Alerta

Selection of random access preamble

Resumen de: GB2642672A

Determination and implementation of a random access channel (RACH) preamble selection policy (PSP). An apparatus such as a distributed unit (DU) 420 of a first radio access technology (RAT) determines a RACH PSP based upon first information. The DU receives from another DU of a second RAT, second information at step (6) and updates the RACH PSP at step (7) based upon the first and second information. At step (8) the RACH PSP is transmitted to a user equipment (UE), 410. The UE selects a RACH preamble based upon the selection policy and transmits the preamble to the DU. The RACH PSP may comprise a probability distribution parameter which may include a type of distribution function, e.g. normal, Gaussian or exponential distribution, a parameter associated with a distribution function or allocation information of RACH preambles. The information may comprise: a mode or state of operation of the apparatus, an arrival rate of random access requests for the apparatus, a number of RACH preamble collisions at the apparatus or load information of the apparatus. A trained machine learning model or algorithm may be used to determine the RACH PSP based on the information to reduce potential RACH preamble collisions.

DETERMINING LABELS OF INHERITANCE DATASETS USING SIMULATED DATA INSTANCES

Resumen de: WO2026015162A1

Disclosed is a method for determining inheritance labels of users based on inheritance datasets of the users. The method includes generating a plurality of reference panels for a plurality of data-inheritance origins, each reference panel corresponding to a data-inheritance origin and comprising reference-panel datasets representative of the data-inheritance origin. The method constructs a plurality of simulated data trees that are built using the reference-panel datasets that are selected from the plurality of reference panels. The method generates a plurality of simulated inheritance datasets representing a plurality of simulated named entities, each representing a descendant named entity in one of the simulated data trees. The method trains a machine learning model to determine inheritance labels of an inheritance dataset.

INTELLIGENT AND REAL-TIME TASK GUIDANCE SYSTEM FOR SURGICAL OPERATING ROOMS

Resumen de: WO2026015586A1

Systems and methods are described for determining and assigning tasks for performing medical procedures. The system may be configured to receive a plurality of data streams related to a medical procedure, wherein the plurality of data streams includes one or more of system data, medical environment data, and indications of personnel performing the medical procedure; analyze, using a task generation machine learning model, the plurality of data streams to generate natural language output relating to one or more tasks to be performed in furtherance of the medical procedure, wherein one or more inputs into the task generation machine learning model includes inputting embeddings of the plurality of data streams; analyze, via a task assignment machine learning model, the one or more tasks to assign the tasks to respective personnel; and provide indications to the respective personnel for performing the respective tasks assigned to the respective personnel.

COMMUNICATION METHOD AND APPARATUS

Resumen de: US20260019345A1

A communication method and apparatus. A first device sends capability information to a second device, so that the second device can send control information to the first device based on the capability information. The control information is usable to indicate that a first machine learning (ML) model corresponding to a first management function (MnF) on the first device is allowed to be trained. The first device trains the first ML model based on the control information.

SYSTEMS AND METHODS FOR PREDICTING PROTEIN-PROTEIN INTERACTION USING MACHINE LEARNING TECHNIQUES

Resumen de: WO2026015381A1

Systems and methods are disclosed for building a regression-based machine learning algorithm to predict native and near-native binding confirmations between two or more proteins. They include receiving training data, the training data, representing a plurality of protein to protein docking poses based on one or more hotspots and a set of descriptors characterizing interaction interface features of one or more of the plurality of protein to protein docking poses. The systems and methods further specify a dedicated training data set based on a variety of single domain interactions. The training data, set comprises an information space required to create dedicated regression and classification models in order to identify near native binding configurations. A specific and diverse set of interaction interface descriptors is calculated to provide the foundation for both, the regression and classification models.

ADDITIVE GRID-BASED FEATURE SELECTIONS FOR TRAINING OF MACHINE LEARNING (ML) MODELS FOR ELECTORNIC DESIGN AUTOMATION (EDA) DESIGN FLOWS

Resumen de: WO2026015127A1

Various systems and method are presented in support of additive grid-based feature selection for training of machine learning (ML) models for electronic design automation (EDA) design flows. The technical aspects presented herein may include determining candidate features for the ML model from an input dataset and scanning the candidate features for different data partition sizes to evaluate training of the ML model through various feature sets and data partitioning parameters. An additive feature grid may generated by adding dimensions through performance of dimension iterations. A selected feature set may be determined based on dimensions of the additive feature grid and used to train the ML model for the EDA design flow.

SYSTEMS AND METHODS FOR MANAGING HEALTH TREATMENT

Resumen de: US20260018293A1

Disclosed herein is a computer implemented method for managing healthcare diagnosis and treatment. The method includes the steps of monitoring at least one data source for a workflow trigger comprising at least one of an order, a test result, an appointment, a patient demographic, a patient status, a patient history, a patient communication, or a condition; and triggering a workflow upon the detection of a workflow trigger. The workflow comprises a first decision-making layer configured to manage at least one of a rule, a patient test, and a patient communication; a second decision-making layer configured to manage at least one workflow, wherein the workflow comprises at least one rule; and a third decision-making layer configured to manage at least one machine learning model, wherein the machine learning model is configured to process data relevant to the workflow and to determine a probability of a condition to be tested.

MACHINE LEARNING DEVICE, MACHINE LEARNING METHOD, AND COMPUTER-READABLE MEDIUM STORING MACHINE LEARNING PROGRAM

Resumen de: US20260017931A1

A machine learning device includes an image set acquiring unit to acquire an image set including images, and an image set selecting unit to select an image set similar to the acquired image set from a plurality of image sets different from the acquired image set. In addition, the machine learning device includes a performance comparison unit, and a preprocessing acquisition unit to select a learning model from a plurality of machine-learned learning models based on a performance comparison result by the performance comparison unit and to acquire preprocessing performed on an image set used for machine learning of the learning model selected. Furthermore, the machine learning device includes a model learning unit to perform the acquired preprocessing on the acquired image set and to cause a learning model that has not yet trained to perform machine learning using the preprocessed image set.

MACHINE LEARNING MODEL CONTINUOUS TRAINING SYSTEM

Resumen de: US20260019655A1

Described is a system for performing a set of machine learning model training operations that include: accessing media content items associated with interaction functions initiated by users of an interaction system, generating training data including labels for the media content items, extracting features from a media content item of the media content items, identifying additional media content items to include in the training data based on the extracted features from the media content item, processing the training data using a machine learning model to generate a media content item output; and updating one or more parameters of the machine learning model based on the media content item output. The system checks whether retraining criteria has been met, and repeats the set of machine learning model training operations to retrain the machine learning model.

APPLICATION OF ARTIFICIAL INTELLIGENCE MACHINE LEARNING (AIML) MODELS ASSOCIATED WITH SAME FUNCTIONALITY

Resumen de: WO2026015208A1

Disclosed are techniques for wireless communication. In an aspect, a processing device may receive, from a server device, a request for an output based on application of a plurality of artificial intelligence machine learning (AIML) models associated with a same functionality. The processing device may apply the plurality of AIML models to obtain a plurality of respective candidate outputs, the plurality of candidate outputs being associated with the functionality. The processing device may transmit the output to the server device in response to the request, the output indicating at least one of the plurality of candidate outputs.

MACHINE INTELLIGENCE ON WIRELESS EDGE NETWORKS

Resumen de: WO2026015743A1

Deep learning has revolutionized image classification, robotics, life sciences, and other fields. However, the exponential growth in deep neural network parameters and data volumes has strained traditional computing architectures, primarily due to the data movement bottleneck. A Machine Intelligence on Wireless Edge Networks (MIWEN) approach for deep learning on ultra-low-power edge devices addresses this data movement bottleneck. MIWEN leverages disaggregated memory access to wirelessly stream machine learning (ML) models to edge devices, mitigating memory and power bottlenecks by integrating computation into the existing radio-frequency (RE) analog chain of the wireless transceivers on edge devices. MIWEN aims to achieve scalable and efficient implementations, significantly reducing energy consumption and latency compared to conventional digital signal processing systems.

SYSTEMS AND METHODS FOR GENERATING DYNAMIC TRANSIT ROUTES

Resumen de: US20260016310A1

A computing device comprising: obtaining telematics data generated by an autonomous vehicle; building, using a machine learning algorithm, a transit model based at least in part upon the telematics data; generating, based at least in part upon the transit model, a dynamic transit route; calculating a potential benefit comprising at least one of an amount of fuel cost savings, reduced travel time, insurance savings, or environmental pollution reduction when the dynamic route is used compared to a different route; transmitting a notification comprising the dynamic route and the potential benefit to a display or touchscreen of the autonomous vehicle; receiving, via the display screen or touchscreen, a selection input indicating acceptance or declination of the dynamic route; when the selection input indicates declination, modifying the route; and when the selection input indicates acceptance, instructing the autonomous vehicle to autonomously drive along the dynamic route.

DETERMINING LABELS OF INHERITANCE DATASETS USING SIMULATED DATA INSTANCES

Resumen de: US20260017284A1

Disclosed is a method for determining inheritance labels of users based on inheritance datasets of the users. The method includes generating a plurality of reference panels for a plurality of data-inheritance origins, each reference panel corresponding to a data-inheritance origin and comprising reference-panel datasets representative of the data-inheritance origin. The method constructs a plurality of simulated data trees that are built using the reference-panel datasets that are selected from the plurality of reference panels. The method generates a plurality of simulated inheritance datasets representing a plurality of simulated named entities, each representing a descendant named entity in one of the simulated data trees. The method trains a machine learning model to determine inheritance labels of an inheritance dataset.

DYNAMIC RESOURCE ALLOCATION FOR A DISTRIBUTED NODE NETWORK

Resumen de: US20260017115A1

Methods and systems for dynamically allocating resources for a distributed node network. A system may receive a workflow comprising computer program code configured to perform one or more processes of the workflow when executed. The system may generate a set of data inputs, each data input being representative of a (a) resource allocation for allocating compute resources to the one or more nodes during performance of the one or more processes and (b) sample data on which to perform the process(es). The system may determine a performance metric value for each data input by executing at least a portion of the workflow to perform the process(es) on the sample data using the specified resource allocation. Using the generated set of data inputs, a machine learning model may be trained to identify a required resource allocation for a given set of data inputs for meeting the target performance value.

AUTOMATED FEATURE SELECTION FOR SPLIT NEURAL NETWORKS

Resumen de: US20260017517A1

A computer-implemented method and apparatus for feature selection using a distributed machine learning (ML) model in a network comprising a plurality of local computing devices and a central computing device is provided. The method includes training, at each local computing device, the ML model during one or more initial training rounds using a group of input features representing a input features layer of the ML model. The method further includes generating, at each local computing device, based on the one or more initial training rounds, feature group values. The method further includes transmitting, from each local computing device, to the central computing device, the generated feature group values. The method further includes receiving, at each local computing device, from the central computing device, central computing device gradients. The method further includes computing, at each local computing device, local computing device gradients, using the received central computing device gradients. The method further includes generating, at each local computing device, a gradient trajectory for each input feature in the group of input features based on the computed local computing device gradients. The method further includes identifying, at each local computing device, based on the generated gradient trajectory, whether each input feature in the group of input features is non-contributing. The method further includes removing, at each local computing device, from the group

AUTOMATED TASK DETECTION USING MACHINE LEARNING AND LARGE LANGUAGE MODEL

Resumen de: US20260017573A1

The invention relates to a method for improving task detection through a combination of machine learning and natural language processing. The method involves preparing data by preprocessing and cleaning to ensure suitability for machine learning algorithms, followed by training a LightGBM model using the prepared data. Task detection results are generated using the trained LightGBM model. The method further includes analyzing feature importance and generating new features using a large language model (LLM). These new features are used to expand the dataset, and the LightGBM model is retrained to enhance task detection performance. This approach automates feature extraction, improves performance, increases adaptability, and enhances the generalizability of task detection methods.

Response Prediction for Electronic Communications

Resumen de: US20260017544A1

Systems, methods, and apparatuses are described herein for performing sentiment analysis on electronic communications relating to one or more image-based communications methods, such as emoji. Message data may be received. The message data may correspond to a message that is intended to be sent but has not yet been sent to an application. Using a first machine learning model, one or more subsets of the plurality of emoji may be determined. The one or more subsets of the plurality of emoji may comprise one or more different types and quantities of emoji, and may each correspond to the same or a different sentiment. Using a second machine learning model, one or more emojis may be selected from the one or more subsets. The one or more emojis selected may correspond to responses to the message.

METHOD FOR PROVIDING INTELLIGENT RESPONSE AGENT BASED ON ADVANCED INFERENCE AND ESTIMATION FUNCTION, AND SYSTEM THEREFOR

Resumen de: EP4679287A1

A method and system for providing an intelligent response agent based on a sophisticated reasoning and speculation function according to an embodiment of the present disclosure can generate and provide response data for queries related to specialized documents using a deep-learning neural network that implements a stepwise process for a sophisticated reasoning and speculation function.

MULTI-TASKING MODEL TRAINING METHOD AND MULTI-TASKING PERFORMING METHOD USING MACHINE LEARNING MODEL TRAINED ON BASIS THEREOF

Resumen de: EP4679330A1

A multi-tasking model training method and a multi-tasking performing method using a machine learning model trained on the basis thereof, according to an embodiment of the present invention, may mutually transfer and learn knowledge data of a latent space for each task through geometric alignment in one integrated latent space in order to process a multi-task for output according to a plurality of domains.

Method for improving accuracy of machine learning models

Nº publicación: GB2642421A 14/01/2026

Solicitante:

SAMSUNG ELECTRONICS CO LTD [KR]
Samsung Electronics Co., Ltd

Resumen de: GB2642421A

Method for training a neuro-symbolic machine learning model, comprising: for each image depicting at least two objects of a training dataset: inputting the image into a neural module 102, 200, 202 to obtain bounding boxes and features therein (digit); inputting each bounding box and object feature into a symbolic module (106, Fig.1; rest of Fig.2) to obtain a plurality of possible labels i.e. partial labels 212 and possible relationships 210 as a new partially-labelled training dataset; and training the neuro-symbolic model (neural module and the symbolic module) by calculating a loss from a ground truth label for the image. The symbolic module may use a set of logical rules to constrain the labels and explanations (R1-R5, Fig.7). The trained neuro-symbolic model may generate a scene graph, perform action recognition, perform visual question answering (Fig.4) or control an autonomous or semi-autonomous electronic device. The electronic device may be a moveable robot or a wearable augmented reality device.