Neural networks

TRAINING A NEURAL NETWORK USING A DATA SET WITH LABELS OF MULTIPLE GRANULARITIES

Absstract of: EP1000000A1

The invention relates to an apparatus (1) for manufacturing green bricks from clay for the brick manufacturing industry, comprising a circulating conveyor (3) carrying mould containers combined to mould container parts (4), a reservoir (5) for clay arranged above the mould containers, means for carrying clay out of the reservoir (5) into the mould containers, means (9) for pressing and trimming clay in the mould containers, means (11) for supplying and placing take-off plates for the green bricks (13) and means for discharging green bricks released from the mould containers, characterized in that the apparatus further comprises means (22) for moving the mould container parts (4) filled with green bricks such that a protruding edge is formed on at least one side of the green bricks.

OBJECT IDENTIFICATIONS IN IMAGES OR VIDEOS

Absstract of: US20260087646A1

0000 An apparatus is provided. The apparatus includes a communications interface to receive raw data from an external source. The raw data includes a representation of a first object and a second object. The apparatus further includes a memory storage unit to store the raw data. In addition, the apparatus includes a neural network engine to receive the raw data. The neural network engine is to generate a segmentation map and a boundary map. The apparatus also includes a post-processing engine to identify the first object and the second object based on the segmentation map and the boundary map.

IMPROVED WELLBORE CONTROL AND MODELS USING IMAGE DATA SYSTEMS AND METHODS

Absstract of: US20260085602A1

Computer implemented methods and systems for testing one or more operational changes in a drill rig includes initiating the one or more operational changes and using, in part, image data of a mechanical mud separation machines (“MMSM”) to detect the impact of the one or more changes. The image data may be processed by a Deep Neural Network to identify objects in the object flow, operational parameters of the MMSM, and wellbore environmental conditions. Additional image data may be selected for additional processing based on the results of the analysis. The results of the test may be used to update the drilling operation or a drilling model.

Neural Networks with Semantic Inference

Absstract of: US20260087304A1

0000 Systems and method of classification are provided. Upon receiving an input, a feature set is defined from the input. A semantic cluster to be associated with the input is defined based on the feature set, the semantic cluster being one of a plurality of semantic clusters each defining a subset of outputs of a neural network based on semantic similarity of the subset. The feature set is applied to a subgraph corresponding to the semantic cluster, the subgraph being one of a plurality of subgraphs each defining a portion of the neural network. A classification for the input is then be determined based on an output of the subgraph.

MACHINE LEARNING TECHNIQUES FOR GENERATING CONTROLLER LOGIC

Absstract of: US20260086524A1

Embodiments of the present disclosure relate to generating controller logic. Indication of a controller logic generation request associated with an asset identifier may be received. A prompt template set associated with a controller logic generation workflow may be identified based on the asset identifier. The prompt template of the prompt template set may comprise one or more instruction sets. The prompt template set may be input into a large language model comprising one or more transformer neural networks and configured to generate a controller logic configuration file for the asset identifier based on the prompt template set and intent classification associated with each prompt template. The controller logic configuration file may be received from the large language model. Performance of one or more prediction-based actions may be initiated based on the controller logic configuration file.

END-TO-END STREAMING KEYWORD SPOTTING

Absstract of: US20260088023A1

A method for training hotword detection includes receiving a training input audio sequence including a sequence of input frames that define a hotword that initiates a wake-up process on a device. The method also includes feeding the training input audio sequence into an encoder and a decoder of a memorized neural network. Each of the encoder and the decoder of the memorized neural network include sequentially-stacked single value decomposition filter (SVDF) layers. The method further includes generating a logit at each of the encoder and the decoder based on the training input audio sequence. For each of the encoder and the decoder, the method includes smoothing each respective logit generated from the training input audio sequence, determining a max pooling loss from a probability distribution based on each respective logit, and optimizing the encoder and the decoder based on all max pooling losses associated with the training input audio sequence.

DEEP LEARNING INTERNAL STATE INDEX-BASED SEARCH AND CLASSIFICATION

Absstract of: US20260088022A1

Systems and methods are disclosed for generating internal state representations of a neural network during processing and using the internal state representations for classification or search. In some embodiments, the internal state representations are generated from the output activation functions of a subset of nodes of the neural network. The internal state representations may be used for classification by training a classification model using internal state representations and corresponding classifications. The internal state representations may be used for search, by producing a search feature from an search input and comparing the search feature with one or more feature representations to find the feature representation with the highest degree of similarity.

DEEP NEURAL NETWORKS (DNN) INFERENCE USING PRACTICAL EARLY EXIT NETWORKS

Absstract of: US20260086912A1

The present disclosure relates to methods and systems for providing inferences using machine learning systems. The methods and systems receive a load forecast for processing requests by a machine learning model and split the machine learning model into a plurality machine learning model portions based on the load forecast. The methods and systems determine a batch size for the requests for the machine learning model portions. The methods and systems use one or more available resources to execute the plurality of machine learning model portions to process the requests and generate inferences for the requests.

Method And Device For Monitoring Port And Ship In Consideration Of Sea Level

Absstract of: US20260087822A1

The present invention relates to a method for monitoring a harbor performed by a computing device, the method for monitoring the harbor according to an aspect of the present invention comprising: obtaining a harbor image having a first view attribute; generating a segmentation image having the first view attribute and corresponding to the harbor image by performing an image segmentation using an artificial neural network trained to output information, from an input image, related to an object included in the input image; generating a transformed segmentation image having a second view attribute from the segmentation image having the first view attribute based on a first view transformation information used to transform an image having the first view attribute into an image having the second view attribute different from the first view attribute; and calculating berthing guide information of the ship based on the transformed segmentation image.

CONTROLLING A FUNCTION VIA GAZE DETECTION

Absstract of: US20260086636A1

0000 Aspects of the present disclosure relate to systems and methods for controlling a function of a computing system using gaze detection. In examples, one or more images of a user are received and gaze information may be determined from the received one or more images. Non-gaze information may be received when the gaze information is determined to satisfy a condition. Accordingly, a function may be enabled based on the received non-gaze information. In examples, the gaze information may be determined by extracting a plurality of features from the received one or more images, providing the plurality of features to a neural network, and determining, utilizing the neural network, a location at a display device at which a gaze of the user is directed.

MEDIA ENGAGEMENT THROUGH DEEP LEARNING

Absstract of: US20260088021A1

Apparatuses, systems, and techniques to facilitate understanding of media content using neural networks to adjust playback speed and volume based on environmental and other factors. In at least one embodiment, playback of media content is slowed down or sped up if audio associated with said media content is difficult to understand based on background noise, accent, difficulty of material, as well as other factors that decrease understandability of media content.

VIRTUAL IMPACTOR-BASED LABEL-FREE PARTICULATE MATTER DETECTION USING HOLOGRAPHY AND DEEP LEARNING

Absstract of: US20260086013A1

A particulate matter detection device takes holographic images of flowing particulate matter concentrated by a virtual impactor, which selectively slows down and guides larger particles to fly through an imaging window. The flowing particles are illuminated by a pulsed laser diode, casting their inline holograms on a CMOS image sensor in a lens-free mobile imaging device. The illumination contains three short pulses with a negligible shift of the flowing particle within one pulse and triplicate holograms of the same particle are recorded at a single frame revealing different perspectives of each particle. A deep neural network classifies the particles based on the acquired holographic images. The device was tested using different types of pollen and achieved a blind classification accuracy of 92.91%. This mobile and cost-effective device weighs ˜700 g and can be used for label-free sensing and quantification of various bio-aerosols over extended periods.

IMAGE COMPRESSION AND DECODING, VIDEO COMPRESSION AND DECODING: METHODS AND SYSTEMS

Absstract of: US20260089329A1

A computer-implemented method for lossy image or video compression, transmission and decoding, the method including the steps of (i) receiving an input image at a first computer system; (ii) encoding the input image using a first trained neural network, using the first computer system, to produce a latent representation; (iii) quantizing the latent representation using the first computer system to produce a quantized latent; (iv) entropy encoding the quantized latent into a bitstream, using the first computer system; (v) transmitting the bitstream to a second computer system; (vi) the second computer system entropy decoding the bitstream to produce the quantized latent; (vii) the second computer system using a second trained neural network to produce an output image from the quantized latent, wherein the output image is an approximation of the input image.

METHOD AND SYSTEM FOR CONTROLLING AND DISTRIBUTING WAVE ENERGY IN OFFSHORE AQUACULTURE

Absstract of: US20260086522A1

0000 Disclosed in the present disclosure is a method and system for controlling and distributing wave energy in offshore aquaculture. The method includes: obtaining an aquaculture cycle of each aquaculture sub-zone of an offshore aquaculture zone, sorting remaining aquaculture cycles of the aquaculture sub-zones from small to large, and obtaining a plurality of work cycles according to sorting results; obtaining a predicted wave energy yield of a next work cycle through a preset neural network model; obtaining an importance coefficient value sorting result of each aquaculture zone through a preset recursive feature elimination (RFE) model; and adjusting operation cycles and operation power of first-type aquaculture apparatuses, second-type aquaculture apparatuses, and third-type aquaculture apparatuses in sequence according to an apparatus type of each aquaculture apparatus, the aquaculture zone where each aquaculture apparatus is located, the predicted wave energy yield, and the importance coefficient value sorting results.

MODEL TRAINING METHOD AND APPARATUS, CONSTRUCTION SAFETY EVALUATION METHOD AND APPARATUS, AND DEVICE

Absstract of: WO2026061185A1

The present application discloses a model training method and apparatus, a construction safety evaluation method and apparatus, and a device. The model training method comprises: acquiring sample data; on the basis of association information between training samples and true value labels, determining label missing ratios of the training samples in different dimensions and first weight values of the training samples in different dimensions; determining second weight values of the training samples on the basis of the label missing ratios; and inputting the sample data, the first weight values, and the second weight values into a preset neural network model for training until a loss value of a target loss function of the preset neural network model meets a model convergence condition, so as to obtain a target prediction model. In this way, by improving a weighted loss function, when sample data having a partially missing label is kept, a loss value of a missing label of a sample is calculated with reference to a weight, thereby solving the problem of reduced prediction accuracy caused by missing samples, improving the accuracy of a model, and simultaneously and accurately predicting prediction values of multiple dimensions.

METHOD FOR CONDENSING TRAINING DATASET, AND IMAGE PROCESSING DEVICE

Absstract of: US20260087787A1

A method of condensing a training dataset and an image processing device are provided. The method of includes generating a cluster set by clustering a training dataset; generating an initial condensed high-resolution (HR) dataset by selecting, for each cluster included in the cluster set, one or more images from among a respective cluster in the training dataset; obtaining a first loss of a first neural network model based on the training dataset and obtaining a second loss of a second neural network model based on the initial condensed HR dataset. The method further includes generating a condensed HR dataset by updating, based on the first loss and the second loss, pixels in each of the one or more images included in each cluster of the initial condensed HR dataset; and executing an operation instruction to transmit the condensed HR dataset to an image processing device.

Method, System, and Device for Wind Speed Prediction and Layout optimization in Wind Power Generation

Absstract of: US20260085661A1

0000 A method, system, and device for wind speed prediction and layout optimization in wind power generation are provided. The method includes: obtaining a basic wind resource dataset of a target region; constructing a physics-informed neural network model based on the basic wind resource dataset; obtaining wind speeds data at a specific location in a velocity field based on the physics-informed neural networks and constructing a training dataset; training the physics-informed neural network model based on the training dataset; reconstructing a wind speed distribution within the velocity field and predicting wind speeds for a next time period with a wind farm using the trained physics-informed neural network model; and optimizing a layout of a wind turbine cluster based on a reconstructed wind speed distribution within the velocity field. The present application reconstructs a two-dimensional velocity field of the wind farm by training the PINN and enables accurate ultra-short-term wind speed prediction.

USE OF A CONVOLUTIONAL NEURAL NETWORK TO AUTO-DETERMINE A FLOOR HEIGHT AND FLOOR HEIGHT ELEVATION OF A BUILDING

Absstract of: US20260087344A1

0000 A method using a convolutional neural network to auto-determine a first floor height (FFH) and a FFH elevation (FFE) of a building. The FFH, and FFE of the building are determined with respect to the terrain or surface of the parcel of land on which the building is located. In turn, by knowing the FFH and/or FFE of the building on the parcel, it is possible to use that information while performing a flood risk assessment to a property without requiring a personal inspection of the parcel by a human.

NEURAL NETWORKS FOR CHANGING CHARACTERISTICS OF VOCALS

Absstract of: EP4715678A2

A messaging system for audio character type swapping. Methods of audio character type swapping include receiving input audio data having a first characteristic and transforming the input audio data to an input image where the input image represents the frequencies and intensities of the audio. The methods further include processing the input image using a convolutional neural network (CNN) to generate an output image and transforming the output image to output audio data, the output audio data having a second characteristic. The input audio and output audio may include vocals. The first characteristics may indicate a male voice and the second characteristics may indicate a female voice. The CNN is trained together with another CNN that changes input audio having the second characteristic to audio having the first characteristic. The CNNs are trained using discriminator CNNs that determine whether audio has a first characteristic or a second characteristic.

INFORMATION GENERATION METHOD AND RELATED APPARATUS

Absstract of: EP4715669A1

0001 Embodiments of this application disclose an information generation method and a related apparatus. The method includes: A second device receives a first message and a third message, and sends a second message to a first device. The first message indicates all or a part of a first generator, the third message indicates all or a part of a third generator, an input supported by the first generator includes first information of a first type, an input supported by the third generator includes fourth information of the first type, and the first generator and the third generator are configured to train a neural network corresponding to a second generator; and the second message indicates all or a part of the second generator, and an input supported by the second generator includes the first information and the fourth information. According to embodiments of this application, information collected in a real scenario may be used to train a generation model, to implement communication-assisted detection and detection-assisted communication, so that a communication network develops towards a more intelligent and adaptive direction.

DOMAIN-AWARE NEUROSYMBOLIC AGENTS FOR IMPROVING PROBLEM-SOLVING ACCURACY AND CONSISTENCY

Absstract of: US20260080313A1

A system receives domain specific questions from users and answers them. The system stores domain specific information comprising domain specific facts and domain specific programs. The system receives an input request to perform a domain specific task for the particular domain. The system provides the input request to a machine learning model trained to predict a score indicating whether the input request should be processed by a symbolic processor or by a neural network. If the score predicted by the machine learning model indicates that the input request should be processed by the symbolic processor, the system determines whether a stored domain specific program can solve the input request. If none of the stored domain specific programs can solve the input request, the system generates a new program for solving the input request using a machine learning based language model and the set of domain specific facts.

UNSUPERVISED PRE-TRAINING OF NEURAL NETWORKS USING GENERATIVE MODELS

Absstract of: US20260080214A1

0000 In various examples, systems and methods are disclosed relating to generating a response from image and/or video input for image/video-based artificial intelligence (AI) systems and applications. Systems and methods are disclosed for a first model (e.g., a teacher model) distilling its knowledge to a second model (a student model). The second model receives a downstream image in a downstream task and generates at least one feature. The first model generates first features corresponding to an image which can be a real image or a synthetic image. The second model generates second features using the image as an input to the second model. Loss with respect to first features is determined. The second model is updated using the loss.

GENERATIVE NEURAL NETWORK SYSTEMS FOR GENERATING INSTRUCTION SEQUENCES TO CONTROL AN AGENT PERFORMING A TASK

Absstract of: US20260080207A1

A generative adversarial neural network system to provide a sequence of actions for performing a task. The system comprises a reinforcement learning neural network subsystem coupled to a simulator and a discriminator neural network. The reinforcement learning neural network subsystem includes a policy recurrent neural network to, at each of a sequence of time steps, select one or more actions to be performed according to an action selection policy, each action comprising one or more control commands for a simulator. The simulator is configured to implement the control commands for the time steps to generate a simulator output. The discriminator neural network is configured to discriminate between the simulator output and training data, to provide a reward signal for the reinforcement learning. The simulator may be non-differentiable simulator, for example a computer program to produce an image or audio waveform or a program to control a robot or vehicle.

SYSTEMS AND METHODS FOR VIDEO-LANGUAGE NEURAL NETWORKS

Absstract of: US20260080681A1

0000 Embodiments described herein provide a vision-language neural network framework that outputs a text response to a user text query relating to the media content of the video input. Specifically, the vision-language neural network may comprise (1) a vision encoder (ViT) transforming each frame input from the video input into a set of tokens, (2) a frame-level tokenizer to reduce the number of tokens, (3) a temporal encoder to build video-level token representations, and (4) an autoregressive LLM generating a text output based on such video tokens and text prompt tokens.

PRETRAINING FRAMEWORK FOR NEURAL NETWORKS

Nº publicación: US20260080533A1 19/03/2026

Applicant:

NVIDIA CORP [US]

Absstract of: US20260080533A1

0000 Apparatuses, systems, and techniques to indicate an extent, to which text corresponds to one or more images. In at least one embodiment, an extent to which text corresponds to one or more images is indicated using one or more neural networks and used to train the one or more neural networks.