Alerta

MACHINE LEARNING BASED OCCUPANCY GRID GENERATION

Resumen de: CN120303583A

In some aspects, a device may receive sensor data associated with a vehicle and a set of frames. The device may aggregate sensor data associated with the set of frames using the first gesture to generate an aggregated frame, where the aggregated frame is associated with the set of cells. The device may obtain an indication of a respective placeholder flag from each cell of the set of cells, where the respective placeholder flag includes a first placeholder flag or a second placeholder flag, and where the set of cells from the set of cells is associated with the first placeholder flag. The device may train a machine learning model using data associated with the aggregated frame to generate a placeholder grid based on a loss function that calculates only losses from respective cells of the set of cells. Numerous other aspects are described.

MULTI-STAGE MACHINE LEARNING MODEL CHAINING

Resumen de: CN120390929A

A skill chain including a set of ML model evaluations is generated, the input is processed with the set of ML model evaluations, and the skill chain is used to eventually generate a model output accordingly. Each ML model evaluates a "model skill" corresponding to the skill chain. The intermediate output generated by the first ML evaluation for the first model skills of the skill chain may then be processed as an input for the second ML evaluation for the second model skills of the skill chain, thereby ultimately generating a model output for the given input. Such a skill chain may include any number of skills according to any of the various structures and do not need to be evaluated using the same ML model.

BENCHMARKING ALGORITHMS FOR DATA QUALITY MONITORING

Resumen de: WO2025221286A1

In a general aspect, benchmarking for data quality monitoring is described. In some embodiments, a system identifies a base data set to be used as input to a machine learning (ML) model. The system generates a modified base data set by causing synthetic anomaly injection operations to be performed on data of the base data set. The system causes the ML model to run, using the base data set as input, to determine a first output of the ML model, and to run, using the modified base data set as input, to determine a second output of the ML model. The system determines a set of performance metrics representing performance of the ML model at detecting data anomalies and outputs a representation of the set of performance metrics.

Benchmarking Algorithms for Data Quality Monitoring

Resumen de: US2025328505A1

In a general aspect, benchmarking for data quality monitoring is described. In some embodiments, a system identifies a base data set to be used as input to a machine learning (ML) model. The system generates a modified base data set by causing synthetic anomaly injection operations to be performed on data of the base data set. The system causes the ML model to run, using the base data set as input, to determine a first output of the ML model, and to run, using the modified base data set as input, to determine a second output of the ML model. The system determines a set of performance metrics representing performance of the ML model at detecting data anomalies and outputs a representation of the set of performance metrics.

MULTISTAGE FEED RANKING SYSTEM WITH METHODOLOGY PROVIDING SCALABLE MULTI-OBJECTIVE MODEL APPROXIMATION

Resumen de: US2025328821A1

Approximating a more complex multi-objective feed item scoring model using a less complex single objective feed item scoring model in a multistage feed ranking system of an online service. The disclosed techniques can facilitate multi-objective optimization for personalizing and ranking feeds including balancing personalizing a feed for viewer experience, downstream professional or social network effects, and upstream effects on content creators. The techniques can approximate the multi-objective model-that uses a rich set of machine learning features for scoring feed items at a second pass ranker in the ranking system-with the more lightweight, single objective model-that uses fewer machine learning features at a first pass ranker in the ranking system. The single objective model can more efficiently score a large set of feed items while maintaining much of the multi-objective model's richness and complexity and with high recall at the second pass ranking stage.

SYSTEMS AND METHODS FOR BATTERY PERFORMANCE PREDICTION

Resumen de: US2025328780A1

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for battery performance prediction. One of the methods includes actions of receiving battery test data of a battery cell. The battery test data includes data of at least one battery cell property of at least two battery tests. Each battery test includes applying pulses on the battery cell during a battery cycle. The battery test data is provided as input to a machine learning system to predict battery cell performance. The machine learning system includes a machine learning model that has been trained using training data includes test data of battery cells that reached respective end of life (EOL) cycles. In response, a prediction result for the battery cell is automatically generated by the machine learning model. The prediction result indicates an EOL cycle of the battery cell. An action is taken based on the prediction result.

IDENTIFICATION AND MITIGATION OF PERFORMANCE ISSUES OF ENTITIES AND AUTOMATED COMPONENTS

Resumen de: US2025328783A1

Systems and methods are described for identifying and resolving performance issues of automated components. The automated components are segmented into groups by applying a K-means clustering algorithm thereto based on segmentation feature values respectively associated therewith, wherein an initial set of centroids for the K-means clustering algorithm is selected by applying a set of context rules to the automated components. Then, for each group, a performance ranking is generated based at least on a set of performance feature values associated with each of the automated components in the group and a feature importance value for each of the performance features. The feature importance values are determined by training a machine learning based classification model to classify automated components into each of the groups, wherein the training is performed based on the respective performance feature values of the automated components and the respective groups to which they were assigned.

BANDWIDTH PREDICTION USING MACHINE LEARNING

Resumen de: WO2025221398A1

Systems, methods, and apparatus, including computer-readable media, for bandwidth prediction using machine learning. In some implementations, a device detects a series of requests for streaming media content. The device generates a set of feature values based on times that the requests for the streaming media content were issued. The device provides the set of feature values as input to a machine learning model that has been trained to predict a time that a future request for media content will be issued. The device receives output of the machine learning model that indicates a predicted time of a subsequent request for the streaming media content or a predicted time to request bandwidth allocation for the subsequent request. Based on the output generated by the machine learning model, the device sends a bandwidth allocation request to allocate bandwidth to transmit data in a wireless network.

GRAPH DATA REPRESENTATION SYSTEMS AND METHODS FOR ELIGIBILITY DETERMINATION AND/OR MONITORING

Resumen de: US2025328787A1

Embodiments of the present disclosure provide methods, apparatus, systems, computing devices, computing entities, and/or the like for processing an inclusion of an entity for an event. In accordance with one embodiment, a method is provided that includes: determining whether a graph representation data object comprises an inbound edge connecting an entity node representing the entity with an event node representing the event; and responsive to determining the graph representation data object comprises the inbound edge, performing an action involving inclusion of the entity for the event. The inbound edge is generated via an inbound edge generator machine learning model configured to: traverse entity and/or inclusion edges of the graph representation data object to identify inclusion and entity edges connected, generate an entity score data object for the entity based at least in part on the inclusion edges, and responsive to the data object satisfying a threshold, generate the inbound edge.

SYSTEMS AND METHODS FOR SELF-LEARNING ARTIFICIAL INTELLIGENCE OF THINGS (AIOT) DEVICES AND SERVICES

Resumen de: US2025328785A1

The invention is generally directed to systems and methods of monitoring or predicting a service event for an industrial asset using an artificial intelligence of things (AIoT) system including an AIoT device, AIoT cloud, and a self-learning AI classification and analytics engine. The device may include one or more sensors and an inference engine for reducing power consumption and detecting anomalies at the edge and sending data associated with anomalies to a signal processor for classification and AI-driven automatic configuration. Classification may be based on narrow-band analysis and/or machine learning models. If an anomaly is detected power may be provided to a communication module to send sensor data to the signal processor for classification and/or further processing. Classifications or determinations made by the signal processor or detected through a work-order system may be used to automatically retrain the inference model on the edge, so that the system is self-learning.

Machine Learning System Using Quantum Computing

Resumen de: US2025328793A1

Methods and systems for training and using a binary classifier implemented using quantum computing techniques are disclosed. The described approach involves deriving, from an input data set, a plurality of training samples, each training sample comprising a data vector having a plurality of features and a class label. Each data vector is processed using a quantum classification process including: encoding the data vector as an Ising Hamiltonian; implementing the Ising Hamiltonian on a set of real or virtual qubits of a quantum processing unit or an emulation thereof to form a quantum system representing the data vector; executing operations on the (emulation of the) quantum processing unit to prepare the ground state of the quantum system; determining one or more properties of the ground state; and identifying one of a set of possible ground states corresponding to the data vector based on the one or more properties. The system then determines, based on the identified ground states and class labels for the training samples, a mapping that maps ground states to class labels. The mapping is stored and used for classifying further data samples.

VALIDATING VECTOR CONSTRAINTS OF OUTPUTS GENERATED BY MACHINE LEARNING MODELS

Resumen de: US2025328822A1

The technology evaluates the compliance of an AI application with predefined vector constraints. The technology employs multiple specialized models trained to identify specific types of non-compliance with the vector constraints within AI-generated responses. One or more models evaluate the existence of certain patterns within responses generated by an AI model by analyzing the representation of the attributes within the responses. Additionally, one or more models can identify vector representations of alphanumeric characters in the AI model's response by assessing the alphanumeric character's proximate locations, frequency, and/or associations with other alphanumeric characters. Moreover, one or more models can determine indicators of vector alignment between the vector representations of the AI model's response and the vector representations of the predetermined characters by measuring differences in the direction or magnitude of the vector representations.

SYSTEMS AND METHODS FOR PREEMPTIVE COMMUNICATION OF ROAD CONDITION DATA

Resumen de: US2025329252A1

A system and method for communicating road condition data. The system and method includes a plurality of inter-changeable housings, including a sensor housing comprising a sensor configured to generate sensor data; a data processing housing comprising a processor configured to receive the sensor data and vehicle-originated data, and apply one or more layers of a machine learning architecture to the sensor data and the vehicle-originated data to generate at least a portion of vehicle instruction data; and a wireless communication housing comprising a wireless interface circuit configured to receive the vehicle-originated data and to transmit the vehicle instruction data generated by the processor.

SYSTEMS AND METHODS FOR BATTERY PERFORMANCE PREDICTION

Resumen de: WO2025221413A1

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for battery performance prediction. One of the methods includes actions of receiving battery test data of a battery cell. The battery test data includes data of at least one battery cell property of at least two battery tests. Each battery test includes applying pulses on the battery cell during a battery cycle. The battery test data is provided as input to a machine learning system to predict battery cell performance. The machine learning system includes a machine learning model that has been trained using training data includes test data of battery cells that reached respective end of life (EOL) cycles. In response, a prediction result for the battery cell is automatically generated by the machine learning model. The prediction result indicates an EOL cycle of the battery cell. An action is taken based on the prediction result.

VALIDATING VECTOR CONSTRAINTS OF OUTPUTS GENERATED BY MACHINE LEARNING MODELS

Resumen de: WO2025221872A1

The technology evaluates the compliance of an AI application with predefined vector constraints. The technology employs multiple specialized models trained to identify specific types of non-compliance with the vector constraints within AI-generated responses. One or more models evaluate the existence of certain patterns within responses generated by an AI model by analyzing the representation of the attributes within the responses. Additionally, one or more models can identify vector representations of alphanumeric characters in the AI model's response by assessing the alphanumeric character's proximate locations, frequency, and/or associations with other alphanumeric characters. Moreover, one or more models can determine indicators of vector alignment between the vector representations of the AI model's response and the vector representations of the predetermined characters by measuring differences in the direction or magnitude of the vector representations.

AGENT PROMPT GENERATION USING TOOL DESCRIPTION EMBEDDINGS

Resumen de: WO2025221523A1

Methods, systems, and apparatuses include receiving, via a conversational interface, user input from a user of an online system. A user input embedding is generated for the user input. A vector store is retrieved including tool description embeddings. A similarity search is performed using the user input embedding and the tool description embeddings. A set of tool descriptions is determined using results of the similarity search. A prompt is generated using the set of tool descriptions and the user input. Machine learning agents are applied to the prompt to cause the machine learning agents to use tools associated with the set of tool descriptions. A response to the prompt is received, from the machine learning agents, in response to the machine learning agents using the tools. An output to the user input based on the response is sent, via the conversational interface, to the user of the online system.

MACHINE LEARNING MODEL ADMINISTRATION AND OPTIMIZATION

Resumen de: US2025190475A1

Systems and methods are configured to generate a set of potential responses to a prompt using one or more data models with data from at least a plurality of data domains of an enterprise information environment that includes access controls. A deterministic response is selected from the set of potential responses based on scoring of the validation data and restricting based on access controls in view profile information associated with the prompt. These enterprise generative AI systems and methods support granular enterprise access controls, privacy, and security requirements. enterprise generative AI providing traceable references and links to source information underlying the generative AI insights. These systems and methods enable dramatically increased utility for enterprise users to information, analyses, and predictive analytics associated with and derived from a combination of enterprise and external information systems.

MITIGATING TEMPORAL GENERALIZATION FOR A MACHINE LEARNING MODEL

Resumen de: US2025322342A1

Mitigation of temporal generalization losses a target machine learning model is disclosed. Mitigation can be based on identifying, removing, modifying, transforming, etc., features, explanatory variables, models, etc., that can have an unstable relationship with a target outcome over time. Implementation of a more stable representation can be initiated. Temporal stability measures (TSMs) for one or more model feature(s) can be determined based on one or more variable performance metrics (VPMs). A group of one or more VPMs can be selected based on features of a model in either a development or production environment. Model feature modification can be recommended based on a TSM, which can prune a feature, transform a feature, add a feature, etc. Temporal stability information can be presented, e.g., via a dashboard-type user interface. Models can be updated based on mutations of a model comprising a feature modification(s), including competitive champion/challenger model updating.

USING MACHINE LEARNING TO PREDICT CELL THERAPY CHARACTERISTICS

Resumen de: WO2025217397A1

Disclosed are systems and methods for improving processes for developing cell therapies by applying machine learning to data including manufacturing process data and clinical measurements (e.g., patient response and treatment data) to determine parameters and settings for a manufacturing process for engineering cells for use in cell therapy. Parameters and settings for a manufacturing process for genetically engineered T-cells including, but not limited to, Chimeric Antigen Receptor (CAR) T cells can be determined. A method can include receiving a set of process parameters of a cell engineering process, predicting a clinical response associated with an output of the cell engineering process by applying a machine learning model on the received set of process parameters, where the machine learning model is trained on process parameter data and clinical response data, and generating a visualization for use in a graphical user interface of the predicted clinical response.

SYSTEM FOR OPTIMAL DECISION-MAKING AND METHODS THEREOF

Resumen de: WO2025215419A1

The present disclosure provides a system and method for optimal decision-making in multi-criteria decision-making (MCDM) problems. The invention addresses limitations of conventional approaches, which rely heavily on subjective expert inputs and biased preprocessing techniques, by introducing a statistically driven framework based on distribution normalization and data-driven weight assignment. The system comprises modules for preprocessing, evaluation, assessment, and output generation, wherein input data is normalized, criteria constraints inverted where necessary, and statistical weights optimally assigned. Decision alternatives are then computed, evaluated, and ranked to derive one or more optimal decisions. This framework ensures unbiased, efficient, and replicable outcomes across applications including Geographic Information Systems (GIS), Data Analysis, Artificial Intelligence, and Machine Learning.

SYSTEMS AND METHODS FOR MACHINE LEARNING-BASED SITE-SPECIFIC THREAT MODELING AND THREAT DETECTION

Resumen de: US2025322269A1

Systems and methods for implementing a threat model that classifies contextual events as threats. The method can include: accessing a threat model; identifying a set of contextual events, wherein each contextual event comprises a set of semantic primitives predicted from a plurality of sensor streams; and determining a threat level for each contextual event based on threat probabilities.

REDUCED POWER MACHINE LEARNING SYSTEM FOR ARRHYTHMIA DETECTION

Resumen de: US2025322958A1

Techniques are disclosed for using feature delineation to reduce the impact of machine learning cardiac arrhythmia detection on power consumption of medical devices. In one example, a medical device performs feature-based delineation of cardiac electrogram data sensed from a patient to obtain cardiac features indicative of an episode of arrhythmia in the patient. The medical device determines whether the cardiac features satisfy threshold criteria for application of a machine learning model for verifying the feature-based delineation of the cardiac electrogram data. In response to determining that the cardiac features satisfy the threshold criteria, the medical device applies the machine learning model to the sensed cardiac electrogram data to verify that the episode of arrhythmia has occurred or determine a classification of the episode of arrhythmia.

SENSOR READING CORRECTION

Resumen de: AU2024243389A1

Disclosed are systems, methods, and devices for correcting or otherwise cleaning sensor data. Sensor readings and metadata or other information about the sensor readings can be collected, and one or more detection rules (e.g., machine learning models or other detection rules) can be automatically generated for modifying subsequent sensor data. Sensor readings can be refined or supplemented by applying applicable detection rules.

SYSTEM AND METHOD OF DYNAMICALLY RECOMMENDING ONLINE ACTIONS

Resumen de: US2025322366A1

The present disclosure generally relates to a computer device, method and system utilizing machine learning for capturing and analyzing profile data communicated across a computing environment including but not limited to: each user's profile, online behaviors and career progression path and provides dynamic recommendations of online actions to be performed to reach a desired target state.

SYSTEMS AND METHODS FOR LEAKAGE DETECTION, PREVENTION, AND MITIGATION

Nº publicación: WO2025215513A2 16/10/2025

Solicitante:

KOTLEAK LTD [IL]
EDLITZ YOCHAI [IL]
KOTLER IDO [IL]
SWED ELI [IL]
KOTLEAK LTD,
EDLITZ, Yochai,
KOTLER, Ido,
SWED, Eli

Resumen de: WO2025215513A2

Disclosed embodiments relate to systems and methods for acoustically detecting leakage of a fluid using one or more acoustic sensors. Techniques include receiving a signal from the one or more acoustic sensors; performing pre-processing on the signal; inputting the pre-processed signal to a machine learning algorithm; receiving, based on the pre-processed signal and the machine learning algorithm, a classification of the pre- processed signal, the classification being associated with an acoustic profile of leakage of a fluid; and providing a prompt associated with the classification to a user device.