022

Solicitante:

DEEP LABS INC [US]

US_11182675_B1

Resumen de: US2022374715A1

The present disclosure relates to systems and methods for creating and training neural networks. The method includes collecting a set of signals from a database; applying a transform to each signal to create a modified set of signals, wherein signals of the modified set of signals are wavelets; iteratively, for each of a subset of the modified signals: training the neural network using a modified signal of the subset by adding at least one node to the neural network in response to an error function of an analysis of the modified signal exceeding a threshold; removing nodes from the neural network with activation rates below an activation rate threshold; and grouping each node into a lobe among a plurality of lobes, wherein nodes belonging to a lobe have a common characteristic.

traducir

AUTOMATED SYSTEM AND METHOD OF MONITORING ANATOMICAL STRUCTURES

NºPublicación: US2022370031A1 24/11/2022

Solicitante:

NGEE ANN POLYTECHNIC [SG]
SINGAPORE HEALTH SERVICES PTE LTD [SG]
KUMAMOTO UNIV [JP]

WO_2021054901_A1

Resumen de: US2022370031A1

Embodiments include a patch-type, ultrasound sensor system and method to monitor the function and motion of a patients anatomical structure, comprising processing at least one received ultrasound image using one or more analytical tools, including radon transformation, higher-order spectra techniques, and/or active contour models, to generate at least one processed ultrasound image; inputting the at least one processed ultrasound image into a deep learning Convolutional Neural Network to obtain an automatic classification result selected from two or more classes indicating the functional state of the anatomical structure. The patch-type, ultrasound sensor system can communicate via a wireless or wired connection. The monitoring can be at rest or during surgery or other procedure or whilst the subject is exposed to any physiological stressors as part of medical examinations, and can be adapted for use in monitoring the function of body structures including the heart, blood vessels, lungs or joints.

traducir

MULTI-ROBOT TRAJECTORY PLANNING METHOD

NºPublicación: WO2022241808A1 24/11/2022

Solicitante:

GUANGZHOU INSTITUTE OF ADVANCED TECH CHINESE ACADEMY OF SCIENCES [CN]

CN_113326872_A

Resumen de: WO2022241808A1

Disclosed in the present invention is a multi-robot trajectory planning method. The method comprises the following steps: obtaining a current trajectory vector by means of analysis during deep Q-learning and by using a state of a multi-robot surrounding environment, designing a reward network of deep Q-learning, and taking both the current trajectory vector and a desired trajectory vector as inputs of the reward network, and an output of the reward network as reward information, and training parameters of a convolutional neural network (CNN) by using the inputs and the reward information; taking the current trajectory vector as an input of the CNN, and the CNN, which has been trained on the basis of the reward information, outputting corresponding action information to environment information by using a CNN algorithm; and then rationally allocating all actions related to a workpiece to multiple robots by using a resource-based multi-robot task allocation algorithm, such that the multiple robots can cooperate with each other without interfering with each other, thereby implementing spatial three-dimensional complex trajectory planning for multiple robots, and thus achieving the high efficiency of the robots cooperatively executing a complex task.

traducir

SYSTEMS AND METHODS FOR ADAPTIVE TRAINING NEURAL NETWORKS

NºPublicación: WO2022245949A1 24/11/2022

Solicitante:

DEEP LABS INC [US]

US_2022374715_PA

Resumen de: WO2022245949A1

The present disclosure relates to systems and methods for creating and training neural networks. The method includes collecting a set of signals from a database; applying a transform to each signal to create a modified set of signals, wherein signals of the modified set of signals are wavelets; iteratively, for each of a subset of the modified signals: training the neural network using a modified signal of the subset by adding at least one node to the neural network in response to an error function of an analysis of the modified signal exceeding a threshold; removing nodes from the neural network with activation rates below an activation rate threshold; and grouping each node into a lobe among a plurality of lobes, wherein nodes belonging to a lobe have a common characteristic.

traducir

SIGNING AND AUTHENTICATION OF DIGITAL IMAGES AND OTHER DATA ARRAYS

NºPublicación: US2022374660A1 24/11/2022

Solicitante:

INT BUSINESS MACHINES CORPORATION [US]

Resumen de: US2022374660A1

Computer-implemented methods and systems are provided for digitally signing predetermined arrays of digital data. Such a method may provide a secret neural network model trained to classify arrays of digital data in dependence on data content of the arrays. The array of the arrays may be signed by supplying the array to the secret neural network model to obtain an initial classification result; and effecting a modification of data in the array to change the initial classification result to a predetermined, secret classification result, the modification being effected via a backpropagation process in the secret neural network model to progressively modify the array in response to backpropagated errors dependent on a difference between a current classification result for the array and the secret classification result.

traducir

ANOMALY DETECTION FOR DEEP NEURAL NETWORKS

NºPublicación: US2022374657A1 24/11/2022

Solicitante:

FORD GLOBAL TECH LLC [US]

CN_115293321_PA

Resumen de: US2022374657A1

An image including a first object can be input to a deep neural network trained to detect objects. The deep neural network can output a first feature vector corresponding to the first object. A first distance can be measured from the first feature vector to a feature vector subspace determined using a k-means single value decomposition algorithm on an overcomplete dictionary of feature vectors. The first object can be determined to correspond to an anomaly based on the first distance.

traducir

LANGUAGE-GUIDED DISTRIBUTIONAL TREE SEARCH

NºPublicación: US2022374723A1 24/11/2022

Solicitante:

NVIDIA CORP [US]

Resumen de: US2022374723A1

Apparatuses, systems, and techniques to perform a language-guided distributional tree search based at least in part on a natural language task. In at least one embodiment, a tree search is performed using one or more neural networks to determine an action to be performed by an autonomous agent.

traducir

HARDWARE-AWARE NEURAL NETWORK DESIGN

NºPublicación: WO2022245238A1 24/11/2022

Solicitante:

HUAWEI TECH CO LTD [CN]
LETUNOVSKIY ALEXEY ALEKSANDROVICH [CN]

Resumen de: WO2022245238A1

The present disclosure relates to improvements of search for architectures suitable for implementation on certain hardware. Design of a search space is improved by determining a search space of architectures with one or more blocks. The determining of the search space is based on a measure including an amount of matrix operations and/or an amount of layer input and/or output data or an amount of vector operations. The searching for the one or more architectures in the determined search space includes a particular scaling which is based on the measured latency of the architecture blocks and predetermined criteria relating to latency.

traducir

Systems and Methods for Contrastive Learning of Visual Representations

NºPublicación: US2022374658A1 24/11/2022

Solicitante:

GOOGLE LLC [US]

US_2021319266_A1

Resumen de: US2022374658A1

Systems, methods, and computer program products for performing semi-supervised contrastive learning of visual representations are provided. For example, the present disclosure provides systems and methods that leverage particular data augmentation schemes and a learnable nonlinear transformation between the representation and the contrastive loss to provide improved visual representations. Further, the present disclosure also provides improvements for semi-supervised contrastive learning. For example, computer-implemented method may include performing semi-supervised contrastive learning based on a set of one or more unlabeled training data, generating an image classification model based on a portion of a plurality of layers in a projection head neural network used in performing the contrastive learning, performing fine-tuning of the image classification model based on a set of one or more labeled training data, and after performing the fine-tuning, distilling the image classification model to a student model comprising a relatively smaller number of parameters than the image classification model.

traducir

FACE SWAPPING WITH NEURAL NETWORK-BASED GEOMETRY REFINING

NºPublicación: US2022374649A1 24/11/2022

Solicitante:

DISNEY ENTPR INC [US]
ETH ZURICH EIDGENOSSISCHE TECHNISCHE HOCHSCHULE ZURICH [CH]

Resumen de: US2022374649A1

Various embodiments set forth systems and techniques for changing a face within an image. The techniques include receiving a first image including a face associated with a first facial identity; generating, via a machine learning model, at least a first texture map and a first position map based on the first image; rendering a second image including a face associated with a second facial identity based on the first texture map and the first position map, wherein the second facial identity is different from the first facial identity.

traducir

NEURAL NETWORKS FOR COARSE- AND FINE-OBJECT CLASSIFICATIONS

NºPublicación: US2022374650A1 24/11/2022

Solicitante:

WAYMO LLC [US]

US_11361187_B1

Resumen de: US2022374650A1

Aspects of the subject matter disclosed herein include methods, systems, and other techniques for training, in a first phase, an object classifier neural network with a first set of training data, the first set of training data including a first plurality of training examples, each training example in the first set of training data being labeled with a coarse-object classification; and training, in a second phase after completion of the first phase, the object classifier neural network with a second set of training data, the second set of training data including a second plurality of training examples, each training example in the second set of training data being labeled with a fine-object classification.

traducir

REAL TIME ENHANCEMENT FOR STREAMING CONTENT

NºPublicación: US2022374714A1 24/11/2022

Solicitante:

NVIDIA CORP [US]

CN_115376035_PA

Resumen de: US2022374714A1

Real time content enhancement can be provided using a solution that is lightweight enough to operate on client devices, even for high resolution, high bitrate content. An enhancement process can include a neural network that upscales the content to a target resolution while also enhancing a visual quality of the content, such as to sharpen visual aspects of the content and reduce a presence of artifacts. Such an approach can enable compressed content to be transmitted in streams across a network, in order to conserve bandwidth and data transmission, while also enabling that content to be upscaled and enhanced at the client device in real time, such that a user or viewer can experience the content at, near, or above its intended or original visual quality.

traducir

REVERSE IMAGE SEARCH BASED ON DEEP NEURAL NETWORK (DNN) MODEL AND IMAGE-FEATURE DETECTION MODEL

NºPublicación: US2022374647A1 24/11/2022

Solicitante:

SONY GROUP CORP [JP]

Resumen de: US2022374647A1

An electronic device and method for reverse image search is provided. The electronic device receives an image. The electronic device extracts, by a DNN model, a first set of image features associated with the image and generates a first feature vector based on the first set of image features. The electronic device extracts, by an image-feature detection model, a second set of image features associated with the image and generates a second feature vector based on the second set of image features. The electronic device generates a third feature vector based on combination of the first and second feature vectors. The electronic device determines a similarity metric between the third feature vector and a fourth feature vector of each of a set of pre-stored images and identifies a pre-stored image based on the similarity metric. The electronic device controls a display device to display information associated with the pre-stored image.

traducir

SELECTING POINTS IN CONTINUOUS SPACES USING NEURAL NETWORKS

NºPublicación: US2022374683A1 24/11/2022

Solicitante:

DEEPMIND TECH LIMITED [GB]

Resumen de: US2022374683A1

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for selecting an optimal feature point in a continuous domain for a group of agents. A computer-implemented system obtains, for each of a plurality of agents, respective training data that comprises a respective utility score for each of a plurality of discrete points in the continuous domain. The system trains, for each of the plurality of agents and on the respective training data for the agents, a respective neural network that is configured to receive an input comprising a point in the continuous domain and to generate as output a predicted utility score for the agent at the point. And the system identifies the optimal point by optimizing an approximation of the shared outcome function that is defined by, for any given point in the continuous domain, a combination of the predicted utility scores generated by the respective neural networks for each of the plurality of agents by processing an input comprising the given point.

traducir

SYNTHESIZING VIDEO FROM AUDIO USING ONE OR MORE NEURAL NETWORKS

NºPublicación: US2022374637A1 24/11/2022

Solicitante:

NVIDIA CORP [US]

CN_115379287_PA

Resumen de: US2022374637A1

Apparatuses, systems, and techniques are presented to reduce an amount of data to be transmitted for media content. In at least one embodiment, one or more neural networks are used to generate video and audio information corresponding to one or more people based, at least in part, on at least one image and voice information corresponding to the one or more people.

traducir

INTER-DOCUMENT ATTENTION MECHANISM

NºPublicación: US2022374479A1 24/11/2022

Solicitante:

MICROSOFT TECH LICENSING LLC [US]

CN_114450681_PA

Resumen de: US2022374479A1

This document relates to natural language processing using a framework such as a neural network. One example method involves obtaining a first document and a second document and propagating attention from the first document to the second document. The example method also involves producing contextualized semantic representations of individual words in the second document based at least on the propagating. The contextualized semantic representations can provide a basis for performing one or more natural language processing operations.

traducir

METHOD AND APPARATUS FOR CONDITIONING NEURAL NETWORKS

NºPublicación: US2022375462A1 24/11/2022

Solicitante:

SAMSUNG ELECTRONICS CO LTD [KR]

Resumen de: US2022375462A1

Broadly speaking, the present techniques provide methods for conditioning a neural network, which not only improve the generalizable performance of conditional neural networks, but also reduce model size and latency significantly. The resulting conditioned neural network is suitable for on-device deployment due to having a significantly lower model size, lower dynamic memory requirement, and lower latency.

traducir

SYSTEM AND METHOD FOR THE FUSION OF BOTTOM-UP WHOLE-IMAGE FEATURES AND TOP-DOWN ENTTIY CLASSIFICATION FOR ACCURATE IMAGE/VIDEO SCENE CLASSIFICATION

NºPublicación: US2022375222A1 24/11/2022

Solicitante:

HRL LABORATORIES LLC [US]

US_2019005330_A1

Resumen de: US2022375222A1

Described is a system and method for accurate image and/or video scene classification. More specifically, described is a system that makes use of a specialized convolutional-neural network (hereafter CNN) based technique for the fusion of bottom-up whole-image features and top-down entity classification. When the two parallel and independent processing paths are fused, the system provides an accurate classification of the scene as depicted in the image or video.

traducir

DISTRIBUTED PLACEMENT OF LINEAR OPERATORS FOR ACCELERATED DEEP LEARNING

NºPublicación: US2022374288A1 24/11/2022

Solicitante:

CEREBRAS SYSTEMS INC [US]

WO_2021084505_A1

Resumen de: US2022374288A1

Techniques in distributed placement of linear operators for accelerated deep learning provide improvements in one or more of accuracy, performance, and energy efficiency. An array of processing elements comprising a portion of a neural network accelerator performs flow-based computations on wavelets of data. Each processing element comprises a compute element to execute programmed instructions using the data and a router to route the wavelets. The routing is in accordance with virtual channel specifiers of the wavelets and controlled by routing configuration information of the router. A software stack determines distributed placement of linear operators based on a description of a neural network. The determined placement is used to configure the routers including usage of the respective colors. The determined placement is used to configure the compute elements including the respective programmed instructions each is configured to execute.

traducir

METHOD AND SYSTEM OF PERFORMING CONVOLUTION IN NEURAL NETWORKS WITH VARIABLE DILATION RATE

NºPublicación: US2022374651A1 24/11/2022

Solicitante:

SAMSUNG ELECTRONICS CO LTD [KR]

KR_20200084808_A

Resumen de: US2022374651A1

A method of performing convolution in a neural network with variable dilation rate is provided. The method includes receiving a size of a first kernel and a dilation rate, determining at least one of size of one or more disintegrated kernels based on the size of the first kernel, a baseline architecture of a memory and the dilation rate, determining an address of one or more blocks of an input image based on the dilation rate, and one or more parameters associated with a size of the input image and the memory. Thereafter, the one or more blocks of the input image and the one or more disintegrated kernels are fetched from the memory, and an output image is obtained based on convolution of each of the one or more disintegrated kernels and the one or more blocks of the input image.

traducir

INFORMATION-AWARE GRAPH CONTRASTIVE LEARNING

NºPublicación: WO2022245491A1 24/11/2022

Solicitante:

NEC LABORATORIES AMERICA INC [US]

Resumen de: WO2022245491A1

A method for performing contrastive learning for graph tasks and datasets by employing an information-aware graph contrastive learning framework is presented. The method includes obtaining (1001) two semantically similar views of a graph coupled with a label for training by employing a view augmentation component, feeding (1003) the two semantically similar views into respective encoder networks to extract latent representations preserving both structure and attribute information in the two views, optimizing (1005) a contrastive loss based on a contrastive mode by maximizing feature consistency between the latent representations, training (1007) a neural network with the optimized contrastive loss, and predicting (1009) a new graph label or a new node label in the graph.

traducir

ROBOT FLEET MANAGEMENT METHOD AND SYSTEM USING A GRAPH NEURAL NETWORK

NºPublicación: WO2022242966A1 24/11/2022

Solicitante:

CONTINENTAL AUTOMOTIVE TECH GMBH [DE]

GB_2606752_PA

Resumen de: WO2022242966A1

In order to improve scalability and manage an arbitrary number of route constraints, the invention provides a robot management system (10) that utilizes a graph neural network system (52). The graph neural network system (52) has a main graph auto encoder network (54), which deals with the system state as a whole and further comprises an auxiliary graph auto encoder network (56) for each route constraint of the robot management system (10), in order to generate commands (50) for the robots (12). The robots (12) move about an environment (14) based on the single-step commands (50).

traducir

GESTURE CLASSIFICATION AND RECOGNITION METHOD AND APPLICATION THEREOF

NºPublicación: WO2022242133A1 24/11/2022

Solicitante:

SHENZHEN INSTITUTES OF ADVANCED TECH CHINESE ACADEMY OF SCIENCES [CN]

CN_113312994_A

Resumen de: WO2022242133A1

A gesture classification and recognition method, relating to the technical field of data classification. The method comprises: obtaining a surface electromyography (sEMG) signal; performing feature extraction on the sEMG signal to obtain a gesture feature sequence and a gesture type; and inputting the gesture feature sequence and the gesture type into a recurrent gate circuit neural network for training to obtain a classification model, and using the classification model to implement gesture classification and recognition. By means of the method, the problems that existing sEMG signal-based gesture classification and recognition algorithms are low in recognition accuracy and have overfitting and underfitting, vanishing gradient, poor robustness, and long training time during model training processes are solved, and the accuracy of predictive classification is improved.

traducir

Using neural networks to perform object detection, instance segmentation, and semantic correspondence from bounding box supervision

Nº publicación: GB2606816A 23/11/2022

Solicitante:

NVIDIA CORP [US]

CN_114972742_PA

Resumen de: GB2606816A

A processor comprises circuits used to run one or more neural networks 106 to segment images 102, 104 using at least partially, bounding boxes. The neural networks may be supervised or unsupervised and may classify one or more objects 108. Instance segmentation 110 and semantic correspondence 112 may be carried out. The neural networks may be trained. The bounding boxes may be annotated. Pixel masks and bounding box stripes can be used. Different neural network versions with updatable parameters may be employed. The networks may be initialised with the same set of parameters. Hungarian algorithms may be used for object correspondence and segmentation maps may be generated. One of the neural networks can comprise one or more conditional random field (CRF) operations. Optimal transport loss and stochastic gradient descent (SGD) operations may be considered. The images may be obtained from an autonomous vehicle.

traducir

Página1 de 4 nextPage por página

punteroimgVolver