Redes Neuronales

LIGHT ESTIMATION USING NEURAL NETWORKS

Resumen de: US20260360076A1

A messaging system performs image processing to estimate lighting properties with neural networks for images provided by users of the messaging system. A method of estimating light properties includes receiving an input image with first lighting properties and processing the input image using a convolutional neural network to generate an estimate of the first lighting properties. The method may further include modifying the input image with an augmentation to generate a modified input image, where the augmentation has second lighting properties, and changing the second lighting properties of the augmentation in the modified input image to the estimate of the first lighting properties.

SYSTEMS AND METHODS FOR COMPUTING FEATURING SYNTHETIC COMPUTING OPERATORS AND COLLABORATION

Resumen de: AU2025261949A1

One embodiment is directed to a synthetic engagement system for process-based problem solving, comprising: a computing system comprising one or more operatively coupled computing resources; and a user interface operated by the computing system and configured to engage a human operator in accordance with a predetermined process configuration toward an established requirement based at least in part upon one or more specific facts; wherein the user interface is configured to allow the human operator to select and interactively engage one or more synthetic operators operated by the computing system to proceed through the predetermined process configuration, and to return result to the human operator selected to at least partially satisfy the established requirement; and wherein each of the one or more synthetic operators is informed by a convolutional neural network informed at least in part by historical actions of a particular actual human operator and a synthetic operator background configuration.

METHODS AND SYSTEMS FOR WORD EDIT DISTANCE EMBEDDING

Resumen de: US20260360355A1

A system for classifying words in a batch of words can include at least one memory device storing instructions for causing at least one processor to create dictionary vectors for each of a plurality of dictionary words using a neural network (NN), store each dictionary vector along with a classification indicator corresponding to the associated dictionary word, and create word vectors for each word in a batch of words for classification using the NN. The closest matching dictionary vectors are found for each word vector and the classification indicators of the closest matching dictionary vector for each word vector in the batch is reported.

Low-Rank Adaptation of Neural Network Models

Resumen de: US20260359642A1

A computer implemented method obtains neural network-based model base model weight matrices for each of multiple neural network layers. First low-rank factorization matrices are added to corresponding base model weight matrices to form a first domain model. The low-rank factorization matrices are treated as trainable parameters. The first domain model is trained with first domain specific training data without modifying base model weight matrices.

NEURAL NETWORK MODEL AND LEARNING METHOD OF THE SAME

Resumen de: US20260359603A1

A neural network model that can perform highly accurate processing on input data is provided. The neural network model includes first and second neutral networks, and the first neural network includes a first layer, a second layer, and a third layer. A feature map output from the first layer is input to the second layer and the second neural network, and a feature map output from the second neural network is input to the third layer. Given that the feature map output from the first layer when first data is input to the first neural network is a correct feature map and that the feature map output from the first layer when second data obtained by adding noise to the first data is input to the first neural network is a learning feature map, the second neural network is learned so that a feature map output from the second neural network matches the correct feature map when the learning feature map is input.

DISTRIBUTED PLACEMENT OF LINEAR OPERATORS FOR ACCELERATED DEEP LEARNING

Resumen de: US20260359145A1

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.

CHANNELWISE FEATURE REORIENTATION AND MERGING WITH NEIGHBOURING FEATURES FOR ROBUST LOW COMPLEXITY DNN BASED AUDIO PROCESSING

Resumen de: AU2025266039A1

There is disclosed an apparatus (10) for processing an input information signal (20), comprising: * a feature extractor (100) for extracting a set of features (102) from the input information signal (20), the set of features (102) having a first shape, * a feature block generator (200, CR1) configured to reshape the set of features (102) from a first shape to a second shape by: - segmenting (210) the set of features (102) into at least 3 feature subsets (212, 212a-212h), each feature subset (212, 212a-212h) having a specific shape, - performing a subsampling (220, 220a, 220b) of the feature subsets (212) to generate a plurality of feature blocks (222, 222a-222d) by stacking a selected set of the subsampled feature subsets feature subsets (212, 212a-212h) along a first dimension, - creating (230) a feature hyper-block (332) of the second shape by stacking the plurality of feature blocks (222, 222a-222d) along a second dimension, * a neural network processor (300, D1 ) configured to process the feature hyper block (332) to obtain a first result (380); and * an output post-processor (P2, 700) configured to post-process the result (380) to construct an output information signal (40) and/or to output a classification result.

Detection System for Alzheimer's Disease Using Brain Structural and/or Functional Magnetic Resonance Image Processing and Machine Learning Techniques

Resumen de: US20260360542A1

Early and accurate identification of dementia, such as Alzheimer's disease (AD) is of utmost importance. However, despite several decades of biomarker research, there is still no single diagnostic test or biomarker to definitively diagnose dementia and its underlying neurodegenerative disorder. A combination of structural MRI (fMRI) and functional MRI (fMRI) derived image analysis may provide the unique ability to capture the dynamic state of change in the degenerating brain. Hence, to capture the overall structural and functional anomalies of brain tissues caused by AD, an exhaustive combinatorial system has been developed using structural and/or functional MRI data followed by rigorous image processing and deep learning-based algorithms to diagnose AD and/or mild cognitive impairment (MCI) patients. Whole brain sMRI and fMRI slices are processed, and brain pixel-based intensity features are fed into a deep learning-based convoluted neural network (CNN) algorithm to calculate the AD/healthy probability of each MRI slice.

CONTENT-AWARE STYLE ENCODING USING NEURAL NETWORKS

Resumen de: US20260360037A1

Apparatuses, systems, and techniques to facilitate application of a style, for which one or more neural networks have not been trained by a training framework, from one image to content of another image. In at least one embodiment, a styled output image is generated by one or more neural networks based on a style contained in a style image and content of a content image where said one or more neural networks have not been trained by a training framework on said style.

SCANNING ELECTRON MICROSCOPE IMAGE PROCESSING USING NEURAL NETWORK

Resumen de: US20260360103A1

A neural network device includes: (1) a pre-processor configured to select target images from scanning electron microscope (SEM) images, based on frequencies respectively corresponding to the SEM images, and crop each of the target images into a plurality of cropped images; (2) a neural network processor configured to generate a crop detection image by inferring a target object from each of the plurality of cropped images by using a segmentation model trained to detect the target object in each of the plurality of cropped images; and (3) a post-processor configured to merge crop detection images with each other in a same size as the SEM images, based on position information of the plurality of cropped images.

MULTI-TASKING ACTION UNIT PREDICTIONS

Resumen de: US20260360183A1

In an example in accordance with the present disclosure, a multi-tasking learning expression tracking system is described. The system includes a feature extractor to extract a feature representation from an image of a user. The system also includes a classification branch having a classification neural network and the feature extractor to, during training, predict expression classes for training images having different expressions. The system also includes a regression branch comprising a regression neural network and the feature extractor to, during training, predict action unit (AU) intensities for the training images and during deployment predict an AU intensity for the image of the user.

METHOD AND APPARATUS FOR INTELLIGENT DATA PLANE BASED ON NEURAL NETWORK

Resumen de: US20260359599A1

A method for an intelligent data plane based on a neural network, includes obtaining a switch classification result by deploying a recurrent neural network on a P4 programmable switch to perform line-speed analysis on traffic; forwarding a first part of traffic to a server for upgraded analysis based on the switch classification result and a classification confidence; and performing the upgraded analysis on the first part of traffic by deploying a Transformer-based model on the server.

SENSOR FUSION

Resumen de: US20260360137A1

In various examples, sensor fusion for autonomous or semi-autonomous systems and applications is described. Systems and methods are disclosed that use data pipelines to process sensor data generated using different types of sensors (e.g., image sensors, RADAR sensors, LiDAR sensors, etc.) in order to generate first data representing information associated with objects surrounding a vehicle. For instance, the information may represent values for parameters associated with the objects, such as locations of the objects, dimensions of the objects, velocities of the objects, orientations of the objects, classifications of the objects, and/or any of parameters. The systems and methods may then process the first data using one or more machine learning models (e.g., one or more deep neural networks) that are trained to fuse the information (e.g., the parameters) and output second data representing final information associated with the objects. The fused output may then be used to perform downstream operations.

MULTICORE-BASED DEEP LEARNING ACCELERATOR AND METHOD FOR OPTIMIZING CORE THEREOF

Resumen de: US20260359620A1

A multicore-based deep learning accelerator includes a neural processing unit including a plurality of cores configured to enable parallel processing of a neural network model and a storage unit storing a first binary configured to cause the plurality of cores to perform inference based on the neural network model and a second binary configured to cause the remaining cores, excluding a portion of the plurality of cores, to perform inference based on the neural network model, wherein the neural processing unit is configured to perform inference based on the neural network model using a core combination configured based on a binary selected from the first binary and the second binary.

METHOD FOR ASSIGNING A SEMANTIC CONCEPT TO A PREVIOUSLY UNSEEN DATA ASSET

Resumen de: EP4838381A1

0001 A computer-implemented method assigns a semantic concept to a new data asset by: enriching (1) the data assets metadata I (tokenizing labels (A1), expanding acronyms (A2), leveraging (A3) a knowledge graph (11)/databases (14), generating descriptions (A4)); retrieving (2) candidate concepts and similarity scores via, e.g., vector database (13) searches comparing enriched metadata embeddings with concept descriptor embeddings; determining (3) a first ranking based on similarity scores; determining (4) at least a second ranking using a neural network (12b) processing the first ranking and enriched metadata; and storing a proposed mapping (O) based on at least the second ranking in a knowledge graph (11).

A METHOD AND APPARATUS FOR ENCODING OR DECODING A PICTURE USING A NEURAL NETWORK

Resumen de: EP4838416A2

0001 Disclosed herein are methods and systems for encoding a picture and decoding a bitstream that may represent an encoded picture. During encoding and decoding, rescaling operations are applied to rescale an input to a size that can be processed by a layer of a neural network. Embodiments disclosed herein provide methods for rescaling that achieve a reduced size of the bitstream, thereby improving compression.

COMPILATION METHOD AND APPARATUS FOR NEURAL NETWORK MODEL, INFERENCE METHOD AND APPARATUS FOR NEURAL NETWORK MODEL, AND DEVICE AND MEDIUM

Resumen de: US20260350344A1

Disclosed are a compilation and inference method for a neural network model, and a device and a medium. The compilation method comprises: acquiring a first model corresponding to a neural network model to be subjected to compilation; using, as a target network layer, any one of at least one network layer which includes a preset operation in the first model, and converting the preset operation of the target network layer into a convolution operation which is supported by a neural network processor, so as to obtain a converted first network layer, wherein the preset operation is an operation comprising a multiply—accumulate operation; on the basis of the converted first network layer, obtaining an optimization model corresponding to the first model; and according to the optimization model, generating a target instruction sequence that is executable by the neural network processor and corresponds to said neural network model.

MULTIRESOLUTION HASH ENCODING FOR NEURAL NETWORKS

Resumen de: US20260350393A1

0000 Neural network performance is improved in terms of training speed and/or accuracy by encoding (mapping) inputs to the neural network into a higher dimensional space via a hash function. The input comprises coordinates used to identify a point within a d-dimensional space (e.g., 3D space). The point is quantized and a set of vertex coordinates corresponding to the point are input to a hash function. For example, for d=3, space may be partitioned into axis-aligned voxels of identical size and vertex coordinates of a voxel containing the point are input to the hash function to produce a set of encoded coordinates. The set of encoded coordinates is used to lookup D-dimensional feature vectors in a table of size T that have been learned. The learned feature vectors are filtered (e.g., linearly interpolated, etc.) based on the coordinates of the point to compute a feature vector corresponding to the point.

INVARIANT STRUCTURE CHANNEL MAPPING

Resumen de: WO2026232172A2

A method for utilizing a Structural Receptive Set (SRS) structure in a trained neural network model includes accessing learned weight tensors of a trained neural network layer comprising input channels and output channels. Semantic Core (SC) channels are identified from the output channels based on the learned weight tensors. For an SC channel, a Structure Auxiliary (AC) channel set is identified based on structural similarity among output-channel weight tensors. An SRS comprising a subset of the input channels structurally used to compute an SC/AC output group is determined for the SC channel and the AC channel set. An Invariant Structure Channel Mapping (ISCM) group comprising the SC channel, the AC channel set, and the SRS is generated and used to produce a sparse input-output dependency structure that enables the SC/AC output group to be computed from fewer than all input channels.

TRAINING LARGE-SCALE VISION TRANSFORMER NEURAL NETWORKS

Resumen de: US20260350975A1

0000 Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training Vision Transformer (ViT) neural networks.

HUMAN-UNDERSTANDABLE INSIGHTS FOR NEURAL NETWORK PREDICTIONS

Resumen de: US20260350360A1

Techniques for mapping human-understandable attributes to a neural network are disclosed. A system generates human-understandable explanations for decisions or predictions of a neural network by mapping human-understandable attributes to the trained neural network. The system may map a human-understandable attribute to a particular label, indicating a correlation between the particular human-understandable attribute and the particular label. In addition, or in the alternative, the system may map the human-understandable attribute to a particular layer of the neural network, such as a particular hidden layer, indicating a correlation between the human-understandable attribute and the particular layer.

DETECTION OF SAFETY STATES DURING MEDICAL IMAGING PROCEDURES

Resumen de: US20260350783A1

0000 Disclosed herein is a medical system (100, 300). The execution of machine executable instructions (120) causes a computational system to (104): receive (200) an overview image (128), receive (202) object locators (130) in response to inputting the overview image into a first neural network (122), construct (204) an object sub image (132) using the object locator, receive (206) object positional data (134) in response to inputting the object sub image for each of the multiple predetermined objects into an object specific neural network (124), construct (208) multiple search images (136) using the object positional data for the multiple predetermined objects, receive (210) locator data (138) descriptive of a position of an item in response to inputting each of the multiple search images into an item specific neural network (126), detect (212) a safety state (140) by comparing the locator data and the object positional data of each of the predetermined objects with respect to a predetermined criterion; and provide (214) a warning signal (144) if the safety state is detected

METHOD AND DEVICE FOR TRAINING AN OBJECT DETECTOR WITH AN ENSEMBLE OF SEVERAL DETECTION HEADS

Resumen de: US20260350394A1

0000 A computer-implemented method for training a neural network. The neural network detects objects in images. The training includes: ascertaining a feature representation of a training image using a backbone of the neural network; ascertaining head outputs based on the feature representation and using heads of the neural network, wherein the head outputs, for regions of the training image, each characterize a presence or absence of an object or characterize an extension of the object or characterize a position of the object or characterize the extension and the position; ascertaining first loss values, wherein each first loss value corresponds to a head output and characterizes a deviation of the head output from a desired head output; adjusting a parameter(s) of the backbone and/or at least one parameter of a head based on a second loss value, wherein the second loss value characterizes a sum of the first loss values.

METHODS AND SYSTEMS FOR PERFORMING ZOOM OPERATIONS ON COMPUTING DEVICE

Resumen de: US20260350739A1

Systems and methods for performing a zoom operation on a display of a computing device, which include identifying a region of interest (ROI) of an image presented on the display based on a zoom user input, determining a scaling factor for a final zoomed presentation of the ROI based on an image rendering requirement and/or an inference speed of a frame generation convolutional neural network (CNN), using the CNN to generate the final zoomed presentation of the ROI to the determined scaling factor, down-scaling the final zoomed presentation to generate a sequence of intermediate size image frames each having dimensions smaller than dimensions of the final zoomed presentation of the ROI and greater than dimensions of the ROI, and rendering an animation of changes in a displayed size of the ROI using the sequence of intermediate image frames of the ROI.

VISUAL PROPERTY DETERMINATIONS USING ONE OR MORE NEURAL NETWORKS

Nº publicación: US20260350770A1 05/11/2026

Solicitante:

NVIDIA CORP [US]
NVIDIA Corporation

Resumen de: US20260350770A1

Apparatuses, systems, and techniques are presented to determine visual properties of an object. In at least one embodiment, shape and visual properties of an object can be determined using differently illuminated images fed to a pipeline of neural networks.