Absstract of: EP4841518A1
0001 A computer-implemented method comprising: performing a signature generation process to generate a signature based on a target knowledge graph, KG, the signature generation process including determining the target KG's graph Laplacian, determining eigenvalues of the graph Laplacian, determining gaps between the eigenvalues, determining an importance score for each edge in the target KG based on the graph Laplacian, and generating the signature as an array of values based on the gaps between the eigenvalues and the importance scores; using a machine learning, ML, classifier, classifying, based on the signature, the target KG as adversarial or clean, wherein the ML classifier has been trained according to a training process including: training the ML classifier to correctly classify training KGs as adversarial or clean based on the corresponding training signatures.
Absstract of: EP4841945A1
0001 A method for training a lighting control machine learning model for deployment in a lighting system operating in an environment. The method comprising the steps of deploying a multi-control machine learning model in the environment, said multi-control machine learning model configured to generate control outputs based on a natural language input, each of said control outputs for controlling the lighting system and/or a non-lighting system, determining whether the output generated by the multi-control machine learning model comprises a control output for controlling the lighting system, creating a training dataset only comprising the control outputs for which it has been determined that the control output generated by the multi-control machine learning model comprises a control output for controlling the lighting system, wherein the training dataset comprises the natural language inputs and the corresponding generated control outputs for controlling the lighting system, and training the lighting control machine learning model with the created training dataset, said lighting control machine learning model configured to only generate control outputs for controlling the lighting system.
Absstract of: EP4841641A1
A Computer-implemented method outputting a pathological location is provided. The method comprises a plurality of steps. One step is directed to receiving patient data. Another step is directed to applying a trained machine learning model to the patient data to determine one or more findings. Another step is directed to mapping the one or more findings to nodes of a clinical knowledge database to determine one or more mapped findings. The nodes of the clinical knowledge database comprise reference findings and/or reference data. Another step is directed to determining the pathological location associated with the one or more mapped findings. Another step is directed to outputting the pathological location.
Absstract of: EP4841545A1
0001 The invention relates to a computer-implemented method, and a corresponding system, of controlling an industrialized fermentation process to produce a personalized fermented food product. The method comprises receiving (S1) user-specific data, via an input interface, said user-specific data including at least one nutritional target comprising a recommended daily value or therapeutic threshold for a nutrient or macronutrient, and accessing (S2) a fermentation knowledge base comprising ingredient composition data specifying macronutrient and/or micronutrient profiles, measured correlations between fermentation parameters and changes in nutrient content, to thereby retrieve food production data for the fermentation process. The method further comprises executing (S3), based on the user-specific data and/or the retrieved food production data, an Artificial Intelligence, AI, based processing module that uses at least one non-linear or machine-learning model trained to predict nutrient retention in response to varying fermentation conditions, and computes fermentation settings to meet said at least one nutritional target within a target interval and/or achieve a target nutrient retention level. The method also comprises applying (S4) the fermentation settings for at least partly controlling the fermentation process.
Absstract of: US20260359734A1
Certain aspects of the present disclosure provide techniques and apparatus for machine learning. In an example method, an input is received at a first generative machine learning model associated with a first vocabulary and being capable of being associated with a plurality of second generative machine learning models, each being associated with a respective second vocabulary. Based on the input, a first token identifier from the second vocabulary associated with an operative second generative machine learning model is mapped to at least one second token identifier associated with the input from the first vocabulary. The operative second generative machine learning model is one of the plurality of second generative machine learning models. A set of draft tokens is generated based on processing the at least one second token identifier using the first generative machine learning model. A response to the input is generated based on the set of draft tokens.
Absstract of: WO2026234646A1
Disclosed herein are methods for determining radio resource management (RRM) information. In one embodiment, a user equipment (UE) can receive a legacy RRM configuration and an inference configuration from a network node. The legacy RRM configuration can be used for determining the RRM information, and the inference configuration can be used for determining the RRM information using artificial intelligence or a machine learning model. The inference configuration can be used when it is applicable or when its performance satisfies a performance threshold. Otherwise, the legacy RRM configuration can be used. The UE can monitor the inferences provided by the inference configuration to identify when the inference configuration is applicable and/or satisfying the performance threshold. The UE can toggle between the legacy RRM configuration and the inference configuration in accordance with the applicability and performance of the inference configuration.
Absstract of: WO2026234436A1
The present invention provides a system (102) and method for the automated prediction of pathological cancer parameters using only preoperative data, thereby enhancing early decision-making in oncology. By curating and preprocessing clinical, radiological, and non-pathological patient data, the system (102) employs multiple machine learning models (112) to predict key pathological features such as tumor stage (pT), nodal stage (pN), depth of invasion (DOI) etc. The use of a voting-based ensemble mechanism improves accuracy and robustness, with the system generating confidence scores and prediction accuracy statistics for each parameter. Based on the predicted pathological features, the invention further calculates the final pathological cancer stage and assists in determining the need for neck node metastasis surgery. The invention reduces reliance on invasive post-surgical histopathology, enhances clinical efficiency, and supports personalized cancer treatment planning. It is scalable, adaptable to various healthcare systems, and supports federated learning for continuous improvement.
Absstract of: US20260359430A1
Systems and methods are disclosed for automating anomaly detection. One method may include retrieving one or more data metrics from a data server, mapping, by a machine learning model, the one or more data metrics using one or more logic rules, determining, by the machine learning model, one or more patterns based on the mapping, determining, by the machine learning model, one or more anomalies based on the one or more patterns, generating, by the machine learning model, a recommendation based on the one or more anomalies, and automatically notifying one or more users of the one or more anomalies and the recommendation.
Absstract of: US20260359274A1
A computer-implemented system automatically detects, analyzes, and resolves data inconsistencies within electronic application datasets using artificial intelligence and rule-based validation. The system performs semantic normalization and validation of multi-document datasets. The system receives diverse documents containing structured and unstructured fields, extracts text, key–value pairs, and contextual features, and generates a canonical semantic representation that encodes syntax, semantics, and relationships. A machine-learning model determines semantic equivalence among fields across documents and maps source values into destination fields of a target document to pre-populate missing data. After pre-fill, a semantic analyzer detects issues including inconsistencies, omissions, contradictions, or formatting defects, assigns severity levels, and presents correction prompts or automatically proposed revisions via a user interface. The system receives and applies corrections, then re-evaluates the dataset with the machine-learning model to confirm resolution and update the semantic representation, enabling iterative completion and validation of complex submissions.
Absstract of: WO2026235927A1
In some examples of the techniques described herein, one or more network settings may be associated with (e.g., mapped to) an identifier. In some approaches, a network entity may indicate one or more identifiers to a wireless device for checking an artificial intelligence or machine learning model on the wireless device. In some approaches, the network may maintain an alignment or correspondence between identifiers and their associated network settings. It may be helpful for a wireless device to have or build relationship information between network settings or associated identifiers. For instance, if the wireless device has information indicating that an identifier X and an identifier Y are associated with a same network setting, the wireless device may utilize the information to determine whether to apply an operation when the network signals a change from identifier X to identifier Y.
Absstract of: AU2026257629A1
A system and method using machine learning prediction that estimates future sport outcomes by mathematically representing current and future match states as a Bayesian network parameterized by underlying machine learning models. ct c t
Absstract of: AU2026257634A1
A computer implemented method for determining entity attributes. The method comprises determining one or more entity identifiers, determining an entity server address of the entity based on the one or more entity identifiers, wherein the entity server address points to an entity server; verifying the entity server address transmitting a message for request for information to the entity server address, receiving entity information from the entity server; and providing, to a machine learning model, the received entity information. The machine learning model is trained to generate a numerical representations of entities based on the entity information. ct c t
Absstract of: US20260359013A1
Methods and systems are presented for providing a computer platform that manages the impacts of government regulations on existing software processes of an online service provider. A regulation document is obtained from a government agency. The regulation document is processed, and legal obligations relevant to an online service provider are extracted from the regulation document. An ensemble machine learning model is used to recommend, for each of the legal obligations, software controls that can be implemented within one or more software processes of the online service provider to mitigate a risk of the legal obligations. The ensemble machine learning model may include an attribute-based model and a text-based model. An explainable visual interface is provided to present the recommended software controls and context that indicates to a user how the software controls are determined for the legal obligations.
Absstract of: US20260359604A1
The invention relates to a black hole prediction and detection system by computer-aided machine learning in software backbone infrastructure networks and an operation method thereof. The black hole prediction and detection system by computer-aided machine learning of the invention predicts and detects outlier network behavior anomalies, thereby generating a warning, detects the black hole anomalies which have not yet been seen without any need to retrain the model, allows the black hole anomalies to be directly and instantly predicted without any need for tagging the data, avoids the problem of learning from imbalanced data in the classification problems caused by data tagging being overcome, without requiring extra time and workload, provides general black hole prediction which may adapt to varying topologies, detects the point, contextual and collective anomalies together for black hole prediction, and analyzes and interprets the instances detected as black holes by computer-aided machine learning, thereby explaining the grounds for detection.
Absstract of: AU2025253676A1
The present disclosure relates to generating suggested responses to customer requests using machine learning models. In one example, a method includes: receiving, from a customer, a customer request via a communication channel; displaying in a customer support user interface the customer request; processing the customer request with a machine learning model to determine a suggested response to the customer request; and displaying in an agent assistance user interface element in the customer support user interface: the suggested response to the customer request; a first user interface element configured to implement the suggested response; and a second user interface element configured to dismiss or modify the suggested response.
Absstract of: US20260359126A1
Systems and methods are disclosed for centralized management and application of distributed parameters through a data generation platform. The platform obtains an aggregate rule set including zone-specific parameter sets and zone classifiers for multiple zones, where the rule set determines allocation metric values associated with transaction values of resources. Upon receiving input data including entity identifiers, resource identifiers, and transaction data, the platform maps the data to zones by comparing zone classifiers, filters the aggregate rule set to generate a zone-specific rule subset, and applies the subset to generate allocation metric values. The platform can process requests synchronously through online message queues and/or asynchronously for bulk processing. Machine learning models can use distribution patterns across multiple resource sets to generate overall allocation metrics and modification suggestions. The platform can provide a graphical user interface for displaying allocation metrics, resource information, and/or supplemental explanations.
Absstract of: US20260359073A1
A method for processing a plugin-based task in an intelligent machine learning algorithm design includes: registering first and second plugins in an execution process of a target task, wherein the execution process of the target task includes generating a task processing result based on historical data related to the target task and a large model, and performing a next round of task processing based on a feedback result of the task processing result; in response to completing execution of a step to which the first plugin subscribes, determining whether to trigger expert participation in decision-making according to a first running state of the target task by invoking the first plugin; and in response to completing execution of a step to which the second plugin subscribes, determining whether to trigger training of the large model according to a second running state of the target task by invoking the second plugin.
Absstract of: US20260359723A1
Machine learning systems, architectures, and methods, capable of varying and adapting topologies including during and after training. The systems and methods analyze information-theoretic measures of information processing by the components of a framework. The systems and methods may also have a set of constraints related to such theoretic measures of information processing. The systems and methods employ such measures to improve or optimize the topology of the framework being used.
Absstract of: GB2706086A
The present invention relates to a way of processing a measurement signal, the measurement signal comprising measurements of a sample by a sensor unit comprising a nanopore, the method comprising: analysing the measurement signal using a machine learning model to determine that a portion of the measurement signal is associated with a first state of the nanopore, the first state being one of a plurality of states of the nanopore; and determining whether to further analyse the portion of the measurement signal in dependence on the determination that the portion of the measurement signal is associated with the first state of the nanopore. In preferred embodiments, the machine learning model identifies states of a nanopore such as those bound to substrate, those not bound to substrate, or those blocked by a substrate. The data can then be filtered such that only data corresponding to specific nanopore states are further processed. In a further aspect, the measurement signal is processed to determine values of a metric using, for example, an autocorrelation of the measurement signal. Figure 3
Absstract of: EP4838534A1
0001 The disclosure relates to a method for determining additional predicted test behavior characteristics of a virtual material, wherein the virtual material is associated with predicted test behavior characteristics with respect to at least one virtual material test, the method performed by a processor, the method comprising: applying the virtual material and the associated predicted test behavior characteristics with respect to at least one virtual material test to a machine-learning model, wherein the machine-learning model is adapted to a plurality of training data sets, and determining the additional predicted test behavior characteristics of the virtual material using the machine-learning model.
Absstract of: EP4838410A1
0001 The invention concerns a method of decision-support for a vehicle-control or vehicle-control related simulation, the method comprises the following phases: an acquisition phase in which a user device (10) acquires contextual data relative to a neurophysiological state of a user, reception phase in which the user device (10) receives a decision-support machine-learning model and a correction machine-learning model from a server (20), a first application phase in which the user device (10) applies the decision-support machine-learning model on the contextual data to obtain decision-support data, a second application phase in which the user device (10) applies the correction machine-learning model on the decision-support data and contextual data to obtain corrected decision-support data, and a supply phase in which the user device (10) provides the corrected decision-support data to the user.
Absstract of: NL4001166A
0001 Abstract The present application discloses a data-driven damage identification and diagnosis method and apparatus for pressure equipment, a device, and a medium, and relates to the field of equipment damage identification. The method 5 includes identifying a damage mode and a damage rate of target pressure equipment based on basic data of the target pressure equipment and a pressure equipment damage prediction model of the same apparatus category as the target pressure equipment. A method for determining the 10 pressure equipment damage prediction model includes: acquiring a sample data set, the sample data set including: basic data of pressure equipment of different apparatus categories and corresponding label data; performing sample data balancing on the sample data set using a synthetic 15 minority over-sampling technique and edited nearest neighbor (SMOTEENN) algorithm to obtain an expanded data set; and performing, based on the expanded data set, optimization training on hyperparameters in a machine learning model.
Absstract of: WO2026218095A1
The described technology pertains to a system for monitoring and controlling industrial or manufacturing processes, including additive manufacturing processes such as laser powder bed fusion. A sensor subsystem captures process data from at least one sensor, and, in some embodiments, contextual process data is also used. A defect monitoring subsystem comprises a sensor fusion module that fuses the captured data into a unified representation using a neural-network-based data pipeline with variational autoencoder models, and a decision-making module that generates a defect classification output and a defect probability score, and employs a defect detection dictionary to generate a defect detection indication. A process control subsystem may dynamically adjust one or more process parameters based on outputs of the defect monitoring subsystem, and a sensor simulation subsystem may validate proposed control parameter changes by generating synthetic sensor signals.
Absstract of: US20260345105A1
A method for intra-operative assessment of a tissue specimen mammogram to predict tumor margin status includes receiving, during breast-conserving surgery, a mammogram of a resected breast tissue specimen. The method further includes providing the mammogram as input to a tumor margin classifier comprising a trained machine learning model. The method further includes GPU MEM producing, based on the input, intra-operatively, in real time, and using the trained machine learning model, a prediction of whether the specimen has a positive tumor margin status or a negative tumor margin status. The method further includes outputting the prediction intra-operatively.
Nº publicación: US20260353119A1 05/11/2026
Applicant:
LG ELECTRONICS INC [KR]
LG ELECTRONICS INC.
Absstract of: US20260353119A1
0000 A method performed by a device supporting artificial intelligence/machine learning (AI/ML) in a wireless communication system, according to at least one of embodiments disclosed in the present specification, comprises: receiving a configuration for an AI/ML model from a network; performing monitoring on performance of the AI/ML model on the basis of outputs from the AI/ML model; and performing AI/ML model management of maintaining the AI/ML model or at least partially changing the AI/ML model on the basis of the monitoring of the performance of the AI/ML model, wherein the monitoring of the performance of the AI/ML model may comprise first monitoring for monitoring performance of one or two or more intermediate outputs obtained before a final output from the AI/ML model, and second monitoring for monitoring performance of the final output obtained on the basis of the one or two or more intermediate outputs.