Neural networks

NEURAL TAXONOMY EXPANDER

Absstract of: US2025259081A1

Systems and methods for automatically placing a taxonomy candidate within an existing taxonomy are presented. More particularly, a neural taxonomy expander (a neural network model) is trained according to the existing, curated taxonomic hierarchy. Moreover, for each node in the taxonomic hierarchy, an embedding vector is generated. A taxonomy candidate is received, where the candidate is to be placed within the existing taxonomy. An embedding vector is generated for the candidate and projected by a projection function of the neural taxonomy expander into the taxonomic hyperspace. A set of closest neighbors to the projected embedding vector of the taxonomy candidate is identified and the closest neighbor of the set is assumed as the parent for the taxonomy candidate. The taxonomy candidate is added to the existing taxonomic hierarchy as a child to the identified parent node.

METHOD FOR GPU MEMORY MANAGEMENT FOR DEEP NEURAL NETWORK AND COMPUTING DEVICE FOR PERFORMING SAME

Absstract of: US2025259058A1

Embodiments disclosed herein relate to a method for GPU memory management that observes the deep learning of a deep neural network performed by a GPU and reduces the amount of GPU memory used, thereby overcoming limitations attributable to the memory size of the GPU and allowing the more effective performance of the deep learning, and a computing device for performing the same. According to an embodiment, there is disclosed a method for GPU memory management for a deep neural network, the method being performed by a computing device including a GPU and a CPU, the method including: generating a schedule for GPU memory management based on the processing of a unit operation, included in the deep neural network, by the GPU; and moving data required for deep learning of the deep neural network between GPU memory and CPU memory based on the schedule.

MACHINE-LEARNING-BASED MEAL DETECTION AND SIZE ESTIMATION USING CONTINUOUS GLUCOSE MONITORING (CGM) AND INSULIN DATA

Absstract of: US2025259727A1

Disclosed is a meal detection and meal size estimation machine learning technology. In some embodiments, the techniques entail applying to a trained multioutput neural network model a set of input features, the set of input features representing glucoregulatory management data, insulin on board, and time of day, the trained multioutput neural network model representing multiple fully connected layers and an output layer formed from first and second branches, the first branch providing a meal detection output and the second branch providing a carbohydrate estimation output; receiving from the meal detection output a meal detection indication; and receiving from the carbohydrate estimation output a meal size estimation.

EXPLOITING INPUT DATA SPARSITY IN NEURAL NETWORK COMPUTE UNITS

Absstract of: US2025258784A1

A computer-implemented method includes receiving, by a computing device, input activations and determining, by a controller of the computing device, whether each of the input activations has either a zero value or a non-zero value. The method further includes storing, in a memory bank of the computing device, at least one of the input activations. Storing the at least one input activation includes generating an index comprising one or more memory address locations that have input activation values that are non-zero values. The method still further includes providing, by the controller and from the memory bank, at least one input activation onto a data bus that is accessible by one or more units of a computational array. The activations are provided, at least in part, from a memory address location associated with the index.

GRAPHICS ARCHITECTURE INCLUDING A NEURAL NETWORK PIPELINE

Absstract of: US2025252650A1

One embodiment provides a graphics processor comprising a block of graphics cores and circuitry including a programmable neural network unit, the programmable neural network unit including one or more neural network hardware blocks, wherein a neural network hardware block includes circuitry to perform neural network operations and activation operations for a layer of a neural network, the programmable neural network unit addressable by cores within the block of graphics cores, wherein the programmable neural network unit is to configure one or more neural network hardware blocks with a meta-shader neural network, the meta-shader neural network to generate a texture for one of multiple types of terrain.

GENERALIST COMBINATIONAL OPTIMIZATION AGENT LEARNER

Absstract of: US2025249583A1

A method for robot navigation includes: receiving a backbone neural network, the backbone neural network trained to solve a set of autonomous navigation tasks including a first autonomous navigation task, where the first autonomous navigation task includes optimizing a first path visiting a first number of first locations when considering first path costs; receiving a second autonomous navigation task, where the second autonomous navigation task includes optimizing a second path visiting a second number of second locations when considering second path costs, where the second locations are locations in an environment of an autonomous machine, where the second number is different from the first number; configuring adapter layers for the backbone neural network for forming a neural network pipeline; and feeding the second locations and the second path costs to the neural network pipeline and determining a path for the autonomous machine based on an output.

COMBINED DENOISING AND UPSCALING NETWORK WITH IMPORTANCE SAMPLING IN A GRAPHICS ENVIRONMENT

Absstract of: US2025252530A1

An apparatus to facilitate combined denoising and upscaling network with importance sampling in a graphics environment is disclosed. The apparatus includes set of processing resources including circuitry configured to: receive, at an input of a density map neural network, a sampled signal of a current frame and a reconstructed sample of the current frame; output, from the density map neural network, a prediction of a density map of samples based on the input of the current frame; provide the density map of samples to a sampler; reproject the density map of samples to a next frame; and apply the reprojected density map of samples to the next frame to generate a next sampled signal.

Control sequence generation system and methods

Absstract of: US2025252313A1

A model receives a target demand curve as an input and outputs an optimized control sequence that allows equipment within a physical space to be run optimally. A thermodynamic model is created that represents equipment within the physical space, with the equipment being laid out as nodes within the model according to the equipment flow in the physical space. The equipment activation functions comprise equations that mimic equipment operation. Values flow between the nodes similarly to how states flow between the actual equipment. The model is run such that a control sequence is used as input into the neural network; the neural network outputs a demand curve which is then checked against the target demand curve. Machine learning methods are then used to determine a new control sequence. The model is run until a goal state is reached.

IMPROVED PEAK DETECTION

Absstract of: WO2025162867A1

The invention provides, amongst other aspects, a method for detecting peaks in an optical spectrum, the method comprising the steps of obtaining the optical spectrum from at least one optical spectrometer, the optical spectrum comprising a wavelength range; and applying a trained neural network, NN, on the optical spectrum to detect the peaks in the optical spectrum, wherein the detecting of peaks relates to output nodes of the NN having been trained w.r.t zones corresponding to subranges of the wavelength range. Further provided is a device carrying out the method, the device preferably comprising a memory including the trained NN. Further provided is a trained NN having been trained w.r.t zones corresponding to subranges of the wavelength range.

3D PHOTONIC-ELECTRONIC NEUROMORPHIC COMPUTING

Absstract of: WO2025165752A1

A combination of photonic neural networks and electronic neural networks integrated in 3D to form 3D Electronic-Photonic Integrated Circuits (3D EPICs) are described toward enabling new brain-derived neuromorphic hardware with energy-efficiency, connectivity, density, and scalability. Described are the construction of the optoelectronic (OE) neurons in the photonic neural network (PNN) including photonic-memristive dendrites, photonic-memristive synapses, photonic axons, and nano-electronic somas. These OE neurons and their hierarchical interconnections are described in constructing 3D EPICs. The use of a scalable PNN neuromorphic computing simulator is described, as well as PNN training.

ELECTRONIC DEVICE AND CONTROL METHOD THEREOF

Absstract of: WO2025164944A1

This electronic device comprises a memory, and a processor connected to the memory, wherein the processor: when a first user voice in a first language is received, obtains a first translated text in a second language corresponding to the first user voice by inputting the first user voice to a first neural network model; obtains a first keyword text in the second language corresponding to the first user voice by inputting the first user voice to a second neural network model; when a second user voice in the first language is received, obtains a second translated text in the second language corresponding to the second user voice by inputting the second user voice and the first keyword text to the first neural network model; and provides the first translated text and the second translated text.

Systems and methods for optimizing hyperparameters for machine learning models

Absstract of: GB2637695A

A combined hyperparameter and proxy model tuning method is described. The method involves iterations for hyperparameters search 102. In each search iteration, candidate hyperparameters are considered. An initial (‘seed’) hyperparameter is determined by initialization function 110, and used to train (104) one or more first proxy models on a target dataset 101. From the first proxy model(s), one or more first synthetic datasets are sampled using sampling function 108. A first evaluation model is fitted to each first synthetic dataset, for each candidate hyperparameter, by applying fit function 106 enabling each candidate hyperparameter from hyperparameter generator 112 to be scored. Based on the respective scores assigned to the candidate hyperparameters, a candidate hyperparameter is selected and used (103) to train one or more second proxy models on the target dataset. Hyperparameter search may be random, grid and Bayesian. Scores by scoring function 114 can be F1 scores. Uses include generative causal model with neural network architectures.

ROBOT NAVIGATION BASED ON MULTI-TASK NEURAL COMBINATORIAL OPTIMIZATION

Absstract of: EP4597374A1

A method for robot navigation is disclosed, the method comprising receiving a backbone neural network, the backbone neural network having been trained to solve a set of robot navigation tasks comprising a first robot navigation task of optimizing a first path visiting a first number of first locations when considering first path costs. The method further comprises receiving a second robot navigation task, wherein the second robot navigation task comprises optimizing a second path visiting a second number of second locations when considering second path costs. The method further comprises configuring adapter layers for the backbone neural network for forming a neural network pipeline, wherein the adapter layers comprise a node adapter, an edge adapter and an output adapter. The method further comprises feeding the second locations and the second path costs to the neural network pipeline to determine a path for an autonomous machine based on the output.

USER SELECTION PREDICTION BASED ON GRAPH NEURAL NETWORK AND USER SELECTION SEQUENCE

Absstract of: MX2025007777A

A method for predicting a next user selection in an electronic user interface includes receiving a sequence of user selections through the electronic user interface, determining a context embedding vector according to the sequence of user selections, querying a knowledge graph, the knowledge graph respective of a plurality of possible user selections, with the context embeddings vector, to obtain a knowledge-enhanced representation of the sequence, determining, with a graph neural network respective of the knowledge graph, based on the knowledge-enhanced representation, a respective representation of each selection in the sequence of user selections, and determining a predicted next user selection according to the respective representations of the selections in the sequence of user selections.

ELECTRONIC DEVICE AND METHOD FOR CONTROLLING SAME

Absstract of: US2025245119A1

An example method for controlling an electronic device includes detecting at least one user and acquiring user information of the detected at least one user; determining a user mode based on the acquired user information; determining a service to be provided to the detected at least one user, by inputting the user information and the determined user mode as input data to a model learned by an artificial intelligence algorithm; and providing the determined service corresponding to the user mode. A method for providing the service by the electronic device may at least partially use an artificial intelligence model learned according to at least one of machine learning, neural network, and deep learning algorithms.

METHOD AND SYSTEM FOR EMERGENCY MONITORING AND ACTING IN DOMESTIC ENVIRONMENTS

Absstract of: US2025246065A1

A method and system for emergency monitoring and acting in domestic environments of a user (10), wherein a data stream capturing device (310) provides data for local processing (320) in a user end device (300). This involves classifying events using a lightweight neural network and triggering actions if emergency is detected based on the events classification. In case of emergency, an external user (20) is notified through an external user end device (330) to manage the situation by selecting actions from the set of triggered actions and provides a validation (370) on whether the actions were correctly taken once the situation is under control. The causality supervisor (340) reviews the log of events and actions (360) and the validation (370) from the external end device (330) for improvements. The local knowledge database (380) is updated to adapt or reinforce behaviour (350) based on the detected emergencies and triggered actions. Also, the user (10) can provide feedback to adapt the system as the causality supervisor (340), for example, if a false negative happens.

- Inference apparatus and learning method thereof using a forward-forward algorithm based on unsupervised neural network

Absstract of: KR20250115745A

비지도 신경망 기반의 포워드-포워드 알고리즘을 사용하는 추론장치 및 그 학습방법에 관한 기술이 개시된다. 추론장치는 입력된 데이터로부터 잠재벡터를 출력하도록 학습된 신경망을 구비하는 복수의 포워드 레이어와, 복수의 포워드 레이어 간에 잠재벡터를 전달하는 잠재벡터 전달부와, 복수의 포워드 레이어 각각에서 출력되는 잠재벡터들을 결합하여 특징벡터를 생성하는 특징벡터 생성부와, 특징벡터를 이용하여 추론 결과를 출력하도록 학습된 출력 레이어를 포함한다. 복수의 포워드 레이어는 비지도 학습 방법에 의하여 잠재벡터를 출력하도록 학습된 비지도 신경망을 포함하고, 잠재벡터 전달부는 먼저 학습되는 전단 포워드 레이어에서 출력되는 잠재벡터를 전단 포워드 레이어의 다음으로 학습되는 후단 포워드 레이어의 입력으로 전달한다. 본 발명에 의하면 특수한 입력과 손실함수를 사용하지 않고도 안정적으로 포워드-포워드 알고리즘을 적용한 추론장치를 학습시킬 수 있다.

KNOWLEDGE GRAPH CONSTRUCTION VIA GENERATIVE ARTIFICIAL INTELLIGENCE

Absstract of: US2025245226A1

Systems or techniques that facilitate knowledge graph construction via generative artificial intelligence are provided. In various embodiments, a system can access a plurality of electronic documents associated with design or fabrication of a medical imaging scanner. In various aspects, the system can construct a knowledge graph representing the plurality of electronic documents, by iteratively executing a generative text-to-text neural network on a design discovery tree associated with the medical imaging scanner. In various instances, the system can access a natural language query regarding the medical imaging scanner and can convert, via execution of another neural network, the natural language query to a structured query. In various cases, the system can execute the structured query over the knowledge graph, thereby yielding an electronic answer to the natural language query.

IMPROVED PEAK DETECTION

Absstract of: EP4592654A1

The invention provides, amongst other aspects, a method for detecting peaks in an optical spectrum, the method comprising the steps of obtaining the optical spectrum from at least one optical spectrometer, the optical spectrum comprising a wavelength range; and applying a trained neural network, NN, on the optical spectrum to detect the peaks in the optical spectrum, wherein the detecting of peaks relates to output nodes of the NN having been trained w.r.t zones corresponding to subranges of the wavelength range. Further provided is a device carrying out the method, the device preferably comprising a memory including the trained NN. Further provided is a trained NN having been trained w.r.t zones corresponding to subranges of the wavelength range.

KNOWLEDGE GRAPH CONSTRUCTION VIA GENERATIVE ARTIFICIAL INTELLIGENCE

Absstract of: EP4592896A1

Systems or techniques that facilitate knowledge graph construction via generative artificial intelligence are provided. In various embodiments, a system can access a plurality of electronic documents (e.g., 106) associated with design or fabrication of a medical imaging scanner (e.g., 104). In various aspects, the system can construct a knowledge graph (e.g., 206) representing the plurality of electronic documents, by iteratively executing a generative text-to-text neural network (e.g., 202) on a design discovery tree (e.g., 204) associated with the medical imaging scanner. In various instances, the system can access a natural language query (e.g., 1602) regarding the medical imaging scanner and can convert, via execution of another neural network (e.g., 1604), the natural language query to a structured query (e.g., 1606). In various cases, the system can execute the structured query over the knowledge graph, thereby yielding an electronic answer (e.g., 1608) to the natural language query.

DEEP NEURAL NETWORK BASED IMAGE COMPRESSION USING A LATENT SHIFT BASED ON GRADIENT OF LATENTS ENTROPY

Absstract of: MX2025003035A

A deep neural network based video compression system in which gradients of entropies with respect to side and main latents are used on decoding side to improve compression efficiency.

SYSTEMS AND METHODS TO IDENTIFY NEURAL NETWORK BRITTLENESS BASED ON SAMPLE DATA AND SEED GENERATION

Absstract of: US2025238288A1

Systems and methods for determining neural network brittleness are disclosed. For example, the system may include one or more memory units storing instructions and one or more processors configured to execute the instructions to perform operations. The operations may include receiving a modeling request comprising a preliminary model and a dataset. The operations may include determining a preliminary brittleness score of the preliminary model. The operations may include identifying a reference model and determining a reference brittleness score of the reference model. The operations may include comparing the preliminary brittleness score to the reference brittleness score and generating a preferred model based on the comparison. The operations may include providing the preferred model.

DATA PROCESSING METHOD AND APPARATUS, AND STORAGE MEDIUM

Absstract of: US2025238658A1

This application relates to a data processing method and apparatus, and a storage medium. The method includes: extracting a feature sequence of target data, where the feature sequence includes T input features, T is a positive integer, and t∈1, T; obtaining T hidden state vectors based on a recurrent neural network, where a tth hidden state vector is determined based on a (t−1)th input feature, a (t−1)th hidden state vector, and a (t−1)th extended state vector, and the (t−1)th extended state vector is obtained by performing lightweight processing based on the (t−1)th hidden state vector; and obtaining a processing result of the target data based on the T hidden state vectors by using a downstream task network.

PARTITIONING-BASED SCALABLE WEIGHTED AGGREGATION COMPOSITION FOR KNOWLEDGE GRAPH EMBEDDING

Absstract of: WO2025151950A1

The disclosure relates to methods and systems of partitioning-based scalable weighted aggregation composition for embeddings learned from knowledge graphs for training neural networks to perform downstream machine-learning tasks. For example, a system may access a knowledge graph comprising a plurality of nodes and partition the knowledge graph into a plurality of partitions based on edge densities between nodes of the knowledge graph. The system may perform partition-wise encoding using compositional message passing between nodes that enables learning from neighboring nodes. The system may generate an embedding for each node and each relation type in each partition based on the partition-wise encoding using compositional message passing. The system may concatenate the generated embeddings from the plurality of partitions. The system may train a global neural network for a downstream prediction task based on the concatenated embeddings using one or more weight matrices.

GENERATIVE PIPELINE USING SINGLE SELFIE INPUT

Nº publicación: US2025238902A1 24/07/2025

Applicant:

SNAP INC [US]
Snap Inc