Alerta

MULTI-DIMENSIONAL RISK OPTIMIZATION OF PREDICTIVE MODELS

Resumen de: US2025190883A1

A computer-implemented method comprising: receiving a set of candidate trained machine learning models and a set of evaluation dimensions; generating risk scores for each of the candidate trained machine learning models over each of the evaluation dimensions; determining correlations between the evaluation dimensions based, at least in part, on the generated risk scores; and performing an optimization calculation to identify a subset of the set of candidate trained machine learning models, wherein each of the candidate trained machine learning models in the subset optimizes an overall risk measure over all of the evaluation dimensions, wherein the optimization calculation is based, at least in part, on the determined correlations.

RAPID DATA ANNOTATION FOR MACHINE LEARNING

Resumen de: US2025190824A1

Systems and Methods are described herein for rapid data annotation for ML. Aspects comprise a method for inferencing with an ensemble of agents (“EoAs”), comprising receiving data for processing by one or more agents of the EoAs; selecting a plurality of Bright pool agents from the EoAs; performing a first inference operation with each Bright pool agent of plurality of Bright pool agents based on the received data to generate a plurality of intermediate outputs; performing ground-truthing on one or more of an intermediate outputs and a final output to generate one or more labeled outputs; and storing the labeled outputs in a data repository.

PROCESSING DATA FOR PREDICTING PIPE FAILURE

Resumen de: US2025190877A1

A system receives and automatically transforms utility pipe attribute data and pipe break data. The missing and/or incorrect entries in the pipe attributes and/or break 5 data is automatically identified and correct values for these entries are is automatically imputed to generate improved datasets of the pipe attribute data and break data. The improved data can be used to build a model with machine learning. Predictions of future likelihood of failure for pipe sections in a network of pipes can be made based on the model. A national database can be created that is filled with environmental data that has been transformed, optimized, merged, and imputed. The national database can be used for many customers to save computational costs. The national database can be used to build the failure prediction model for utility companies thereby saving computational costs.

GENERATING ENCODED TEXT BASED ON SPOKEN UTTERANCES USING MACHINE LEARNING SYSTEMS AND METHODS

Resumen de: US2025190686A1

Systems and methods for generating encoded text representations of spoken utterances are disclosed. Audio data is received for a spoken utterance and analyzed to identify a nonverbal characteristic, such as a sentiment, a speaking rate, or a volume. An encoded text representation of the spoken utterance is generated, comprising a text transcription and a visual representation of the nonverbal characteristic. The visual representation comprises a geometric element, such as a graph or shape, or a variation in a text attribute, such as font, font size, or color. Analysis of the audio data and/or generation of the encoded text representation can be performed using machine learning.

SYSTEMS AND METHODS FOR ARTIFICIAL INTELLIGENCE ANALYSIS OF SECURITY ACCESS DESCRIPTIONS

Resumen de: US2025193247A1

Some implementations described herein relate to a system for artificial intelligence analysis of security access descriptions. The system identifies a security access description. The system determines metadata information associated with the security access description. The system determines, by processing the security access description using a first set of one or more machine learning models, a descriptive quality label associated with the security access description. The system determines, by processing the security access description using a second set of one or more machine learning models, one or more descriptive components associated with the security access description and one or more descriptive component labels that correspond to the one or more descriptive components. The system provides the metadata information, the descriptive quality label, the one or more descriptive components, and/or the one or more descriptive component labels.

SYSTEM AND METHOD FOR CONTACTLESS PREDICTIONS OF VITAL SIGNS, HEALTH RISKS, CARDIOVASCULAR DISEASE RISK AND HYDRATION FROM RAW VIDEOS

Resumen de: US2025185924A1

A system and method for contactless predictions of one of vital signs, health risk for a disease or condition, blood biomarker values, and hydration status, the method executed on one or more processors, the method including: receiving a raw video capturing a human subject; determining one of vital signs, health risk for a disease or condition, blood biomarker values, and hydration status using a trained machine learning model, the machine learning model taking the raw video as input, the machine learning model trained using a plurality of training videos where ground truth values for the vital signs, the health risk for a disease or condition, the blood biomarker values, or the hydration status were known during the capturing of the training video; and outputting the predicted vital signs, health risk for a disease or condition, blood biomarker values, or hydration status.

SYSTEMS AND METHODS FOR COMPREHENSIVE AND STANDARDIZED IMMUNE SYSTEM PHENOTYPING AND AUTOMATED CELL CLASSIFICATION

Resumen de: AU2023389234A1

Method and systems for generating an immune profile for a subject are described. In some instances, the methods comprise contacting at least a first aliquot of a sample from the subject with at least a first immunophenotyping panel to fluorescently-label cells contained within the sample; processing the fluorescently-labeled cells using a full spectrum flow cytometer to generate fluorescence intensity data, or data derived therefrom, for fluorescently-labeled cells from the sample; providing at least a subset of the fluorescence intensity data, or data derived therefrom, for the fluorescently-labeled cells as input to an ensemble machine learning model configured to process the data and classify individual cells as belonging to one of a plurality of distinct immune cell sub-populations; and outputting a total cell count or cell frequency for each of the plurality of distinct immune cell sub-populations in the sample as part of an immune profile for the subject.

ARC FAULT DETECTION USING MACHINE LEARNING

Resumen de: US2025192537A1

In aspects of the present disclosure, a circuit interrupter includes a housing, a conductive path, a switch which selectively interrupts the conductive path, sensor(s), memory, and a controller within the housing. The sensor(s) measure electrical characteristic(s) of the conductive path. The memory stores an arc detection program that implements a machine learning model and includes a field-updatable program portion and a non-field-updatable program portion, where the field-updatable program portion includes program parameters used by the non-field-updatable program portion to decide between presence or absence of an arc fault. The controller executes the arc detection program to compute input data for the machine learning model based on the sensor measurements, decide between presence of an arc event or absence of an arc event based on the input data, and cause the switch to interrupt the conductive path when the decision indicates presence of an arc event.

METHODS AND SYSTEMS OF FACILITATING A HEALTH ASSESSMENT

Resumen de: WO2025120557A1

The present disclosure provides a method of facilitating a health assessment. Further, the method may include receiving a medical inquiry. Further, the medical inquiry may be generated by a user. Further, the method may include analyzing the medical inquiry. Further, the method may include generating an assessment query based on the analyzing. Further, the generation of the assessment query may be based on one or more of a template and a first machine learning model. Further, the template includes a standardized assessment query. Further, the method may include transmitting the assessment query. Further, the method may include receiving a response from the healthcare communication infrastructure. Further, the response corresponds. Further, the method may include analyzing the response. Further, the method may include determining a diagnosis data based on the analyzing. Further, the method may include transmitting the diagnosis data.

HYDRAULIC FRACTURING FRAMEWORK

Resumen de: WO2025122496A1

A method can include receiving input for a group of wells in a subsurface region, where the group of wells defines a hydraulically fractured production unit; predicting production data for the group of wells using a machine learning model; and outputting the predicted production data.

Systems and methods of property valuation

Resumen de: GB2636300A

The disclosure features a method which includes inputting or receiving information on one or more features of a plurality of residential properties and prices of the residential properties including a marketed price, a listing price, and a closing price, providing the information to a Machine Learning Algorithm to determine the relationship between the one or more features and the prices of the residential properties to create a Machine Learned Model, inputting or receiving information on one or more features of a new residential property into the Machine Learned Model, and predicting a base price of the new residential property from the Machine Learned Model based on the one or more features of the new residential property. The disclosure also features one or more non- transitory, computer-readable storage media storing instructions capable of performing the method and a computer or computer system capable of performing the method.

MACHINE LEARNING MODEL CORRECTION

Resumen de: US2024046349A1

A method, in some implementations, may include obtaining output from a machine learning (ML) model responsive to input data, obtaining initial training data representing training data used to train the ML model, generating, based on the output from the ML model and the initial training data, correction training data that represents a desired alteration to the output from the ML model responsive to one or more particular subgroups in the input data, generating, based on the correction training data, a correction ML model configured to receive, as input, the input data and to output correction values which, when combined with the output from the ML model, perform the desired alteration, and generating corrected output as a combination of the output from the ML model and the output correction values from the correction ML model, and providing, for display, the corrected output.

SYSTEMS AND METHODS FOR PROGRAMMATIC LABELING OF TRAINING DATA FOR MACHINE LEARNING MODELS VIA CLUSTERING AND LANGUAGE MODEL PROMPTING

Resumen de: AU2023383086A1

Embodiments introduce an approach to semi-automatically generate labels for data based on implementation of a clustering or language model prompting technique and can be used to implement a form of programmatic labeling to accelerate the development of classifiers and other forms of models. The disclosed methodology is particularly helpful in generating labels or annotations for unstructured data. In some embodiments, the disclosed approach may be used with data in the form of text, images, or other form of unstructured data.

MACHINE LEARNING BASED OCCUPANCY FORECASTING

Resumen de: WO2025117106A1

Embodiments determine a final occupancy prediction for a check-in date for a plurality of hotel rooms. Embodiments receive historical reservation data including a plurality of booking curves for the hotel rooms corresponding to a plurality of reservation windows, the historical reservation data including a plurality of features. Based on the historical reservation data, embodiments generate a first occupancy prediction for the check-in date using a first model and generate a second occupancy prediction for the check-in date using a second model. Embodiments determine a best performing model from at least the first model and the second model uses a corresponding occupancy prediction corresponding to the best performing model as the final occupancy prediction for the check-in date.

METHOD AND SYSTEM FOR PREDICTING ABDOMINAL AORTIC ANEURYSM (AAA) GROWTH

Resumen de: AU2023380279A1

There are provided methods, systems and non-transitory storage mediums for predicting growth of an abdominal aortic aneurysm (AAA) of a patient having been diagnosed with AAA. Segmented regions of interest (ROI) comprising the aorta and adjacent structures are received by segmenting a set of images. A wall shear stress parameter and intraluminal thickness parameter is determined. A 3D parametric mesh comprising a plurality of concentric 3D mesh layers is generated, where each concentric 3D mesh layer includes a same predetermined number of nodes. The generation includes encoding the segmented ROIs, the wall shear stress parameter and the intraluminal thickness parameter as features at respective node locations in the 3D parametric mesh. A trained growth prediction machine learning model predicts, based at least on a subset of features of the 3D parametric mesh, if the given patient will show AAA growth. The training of the growth prediction model is also disclosed.

METHODS, SYSTEMS, AND DEVICES TO VALIDATE IP ADDRESSES

Resumen de: US2025184345A1

Aspects of the subject disclosure may include, for example, obtaining a first group of Internet Protocol (IP) addresses from a group of network devices, and determining a second group of IP addresses from the first group of IP addresses includes possible malicious IP addresses utilizing a machine learning application. Further embodiments can include obtaining a first group of attributes of malicious IP addresses from a first repository, and determining a third group of IP addresses from the second group of IP addresses includes possible malicious IP addresses based on the first group of attributes. Additional embodiments can include receiving user-generated input indicating a fourth group of IP addresses from the third group of IP addresses includes possible malicious IP addresses, and transmitting a notification to a group of communication devices indicating that the fourth group of IP address includes possible malicious IP addresses. Other embodiments are disclosed.

UTILIZING MACHINE LEARNING AND A SMART TRANSACTION CARD TO AUTOMATICALLY IDENTIFY OPTIMAL PRICES AND REBATES FOR ITEMS DURING IN-PERSON SHOPPING

Resumen de: US2025182156A1

A device may receive, from a client device of a customer, item data identifying a price of an item and customer data identifying the customer, where the item data may be received by a transaction card from a price tag of the item. The device may receive price data identifying prices associated with multiple items and other data identifying locations, availabilities, and terms of the multiple items, and may process the item data, the price data, and the other data, with a machine learning model, to identify an optimal price for the item. The device may provide, to the client device, data identifying the optimal price and data identifying a merchant associated with the optimal price, and may receive transaction data identifying the item, the optimal price, and the merchant when the customer purchases the item. The device may perform actions based on the transaction data.

MODULAR MACHINE LEARNING SYSTEMS AND METHODS

Resumen de: US2025181676A1

A computer system is provided that is designed to handle multi-label classification. The computer system includes multiple processing instances that are arranged in a hierarchal manner and execute differently trained classification models. The classification task of one processing instance and the executed model therein may rely on the results of classification performed by another processing instance. Each of the models may be associated with a different threshold value that is used to binarize the probability output from the classification model.

Method And System For Key Predictors And Machine Learning For Configuring Cell Performance

Resumen de: US2025183392A1

A method of managing battery performance may include obtaining, via a measurement device, measurements of one or more parameters relating to one or more cells; generating or updating, based on the measurements, a machine learning model; and generating, using the machine learning model, cell performance prediction data for use in managing at least one cell. Each cell includes a cathode, a separator, and a silicon-dominant anode. The measurements of the one or more parameters correspond to a plurality of different types of data. The measurements include one or more of: measurements of cells or cell components before formation or cycling, measurements from formation cycles for one or more cells, measurements from a number of cycles after formation for one or more cells, and measurements of characteristics of cell components prior to cell assembly.

DYNAMIC FILTER RECOMMENDATIONS

Resumen de: US2025181587A1

A user preference hierarchy is determined from user response to images. Images may be tagged using machine learning models trained to determine values for images. Products are clustered according to product vectors. Images of products within a cluster are clustered according to composition and groups of images are selected from image clusters for soliciting feedback regarding user preference for products of a cluster. Feedback is used to train a user preference model to estimate affinity for a product vector. A user may provide feedback regarding a price point and products are weighted according to a distribution about the price point. The distribution may be asymmetrical according to direction of movement of the price point. Filters may be dynamically defined and presented to a user based on popularity and frequency of occurrence of attribute-value pairs of search results and based on feedback regarding the search results.

FEATURE DIMENSIONALITY REDUCTION FOR MACHINE LEARNING MODELS

Resumen de: US2025181991A1

Provided is a method, system, and computer program product for performing automated feature dimensionality reduction without accuracy loss. A processor may determine a first training value associated with a first dataset of a machine learning model. The processor may rank features of the first dataset in relation to the first training value. The processor may compare the ranked features of the first dataset to a predetermined threshold. The processor may generate a second dataset from the first dataset by removing a third dataset, the third dataset having a set of features that did not meet the predetermined threshold. The processor may determine a second training value associated with the second dataset. The processor may compare the first training value to the second training value. In response to the second training value being lower than the first training value, the processor may analyze the third dataset with a dimensionality reduction algorithm.

METROLOGY AND PROCESS CONTROL FOR SEMICONDUCTOR MANUFACTURING

Resumen de: US2025181941A1

A semiconductor metrology system including a spectrum acquisition tool for collecting, using a first measurement protocol, baseline scatterometric spectra on first semiconductor wafer targets, and for various sources of spectral variability, variability sets of scatterometric spectra on second semiconductor wafer targets, the variability sets embodying the spectral variability, a reference metrology tool for collecting, using a second measurement protocol, parameter values of the first semiconductor wafer targets, and a training unit for training, using the collected spectra and values, a prediction model using machine learning and minimizing an associated loss function incorporating spectral variability terms, the prediction model for predicting values for production semiconductor wafer targets based on their spectra.

TELECOMMUNICATION SYSTEM FAILURES PREDICTION THROUGH MACHINE LEARNING AND ARTIFICIAL INTELLIGENCE

Resumen de: US2025184212A1

In an embodiment, a method may be implemented in a computer system comprising a processor, memory accessible by the processor, and computer program instructions stored in the memory and executable by the processor, the computer system interconnected with a telecommunications system, the method comprising: receiving, at the computer system, data relating to operation of the telecommunication system, obtaining, at the computer system, at least one machine learning model trained to detect and predict faults in the operation of the telecommunication system, selecting, at the computer system, computing infrastructure upon which to execute the at least one machine learning model, wherein the selected computing infrastructure comprises a mesh of interconnected micro-applications;, executing, at the computer system, the at least one machine learning model using the selected computing infrastructure to detect and predict faults in the operation of the telecommunication system, and automatically correcting at least some of the detected faults.

CONCURRENT OPTIMIZATION OF MACHINE LEARNING MODEL PERFORMANCE

Resumen de: US2025181978A1

Certain aspects of the present disclosure provide techniques for concurrently performing inferences using a machine learning model and optimizing parameters used in executing the machine learning model. An example method generally includes receiving a request to perform inferences on a data set using the machine learning model and performance metric targets for performance of the inferences. At least a first inference is performed on the data set using the machine learning model to meet a latency specified for generation of the first inference from receipt of the request. While performing the at least the first inference, operational parameters resulting in inference performance approaching the performance metric targets are identified based on the machine learning model and operational properties of the computing device. The identified operational parameters are applied to performance of subsequent inferences using the machine learning model.

PIPELINED EXECUTION OF GENERATIVE ARTIFICIAL INTELLIGENCE MODELS

Nº publicación: WO2025111787A1 05/06/2025

Solicitante:

QUALCOMM INC [US]
ZHANG YIFEI [CA]
SANG XIAOJUN [CN]
GUPTA SAHIL [US]
ARULESAN VELUPPILLAI [CA]
GEHLHAAR JEFFREY BAGINSKY [US]
WANG JIAN [CA]
PRABHUDESAI PRATHAMESH PRAKASH [CA]
BEYKUN ALEXANDER [CA]
VANGALA PRAVEEN KUMAR [IN]
QUALCOMM INCORPORATED,
ZHANG, Yifei,
SANG, Xiaojun,
GUPTA, Sahil,
ARULESAN, Veluppillai,
GEHLHAAR, Jeffrey Baginsky,
WANG, Jian,
PRABHUDESAI, Prathamesh Prakash,
BEYKUN, Alexander,
VANGALA, Praveen Kumar

Resumen de: WO2025111787A1

Techniques and apparatus for efficiently generating a response to an input query using a generative artificial intelligence model in a pipelined execution environment. An example method generally includes loading a first portion of a machine learning model, wherein the first portion of the machine learning model is associated with a first inference; loading a second portion of the machine learning model, wherein the second portion of the machine learning model is associated with a second inference; and while loading the second portion of the machine learning model, generating the first inference based on an input data set and the first portion of the machine learning model.