Alerta

FRAMEWORK FOR FOCUSED TRAINING OF LANGUAGE MODELS AND TECHNIQUES FOR END-TO-END HYPERTUNING OF THE FRAMEWORK

Absstract of: US2025218428A1

Techniques are disclosed herein for focused training of language models and end-to-end hypertuning of the framework. In one aspect, a method is provided that includes obtaining a machine learning model pre-trained for language modeling, and post-training the machine learning model for various tasks to generate a focused machine learning model. The post-training includes: (i) training the machine learning model on an unlabeled set of training data pertaining to a task that the machine learning model was pre-trained for as part of the language modeling, and the unlabeled set of training data is obtained with respect to a target domain, a target task, or a target language, and (ii) training the machine learning model on a labeled set of training data that pertains to another task that is an auxiliary task related to a downstream task to be performed using the machine learning model or output from the machine learning model.

COMPUTER-READABLE RECORDING MEDIUM STORING MACHINE LEARNING PROGRAM, COMPUTER-READABLE RECORDING MEDIUM STORING DETERMINATION PROGRAM, AND MACHINE LEARNING DEVICE

Absstract of: US2025217435A1

A non-transitory computer-readable recording medium storing a machine learning program for causing a computer to execute a process including training a machine learning model by machine learning that uses a cost function in which each element of a matrix obtained by relaxing a discrete variable to be optimized to a continuous matrix becomes a discrete optimization problem as a cost function in a search process that performs a search by adopting continuous relaxation into the discrete optimization problem.

BIG DATA INTELLIGENT DECISION ANALYSIS METHOD AND SYSTEM BASED ON MACHINE LEARNING

Absstract of: WO2025139187A1

The present invention relates to the technical field of data analysis, and specifically relates to a big data intelligent decision analysis method and system based on machine learning. The system comprises an interaction layer, an analysis layer, and a decision layer. Real-time position information of users and building parameters of an area where the users are located are captured by the interaction layer, and seismic information is synchronously collected by the interaction layer in real time; and the analysis layer receives the seismic information collected by the interaction layer in real time. In the present invention, on the basis of the real-time position information of the users and the seismic information, a better earthquake risk mitigation and maintenance effect is brought to the users, so as to ensure that an unorganized user population, when an earthquake strikes, can more orderly evacuate from a building on the basis of data provided by the system, thereby effectively improving the escape probability of indoor users when an earthquake strikes; and moreover, when indoor users cannot escape, data support is provided for rescue personnel in a position information feedback mode, so that rescue work can be carried out more efficiently, and a safety guarantee is further provided for the users.

DATA PRIVACY PRESERVATION IN MACHINE LEARNING TRAINING

Absstract of: WO2025144544A1

A first computing system includes a data store with a sensitive dataset. The first computing system uses a feature extraction tool to perform a statistical analysis of the dataset to generate feature description data to describe a set of features within the dataset. A second computing system is coupled to the first computing system and does not have access to the dataset. The second computing system uses a data synthesizer to receive the feature description data and generate a synthetic dataset that models the dataset and includes the set of features. The second computing system trains a machine learning model with the synthetic data set and provides the trained machine learning model to the first computing system for use with data from the data store as an input.

SYSTEMS AND METHODS FOR DETERMINING LIFETIME VALUE OF WEBSITE VISITOR THROUGH MACHINE LEARNING

Absstract of: US2025217826A1

A vehicle data system receives a lead submission through a website supported by the vehicle data system and determines, utilizing a machine learning model, a user value for a lead associated with the lead submission. The user value represents a probability of the lead purchasing a vehicle from a dealer through the website. The vehicle data system determines a user lifetime value for the lead based at least on the user value for the lead. Subsequently, the vehicle data system obtains clickstream identifiers from a search engine and assigns a corresponding user lifetime value to each clickstream identifier. The vehicle data system aggregates the clickstream identifiers and corresponding user lifetime values in a single file and communicates the single file to a search server for consumption. The user lifetime values are utilized by the search engine in search engine marketing processes.

System and Method for Providing Advisory Notifications to Mobile Applications

Absstract of: US2025217175A1

A system and method are provided for providing advisory notifications to mobile applications. The method includes interfacing the server device with at least one endpoint within an enterprise system and storing a model trained by a machine learning engine to automatically determine advisory notifications relevant to client data sets stored by the endpoint(s) and/or the at least one endpoint. The method also includes determining a current state of a client account, using the model to determine an advisory notification for the client account based on the current state, referring to a set of rules to determine when to provide the advisory notification in the mobile application, and in what portion of the mobile application to display the notification; and sending the advisory notification via the communications module to a client device to display the advisory notification in the mobile application.

MACHINE-LEARNING MODELS FOR TUMOR GRADING USING RANK-AWARE CONTEXTUAL REASONING ON WHOLE SLIDE IMAGES

Absstract of: WO2025145040A1

This specification discloses systems, methods, and techniques for classifying a grade of cancer represented in a tumor tissue sample. The techniques include obtaining a whole-slide image (WSI) of the tumor tissue sample and partitioning the WSI into a collection of patches. For each patch, (i) a contextualized feature representation of the patch is generated based on intrinsic features of the patch, local dependencies between the patch and a subset of local patches of the WSI, and non-local dependencies between the patch and a subset of non-local patches of the WSI; and (ii) an attention weight is determined for the patch. A WSI-level cancer grade for the tumor tissue sample is predicted based on the contextualized feature representations and the attention weights for the plurality of patches.

SYSTEM AND METHOD FOR CONSTRUCTING DIGITAL DOCUMENTS

Absstract of: US2025217579A1

In some aspects described herein, a computer-based system that is capable of constructing digital documents is provided. In some implementations, a machine learning system is provided that learns certain terms within a document. The terms may be, for example, part of a document that forms a legally-binding contract between two entities. In one implementation of the machine learning system, the machine learning system interoperates within a user interface to show predictions of certain terms within the document to the user. Further, the machine learning system may capture user answers relating to certain terms and provide feedback into the system that learns during operation of the system, improving user interactions, accuracy and reducing the number of user interactions.

MACHINE LEARNING PROGRAM, DETERMINATION PROGRAM, MACHINE LEARNING METHOD, DETERMINATION METHOD, MACHINE LEARNING DEVICE, AND DETERMINATION DEVICE

Absstract of: EP4579535A1

A machine learning program for causing a computer to execute a process includes training a machine learning model by machine learning that uses a cost function in which each element of a matrix obtained by relaxing a discrete variable to be optimized to a continuous matrix becomes a discrete optimization problem as a cost function in a search process that performs a search by adopting continuous relaxation into the discrete optimization problem.

REAL-TIME CONTROL OF OFF-LINING OF LOCOMOTIVES FOR ENERGY MANAGEMENT

Absstract of: AU2023395626A1

A method of controlling one or more locomotives in a train (102) includes using a machine learning engine and a virtual system modeling engine (324) to model and classify sections of track (106) along which the train (102) is traveling according to the tractive power needs for the train traversing each section of track (106) as a function of an effective weight profile for the train in the section and an effective friction profile for the train in the section of track. The method includes using the results of the effective weight profile, the effective friction profile, and an effective power availability profile to train the virtual system modeling engine (324) using the machine learning engine to model designated areas of the track where the total tractive effort force or dynamic braking force applied by all of the locomotives in the train is less than a tractive effort force or dynamic braking force, respectively, that can be provided by a subset of the available locomotives in the train (102).

SYSTEMS AND METHODS FOR FACILITATING CONVERSATION BASED ON TEXT DATA AND SPEECH DATA

Absstract of: WO2025133181A1

A method includes receiving a user input and generating a set of user input tokens based on the user input. The method also includes generating a set of enhanced input tokens by providing the set of user input tokens as input to a first machine learning model. A state is determined based on a previous state and at least one of the set of user input tokens or the set of enhanced input tokens. Predetermined data is retrieved from a database based on the state and at least one of the set of user input tokens or the set of enhanced input tokens. The method also includes generating a set of response tokens by providing the set of user input tokens and the predetermined data as input to a second machine learning model. Based on the set of response tokens, a response is sent to a user device.

SYSTEM AND METHOD OF HIERARCHICAL INFERENCING IN MACHINE LEARNING SYSTEMS

Absstract of: WO2025137194A1

A method of hierarchical inferencing which may include: by a camera, capturing an image of a scene; by at least one computing device: based on a first reduced resolution image of the scene, determining a first probability value of a presence of a specified feature in the reduced resolution image; if the first probability value is greater than a first probability threshold: based on the image or a second reduced resolution image of the scene, determining a second probability value of a presence of the specified feature in the image or the second resolution image; and if the second probability value is greater than a second probability threshold which is greater than the first probability threshold, transmitting a notification that the specified feature is present in the image or the second resolution image.

SYSTEMS AND METHODS FOR FEDERATED LEARNING

Absstract of: US2025209383A1

In one embodiment, one or more computing systems may generate a knowledge graph representing relationships between a global model and a number of local models for updating the global model. Each local model may have access to a local dataset for machine-learning training. The systems may access, from the knowledge graph, a sharing policy associated with the local dataset of a first local model. The systems may determine, based on the knowledge graph and one or more pre-determined criteria, that the sharing policy permits the local dataset of the first local model to be shared with a second local model. The system may cause the second local model to be trained using at least the local dataset of the first local model and the local dataset of the second local model. The system may update the global model using the trained second local model.

COMPUTER IMPLEMENTED METHOD FOR DEFECT RECOGNITION IN AN IMAGING DATASET OF A WAFER, CORRESPONDING COMPUTER READABLE-MEDIUM, COMPUTER PROGRAM PRODUCT AND SYSTEMS MAKING USE OF SUCH METHODS

Absstract of: US2025209603A1

A computer implemented method for defect recognition in an imaging dataset of a wafer in a charged particle beam system comprising an embedded system, the method comprising: i) obtaining an imaging dataset of a wafer; ii) obtaining model data for a model architecture of a machine learning model for defect recognition in the imaging dataset of the wafer, the model architecture being implemented in the embedded system; iii) transferring the model data to a programmable memory of the embedded system; and iv) applying the machine learning model to an imaging dataset of a wafer to recognize defects, comprising executing the embedded system implemented model architecture with the transferred model data.

Knowledge edit in a text-to-image model

Absstract of: GB2636608A

Identifying locations within an text to image machine generative AI model that support the model’s ability to generate a specific visual attribute in the image, editing the generative artificial intelligence model to cause removal of the trained ability, and generating a subsequent digital image with the model in which the visual attribute is removed. Visual attribute may be object, style, colour, viewpoint, action, concept, etc. A generative text-to-speech model is corrupted by adding noise, preferably by applying Gaussian noise to a symmetric encoder-decoder machine learning model or to a text encoder machine learning model, then parts of the model are restored, the partially restored model is used, and the result is compared with the original model. Differences in the image, i.e. visual attribute changes, are detected in the comparison and based on the attributes that are the same, the functionality of the restored parts of the model is associated with the attributes. Parts if the model that are associated with attributes can be selectively edited out of the model to tailor the image generation of the model by removal of visual attributes.

DYNAMIC GENERATION ON ENTERPRISE ARCHITECTURES USING CAPACITY-BASED PROVISIONS

Absstract of: EP4576707A2

The present disclosure is directed to systems and methods for generating an enterprise architecture for an enterprise network. As one example, a method may include: receiving historical information from a plurality of enterprise networks, the historical information comprising information about an enterprise architecture of each of the enterprise networks; analyzing the historical information from the plurality of enterprise networks to generate a network health score for each of the plurality of enterprise networks; training a machine learning model using a plurality of machine learning algorithms based on the historical information and the network health score of each the plurality of enterprise networks; and generating, using the machine learning model, an enterprise architecture for a first enterprise network, the first enterprise network being a new enterprise network or an existing enterprise network from among the plurality of enterprise networks.

METHOD AND APPARATUS FOR PROCESSING MODEL GENERATION RESULT, ELECTRONIC DEVICE AND STORAGE MEDIUM

Absstract of: EP4575884A1

The present disclosure provides a method and apparatus for processing a model generation result, an electronic device and a storage medium, and relates to the field of artificial intelligence technologies, such as machine learning technologies, natural language processing technologies. An implementation includes: disassembling a text generation result of a generative large model to obtain a plurality of result logic units; wherein each result logic unit includes a segment in the text generation result; each segment is capable of independently identifying one premise or conclusion in a logical inference relationship of the text generation result; and the text generation result is a response result generated by the generative large model based on text input information; generating a logical inference graph capable of characterizing a logical inference relationship among the plurality of result logic units based on the plurality of result logic units; and determining whether logical inference of generation of the text generation result by the generative large model is correct or not based on the logical inference graph. With the technology according to the present disclosure, whether logical inference of the generative large model is correct or not can be efficiently and accurately determined.

AUTOMATIC DETERMINATION OF STIMULATION SETTINGS AND METRICS

Absstract of: WO2025128472A1

A system and method for automated neurostimulation programming uses machine learning to determine target stimulation or discard suboptimal parameters based on actual or anticipated clinical effects. The system includes processors to generate a database of previously acquired lead and/or brain images, tested stimulation settings and associated clinical effects. A region of interest surrounding at least one stimulation lead is identified. Stimulation settings linked to specific clinical effects, including beneficial or detrimental effects, are determined for the region of interest. A trained machine learning model identifies desirable, undesirable, target, or other stimulation parameter values for a new patient by classifying or regressing raw imaging data, from the region of interest. The model is trained to recommend settings that produce desired clinical effects. An output provides the target stimulation parameter values predicted to maximize therapeutic benefit and minimize side effects. This enables automated programming of neurostimulation devices through data-driven machine learning based on clinical outcomes.

DIGITAL TWIN GENERATION AND LOGGING FOR A VEHICLE

Absstract of: US2025201030A1

The present disclosure generates a digital twin of the interior of a vehicle to initiate and track maintenance issues. In one aspect, the digital twin is formed using multiple captured images of the interior of the vehicle where multiple components in those images are identified using a machine learning (ML) model. The components identified by the ML model are then mapped to a model (e.g., a 3D model) of the components that lists their location in the vehicle and an identifier (e.g., a part number or serial number). In this manner, the digital twin can identify, using the identifiers, the various components in the images captured by a camera.

DETECTING SYNTHETIC USER ACCOUNTS USING SYNTHETIC PATTERNS LEARNED VIA MACHINE LEARNING

Absstract of: US2025200193A1

The present disclosure relates to systems, methods, and non-transitory computer-readable media that detect synthetic user accounts of a digital system via machine learning. For instance, the disclosed systems can utilize a machine learning model to analyze account features that are related to a user account and generate an indication that the user account is synthetic based on the analysis. The disclosed systems can further disable (e.g., suspend or close) the user account based on determining that the user account is synthetic. In some cases, the machine learning model provides a precision score that indicates a likelihood that the user account is synthetic, and the disclosed systems disable the user account if the precision score satisfies a threshold. In some implementations, the disclosed systems generate the machine learning model using synthetic user accounts detected via one or more rules and other user accounts that are associated with those synthetic user accounts.

INFORMATION PROCESSING DEVICE, INFORMATION PROCESSING METHOD, AND PROGRAM

Absstract of: EP4571601A1

A new and improved technology capable of further improving inference accuracy while securing security of a machine learning model by federated learning is proposed.Provided is an information processing apparatus including: a learning unit that learns an inference model by federated learning; and an acquisition unit that acquires, from a plurality of terminals, privacy-protected data obtained by executing privacy protection processing on local data obtained by each of the plurality of terminals, in which the learning unit is configured to: perform learning of the inference model on the basis of the privacy-protected data; and distribute information regarding the inference model including a hyperparameter of the inference model on the basis of a result of the learning to the plurality of terminals, the acquisition unit is configured to acquire, from the plurality of terminals, update information of the inference model obtained by performing the learning of the inference model using the local data as learning data, and the learning unit is configured to update the inference model using the update information.

BATCH SELECTION POLICIES FOR TRAINING MACHINE LEARNING MODELS USING ACTIVE LEARNING

Absstract of: EP4571591A1

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training a machine learning model. In one aspect, a method comprises: generating a set of candidate batches of model inputs; generating, for each candidate batch of model inputs, a respective score for the candidate batch of model inputs that characterizes: (i) an uncertainty of the machine learning model in generating predicted labels for the model inputs in the candidate batch of model inputs, and (ii) a diversity of the model inputs in the candidate batch of model inputs; and selecting the current batch of model inputs from the set of candidate batches of model inputs based on the scores; and training the machine learning model on at least the current batch of model inputs.

COMPUTER-BASED SYSTEMS AND METHODS FOR MACHINE LEARNING BASED EXCEPTION PREDICTIONS

Absstract of: WO2025120589A1

A failure prediction method including a predicting flow and a model training flow, the predicting flow including receiving a natural language input from a client computer, translating the input into a task by a LLM, selecting a ML model dedicated to the task, receiving first data, converting the first data to second data of a predetermined format, immediately applying, the ML model on the second data for predicting an output and providing a corresponding explanation, storing the second data and the output into historical data in a storage layer, translating the output and the explanation into a prediction in the natural language by the LLM, and transmitting the prediction to the client computer and iterating the predicting flow for a predetermined number of time; and the model training flow including retrieving the historical data from the storage layer, and training the ML model on the historical data.

ARC FAULT DETECTION USING MACHINE LEARNING

Absstract of: 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.

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

Nº publicación: US2025185924A1 12/06/2025

Applicant:

NURALOGIX CORP [CA]
NURALOGIX CORPORATION

Absstract of: 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.