Alerta

MITIGATING BIAS IN MACHINE LEARNING WITHOUT POSITIVE OUTCOME RATE REGRESSIONS

Resumen de: WO2025090145A1

A computer obtains multipliers of a sensitive feature. From an input that contains a value of the feature, a probability of a class is inferred. Based on the value of the feature in the input, one of the multipliers of the feature is selected. The multiplier is specific to both of the feature and the value of the feature. The input is classified based on a multiplicative product of the probability of the class and the multiplier that is specific to both of the feature and the value of the feature. In an embodiment, a black-box tri-objective optimizer generates multipliers on a three-way Pareto frontier from which a user may interactively select a combination of multipliers that provides a best three-way tradeoff between fairness and accuracy. The optimizer has three objectives to respectively optimize three distinct validation metrics that may, for example, be accuracy, fairness, and favorable outcome rate decrease.

PREDICTIVE, MACHINE-LEARNING, EVENT-SERIES COMPUTER MODELS WITH ENCODED REPRESENTATION

Resumen de: US2025139528A1

Provided is a process including: obtaining, for a plurality of entities, entity logs, wherein: the entity logs comprise events involving the entities, a first subset of the events are actions by the entities, at least some of the actions by the entities are targeted actions, and the events are labeled according to an ontology of events having a plurality of event types; training, with one or more processors, based on the entity logs, a predictive machine learning model to predict whether an entity characterized by a set of inputs to the model will engage in a targeted action in a given duration of time in the future; and storing the trained predictive machine learning model in memory.

Method and system for diagnosing machine learning model underperformance or failure

Resumen de: GB2634926A

Diagnosing machine learning model underperformance or failures, comprising: embedding validation data samples and text string phrase candidates in a joint embedding space to generate data embeddings and text embeddings S2. A machine learning model trained to perform a number of tasks is received S4 and inference is conducted with the validation dataset using the trained machine learning model to obtain numerical metrics per validation data sample S6. A spatio-temporal relationship representation is generated for the validation data samples S8. A plurality of data slices are generated S10, and a semantic description for each data slice is generated based on the text embeddings S12. The number of tasks may be selected from a group comprising object recognition, object identification, object detection, pose estimation, and semantic segmentation. The spatio-temporal relationship representation may be a graph comprising nodes representing validation data samples, and edges with weights relative to the relationship between data samples.

Machine learning for modeling preference data to optimize interactions with candidate partners

Resumen de: GB2635074A

Technologies are provided for optimizing candidate partners for a user interaction. The technologies can facilitate a trust in facts and identify mutual interest. The technologies can identify the location of users, share personalized information, provide tools for matching users to candidates, exchange data, advertise to users with tracking algorithms, create avatars, host digital interactions between users, and provide user assessments of other users. Nodes of users, and information on user behavior patterns can help identify matches. Users can share their current moods to communicate with other users. Machine learning is included for modeling a user's own feedback from actual interactions in a pool of candidates. The input to the model includes sets of interaction data on users, and the output from the model is an improved, modeled set of user preferences to improve the user's candidate pool. Images of virtual candidates can be created at each iteration of the model.

CLASSIFYING A TYPE OF APPLIANCE ATTACHMENT

Resumen de: WO2025082993A1

The subject-matter of the present disclosure relates to a computer-implemented method of classifying a type of attachment installed on an appliance. The computer-implemented method comprises: sensing (S200), by a sensor (18) of the appliance, data representing physical parameters associated with operating the appliance; classifying (S202), using a machine learning model, a type of attachment installed on the appliance based on the sensed data; and outputting (S204) a signal indicating the type of attachment installed on the appliance based on the detection.

MACHINE LEARNING BASED DEFECT EXAMINATION AND RANKING FOR SEMICONDUCTOR SPECIMENS

Resumen de: US2025124307A1

There is provided a system and method of defect examination on a semiconductor specimen. The method comprises: obtaining an inspection dataset informative of a group of defect candidates and attributes thereof resulting from examining the specimen by an inspection tool; classifying, by a classifier, the group of defect candidates into a plurality of defect classes such that each defect candidate is associated with a respective defect class; and ranking, by a decision model, the group of defect candidates into a total order using a sorting rule. Each defect candidate is associated with a distinct ranking in the total order representative of the likelihood of the defect candidate being a defect of interest (DOI). The decision model is previously trained to learn the sorting rule pertaining to the plurality of defect classes associated with the group of defect candidates and a series of attributes in the inspection data.

ENSEMBLE MACHINE LEARNING SYSTEMS AND METHODS

Resumen de: US2025132051A1

Techniques for responding to a healthcare inquiry from a user are disclosed. In one particular embodiment, the techniques may be realized as a method for responding to a healthcare inquiry from a user, according to a set of instructions stored on a memory of a computing device and executed by a processor of the computing device, the method comprising the steps of: classifying an intent of the user based on the healthcare inquiry; instantiating a conversational engine based on the intent; eliciting, by the conversational engine, information from the user; and presenting one or more medical recommendations to the user based at least in part on the information.

SYSTEM AND METHOD FOR MANAGING AN ON-SENSOR MACHINE LEARNING (ML) MODEL

Resumen de: WO2025083451A1

The present disclosure discloses a system and method for managing an on-sensor machine learning (ML) model is disclosed. The method comprises the step of monitoring performance parameters of plurality of on-sensor ML models present in an industrial plant. The method further comprises the step of detecting a degradation of at least one on-sensor ML model based on the monitored ML model performance parameters of the plurality of on-sensor ML models. The degradation of the at least one on-sensor ML model comprises at least one of: data distribution change, training serving skew, model drift, occurrence of outlier event, and data quality issue. The method finally discloses the step of updating the at least one on-sensor ML model based on the ML model upgradation parameters retrieved from one of the plurality of sources.

SYSTEM AND METHODS FOR HETEROGENEOUS CONFIGURATION OPTIMIZATION FOR DISTRIBUTED SERVERS IN THE CLOUD

Resumen de: US2025130845A1

A system may forecast a workload for a cluster of nodes in a database management system. The system may generate a reconfiguration plan based on the forecasted workload. The system may obtain a heterogenous configuration set. The heterogenous configuration set may include respective configuration sets for the complete sets of nodes. The system may forecast, based on a first machine learning model, respective performance metrics for nodes in each of the complete sets. The system may forecast a cluster performance metric for the entire cluster of nodes based on a second machine learning model. The system may include, in response to satisfaction of an acceptance criterion, the heterogenous configuration set in the reconfiguration plan. The system may cause the cluster of nodes to be reconfigured based on the reconfiguration plan.

AI-POWERED AUTONOMOUS PLANT-GROWTH OPTIMIZATION SYSTEM THAT AUTOMATICALLY ADJUSTS INPUT VARIABLES TO YIELD DESIRED HARVEST TRAITS

Resumen de: US2025131268A1

Inputs from sensors (e.g., image and environmental sensors) are used for real-time optimization of plant growth in indoor farms by adjusting the light provided to the plants and other environmental factors. The sensors use wireless connectivity to create an Internet of Things network. The optimization is determined using machine-learning analysis and image recognition of the plants being grown. Once a machine-learning model has been generated and/or trained in the cloud, the model is deployed to an edge device located at the indoor farm to overcome connectivity issues between the sensors and the cloud. Plants in an indoor farm are continuously monitored and the light energy intensity and spectral output are automatically adjusted to optimal levels at optimal times to create better crops. The methods and systems are self-regulating in that light controls the plant's growth, and the plant's growth in-turn controls the spectral output and intensity of the light.

AGENTIC ARTIFICIAL INTELLIGENCE WITH DOMAIN-SPECIFIC CONTEXT VALIDATION

Resumen de: US2025131028A1

An agent-based website search interface utilizes a multimodal model to enhance enterprise operations. Data agents collect and process diverse inputs, while an orchestrator manages these agents. The system leverages machine learning models to generate insights and automate decision-making processes. It includes tools for data visualization and validation, ensuring accuracy and reliability. By integrating generative AI, the interface provides advanced search functionalities, improving user experience and operational efficiency. This facilitates seamless interaction to answer context specific questions from complex data, offering a robust solution for enterprise-level search and analysis.

Machine Learning Based Overbooking Modeling

Resumen de: US2025131342A1

Embodiments optimize hotel room reservations for a hotel. For a first day of a plurality of future days, embodiments automatically determine, based on an objective function, an overbooking limit for each category of hotel rooms for the hotel, where the hotel includes a plurality of different room categories. Embodiments receive a first reservation request for the first day for a first category room. When the determined overbooking limit for the first category room has not been reached, embodiments accept the first reservation request. When the accepted first reservation request is being checked in to the hotel on the first day, embodiments automatically determine, based on the objective function, to reject the first reservation request, accept the first reservation request, or upgrade the first reservation request to a higher category room.

SYSTEM AND METHOD FOR SELECTING MACHINE LEARNING TRAINING DATA

Resumen de: US2025131336A1

Systems and methods are provided for selecting training examples to increase the efficiency of supervised active machine learning processes. Training examples for presentation to a user may be selected according to measure of the model's uncertainty in labeling the examples. A number of training examples may be selected to increase efficiency between the user and the processing system by selecting the number of training examples to minimize user downtime in the machine learning process.

CROSS-MODEL SCORE NORMALIZATION

Resumen de: US2025131337A1

Computer-implemented techniques encompass using distinct machine learning sub-models to score respective types of candidate content for the purpose of providing personalized content suggestions to end-users of a content management system. The relevancy scores generated by the distinct sub-models are mapped to expected end-user interaction scores of the candidate content scored. Content suggestions are provided at end-users' computing devices where the suggested content is selected from the candidate content based on the expected end-user interaction scores of the candidate content. For each distinct sub-model, a normalizing mapping function is solved using an optimizer that maps the relevancy scores generated by the sub-model for the candidate content to expected end-user interaction scores for the candidate content. The expected end-user interaction scores are comparable across the distinct sub-models and can be used to rank content suggestions across the distinct sub-models.

SMART SKILL COMPETENCY EVALUATION SYSTEM

Resumen de: WO2025085353A1

A computing device and methods of making and using a computing device having machine learning capabilities to analyze course text content based on prompting to generate a list of course learning objectives, and in particular embodiments, having machine learning capabilities to analyze presentation content text against each of the course learning objectives to generate a course competency score with supportive reasoning for each course learning objective and an overall presentation score.

AUTOMATIC WELL TEST VALIDATION

Resumen de: US2025131168A1

A method for validating a well test includes receiving historical well test data. The historical well test data includes one or more accepted flags and one or more rejected flags. The method also includes training a machine-learning (ML) model based upon the historical well test data to produce a trained ML model. The method also includes receiving new well test data. The new well test data does not include the one or more accepted flags and the one or more rejected flags. The method also includes determining whether the new well test data meets or exceeds a predetermined validation threshold using the trained ML model.

CLASSIFYING A TYPE OF APPLIANCE ATTACHMENT

Resumen de: EP4542451A1

The subject-matter of the present disclosure relates to a computer-implemented method of classifying a type of attachment installed on an appliance. The computer-implemented method comprises: sensing (S200), by a sensor (18) of the appliance, data representing physical parameters associated with operating the appliance; classifying (S202), using a machine learning model, a type of attachment installed on the appliance based on the sensed data; and outputting (5204) a signal indicating the type of attachment installed on the appliance based on the detection.

CONFIGURING AN ELECTRONIC DEVICE USING ARTIFICIAL INTELLIGENCE

Resumen de: EP4542395A2

The devices, systems, and methods described herein enable automatically configuring an electronic device using artificial intelligence (AI). The devices, systems, and methods enable accessing telemetry data representing device usage data, inputting the accessed telemetry data into machine learning models that are matched to device metadata, and determining notifications to publish to components of the electronic device. The notifications represent events predicted to occur on the electronic device. The notifications are published to the components of the electronic device such that the electronic device is configured according to the published notifications. The determined notifications enable the identification of optimal settings for the electronic device based on the usage pattern of the device and enable components of the electronic device to preemptively take action on events which are predicted to occur in the future.

METHODS, SYSTEMS, AND COMPUTER READABLE MEDIA FOR USING A MACHINE LEARNING (ML) MODEL IN BATTERY MANAGEMENT

Resumen de: US2025124336A1

One example method for using a machine learning (ML) model in battery management comprises: receiving one or more selection inputs for selecting an ML model for providing battery management information, wherein the one or more selection inputs include a state of health (SOH) value associated with a battery system; selecting, using the selection inputs, the ML model from a plurality of ML models; obtaining, using model inputs and the ML model, the battery management information associated with the battery system; and performing, using the battery management information, a battery management decision for managing the battery system.

MANAGING INFORMATION FOR MODEL TRAINING USING DISTRIBUTED BLOCKCHAIN LEDGER

Resumen de: US2025124311A1

Embodiments are directed to generating and training a distributed machine learning model using data received from a plurality of third parties using a distributed ledger system, such as a blockchain. As each third party submits data suitable for model training, the data submissions are recorded onto the distributed ledger. By traversing the ledger, the learning platform identifies what data has been submitted and by which parties, and trains a model using the submitted data. Each party is also able to remove their data from the learning platform, which is also reflected in the distributed ledger. The distributed ledger thus maintains a record of which parties submitted data, and which parties removed their data from the learning platform, allowing for different third parties to contribute data for model training, while retaining control over their submitted data by being able to remove their data from the learning platform.

MACHINE LEARNING BASED DEFECT EXAMINATION AND RANKING FOR SEMICONDUCTOR SPECIMENS

Resumen de: US2025124307A1

There is provided a system and method of defect examination on a semiconductor specimen. The method comprises: obtaining an inspection dataset informative of a group of defect candidates and attributes thereof resulting from examining the specimen by an inspection tool; classifying, by a classifier, the group of defect candidates into a plurality of defect classes such that each defect candidate is associated with a respective defect class; and ranking, by a decision model, the group of defect candidates into a total order using a sorting rule. Each defect candidate is associated with a distinct ranking in the total order representative of the likelihood of the defect candidate being a defect of interest (DOI). The decision model is previously trained to learn the sorting rule pertaining to the plurality of defect classes associated with the group of defect candidates and a series of attributes in the inspection data.

SYSTEMS AND METHODS FOR DYNAMICALLY GENERATING NEW DATA RULES

Resumen de: US2025124042A1

Systems, apparatuses, methods, and computer program products are disclosed for dynamically generating a new data rule. An example method includes receiving a data rule including an operation to be applied to a data entry and determining regarding whether the data rule belongs to a set of previously-reviewed data rules. The example method further includes, when the data rule does not belong to the set of previously-reviewed data rules, providing the data rule to a user for human validation and receiving human validation. The example method further includes, causing an update, by rule generation circuitry, of an element of a machine learning model, where the element corresponds to the data rule where the update changes a weight value based on the validation and generating the new data rule based on the updated machine learning model. The example method further includes replacing the data rule with the new data rule.

CLASSIFYING AN ENTITY FOR FOLFOX TREATMENT

Resumen de: US2025125034A1

Comprehensive molecular profiling provides a wealth of data concerning the molecular status of patient samples. Such data can be compared to patient response to treatments to identify biomarker signatures that predict response or non-response to such treatments. This approach has been applied to identify biomarker signatures that strongly correlate with response of colorectal cancer patients to FOLFOX. Described herein are data structures, data processing, and machine learning models to predict effectiveness of a treatment for a disease or disorder of a subject having a particular set of biomarkers, as well as an exemplary application of such a model to precision medicine, e.g., to methods for selecting a treatment based on a molecular profile, e.g., a treatment comprising administration of 5-fluorouracil/leucovorin combined with oxaliplatin (FOLFOX) or with irinotecan (FOLFIRI).

HIERARCHICAL MACHINE LEARNING TECHNIQUES FOR IDENTIFYING MOLECULAR CATEGORIES FROM EXPRESSION DATA

Resumen de: US2025125013A1

Described herein in some embodiments is a method comprising: obtaining expression data previously obtained by processing a biological sample obtained from a subject; processing the expression data using a hierarchy of machine learning classifiers corresponding to a hierarchy of molecular categories to obtain machine learning classifier outputs including a first output and a second output, the hierarchy of molecular categories including a parent molecular category and first and second molecular categories that are children of the parent molecular category in the hierarchy of molecular categories, the hierarchy of machine learning classifiers comprising first and second machine learning classifiers corresponding to the first and second molecular categories; and identifying, using at least some of the machine learning classifier outputs including the first output and the second output, at least one candidate molecular category for the biological sample.

MACHINE EVALUATION OF CONTRACT TERMS

Nº publicación: US2025124530A1 17/04/2025

Solicitante:

COUPA SOFTWARE INCORPORATED [US]
Coupa Software Incorporated

Resumen de: US2025124530A1

Embodiments of the present disclosure provide a method that may include defining an object model containing a structural representation of events and artifacts through which contracts are created, changed, and brought to an end. The method may include accessing a machine learning classifier comprising a plurality of rule sets. The method may include applying the plurality of rule sets to one or more words of each corresponding contract document. The method may include linking identified one or more core attributes and one or more words of each corresponding contract document to an applicable object of the object model, determining prevailing terms of each corresponding contract document, and evaluating contract data variables and assigning a contract data risk value to one or more of contract data values. The method may include communicating an alert via email or text message when a contract risk exceeds a threshold value.