155 results about "Learning data" patented technology

The present application belongs to the technical field of hydrological forecasting, and specifically relates to a multi-source adaptive transfer learning data-scarce watershed daily runoff prediction method, which comprises the following steps: selecting multiple source watersheds with long sequence observation data, independently training a basic prediction model based on a Transformer architecture by using historical hydrological and meteorological data of the source watersheds, and constructing a source watershed prior knowledge base containing diversified production and confluence mechanisms; adopting a parameter freezing and differential fine-tuning strategy, transferring the source watershed basic model to a data-scarce target watershed, fine-tuning the output layer by using limited measured daily runoff samples of the target watershed on the basis of keeping the parameters of the Transformer feature extraction layer fixed, and constructing multiple transfer branch models; establishing a dynamic attention module to realize real-time sensing of the evolution of meteorological environmental elements of the target watershed and self-adaptive calculation of the dynamic confidence weight of each transfer branch model at the current time; and based on the weight, weighting and integrating the preliminary prediction values of each branch model to obtain the final daily runoff prediction result of the target watershed.

The application provides an intelligent review reminding method and system fusing a knowledge graph and a forgetting curve, and belongs to the technical field of intelligent learning. The intelligent review reminding method comprises the following steps: collecting learning data information of a user from a target data source, and preprocessing the learning data information to obtain preprocessed learning data; performing edge weight dynamic updating processing on a knowledge graph that has been completed and constructed by using the preprocessed learning data, and obtaining an updated knowledge graph; determining actual forgetting rates of each knowledge point of the user by using the updated knowledge graph, and forming a forgetting curve corresponding to the knowledge point according to the actual forgetting rates of each knowledge point of the user; setting a knowledge point review reminding priority ranking list and an associated review data package according to the forgetting curve corresponding to the knowledge point of the user, and pushing the knowledge point and the associated review data package to the user according to the knowledge point review reminding priority ranking list, so that review reminding is realized. The system comprises modules corresponding to the method steps.

A system capable of optimizing early alarm operation by introducing an alarm filtering technology so as not to cause false alarm to an abnormal signal due to signal noise is disclosed. The system comprises: a data acquisition unit for selectively acquiring historical data according to the operation of a electric power facility by selecting a preset multivariate input tag; a recommendation unit for generating learning data, which is a multivariate signal for machine learning, based on a preset similarity from the historical data; a compression unit for generating re-sampling data by compressing the learning data; a learning unit for performing real-time monitoring by using the re-sampling data as input data of a pre-designed early warning learning model; and a calculation unit for generating a final warning when a predetermined condition predefined through the real-time monitoring is satisfied.

This art asset value evaluation system using artificial intelligence may comprise: a memory for storing instructions; and a processor. When executed by the processor, the instructions can instruct the system to: collect, from a plurality of data sources, artwork-related data including artist information, artwork characteristic information and market transaction information; normalize the collected data to generate learning data; use the learning data to train the current value prediction model, a price change prediction model, and an investment value evaluation model; use the trained models to calculate the current market value, future price change prediction, and a long-term investment value of a specific artwork; and integrate the calculated results to generate a final evaluation result.

An apparatus and method for efficiently performing efficient data storage and data transfer of machine learning data. In various implementations, a host processing circuit of a computing system executes a machine learning (ML) application. The application includes a computational graph that indicates the computational order of the ML nodes, layers, and stages of the ML model. The host processing circuit translates function calls in the application to commands particular to an accelerator circuit. The accelerator circuit preloads weights to be used by ML nodes of the ML model by retrieving compressed weights from a storage device different from system memory. The accelerator circuit uses a streaming application programming interface (API) and bypasses the host processing circuit to retrieve the compressed weights from the storage device. The accelerator circuit decompresses the retrieved weights and executes the ML node using the decompressed weights.

This contributes to the realization of a dialogue system that makes it easier for users to obtain the answers they expect. [Solution] The learning system (1) comprises an acquisition means (11) for acquiring driver information relating to the actions of the vehicle driver, a generation means (12) for generating learning data based on the driver information, and a learning means (13) for fine-tuning a base model relating to a dialogue system that performs voice-based dialogue with the driver using the learning data.

This invention relates to an industrial shop floor machine learning system, comprising: a machine learning model providing machine learning data; an industrial shop floor providing shop floor data; and an abstraction layer connecting the machine learning model and the industrial shop floor, wherein the abstraction layer is configured to provide standardized communication between the machine learning model and the industrial shop floor using a machine learning markup language.

In various examples, machine learning data mining for autonomous or semi-autonomous systems and applications is described herein. Systems and methods are disclosed that use neural networks to perform one or more data mining processes. For instance, a first neural network(s) may process input data (e.g., image data) to remove data samples (e.g., images) that are associated with a first object classification(s) and / or a second neural network(s) may process the input data to retrieve data samples (e.g., images) that are associated with a second classification(s). Next, a third neural network(s) may process filtered input data (e.g., the input data not removed by the first neural network(s) and / or the input data retrieved by the second neural network(s)) to determine uncertainty classifications associated with the data samples and a fourth neural network(s) may process the filtered input data to determine final object classifications associated with the data samples.

The information processing device acquires learning data, which includes: structural information related to the structure of the system, required performance information related to the action of the object device, setting information related to the setting of the evaluation function, calculation method selection information regarding the selection of a parameter calculation method from multiple parameter calculation methods, and evaluation information related to the evaluation results of the servo motor's driving performance on the object device. By using machine learning with multiple learning data, an estimation model is generated that estimates at least one of the optimal setting of the evaluation function and the optimal selection of the parameter calculation method based on the structural information and required performance information.

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.

Increase the amount of learning data about users based on their behavioral history. [Solution] The information processing device (10, 10a) includes an input means (111) for inputting a prompt to a first large-scale language model that includes historical information relating to the user's behavior history and facility information relating to one or more facilities; an acquisition means (112) for obtaining a response from the first large-scale language model indicating whether or not the user will visit one or more facilities; and a generation means (113) for generating training data for machine learning of a model different from the first large-scale language model based on the historical information and the response.

An information providing apparatus includes a debate unit and a judging unit. The debate unit performs a debate regarding a legal case by a plurality of lawyer accounts that make respectively different arguments, the lawyer accounts being set in a language model that is generated by learning data stored in a database regarding legal cases and that generates a response to an input prompt. The judging unit judges superiority or inferiority of the arguments based on a result of the debate by the debate unit, by a judging account set in the language model.

To improve prediction accuracy of a model. The present disclosure provides a model creation method executed by an information processing apparatus that creates a model to associate input data included in learning data with output data. The model creation method includes a first exclusion process of excluding the learning data in which the input data is an outlier, from a learning data set; a second exclusion process of excluding the learning data in which the input data and the output data are outliers, from the learning data set; a first model creation process of creating a first model using the learning data set from which the learning data is excluded in the first exclusion process; a second model creation process of creating a second model using the learning data set from which the learning data is excluded in the second exclusion process; and a model selection process of adopting the model having better prediction accuracy between the first model and the second model.

The present disclosure relates to a user learning data prediction method, device, medium, equipment and program product, comprising: obtaining a learning data time sequence, the learning data time sequence comprising actual data of a target analysis dimension at a plurality of historical time nodes; generating an interval grey number sequence according to the learning data time sequence, the interval grey number sequence comprising interval grey numbers of the target analysis dimension at the plurality of historical time nodes, the interval grey numbers being determined according to corresponding actual data; and predicting the target analysis dimension of a future time node according to the interval grey number sequence to obtain a prediction interval of the target analysis dimension. The present disclosure represents the historical learning data of a user as a learning data time sequence, represents the uncertainty variable of the target analysis dimension by an interval grey number to obtain an interval grey number sequence, and thus predicts the target analysis dimension of a future time node, so that the prediction interval of the target analysis dimension is more in line with the small sample and uncertainty characteristics of individual learning data.

The application provides a power edge lightweight privacy protection method and system supporting a national encryption algorithm, wherein the national encryption algorithm supported by a power edge node is used for initializing a random number generator, ensuring that the randomness of a full homomorphic encryption algorithm calculation bottom layer meets the national encryption standard, realizing the compatibility of the two, guaranteeing the safe and trusted transmission of federal learning data, introducing a mean square error gradient can avoid the ciphertext domain nonlinear calculation of the full homomorphic encryption algorithm, adapting to the situation that the power edge node resources are limited, and introducing a dynamic differential privacy noise mechanism considering annealing attenuation and node cooperative regulation can avoid the inverse decryption analysis of the initiator on the global characteristics of the data, realizing the blocking of data reconstruction attacks while guaranteeing the precision and convergence efficiency of the local model, so as to better support the calculation demand of the power edge scene.

The application provides a multi-connected air conditioner, an expansion valve control method and device thereof; in the expansion valve control method of the multi-connected air conditioner, learning data is obtained by clustering detection data when the multi-connected air conditioner is stably running, and a prediction model is updated according to the learning data, so that the deviation problem caused by the detection data is avoided, the learning deviation is reduced, and the updated prediction model can be applied to any working condition. For the scenes of multi-connected air conditioner equipment difference, environmental difference, system characteristic change after the multi-connected air conditioner is used for many years and the like, adaptive adjustment of the prediction model can be realized, and the general requirement is met.

The application discloses a campus intelligent guiding learning data security monitoring system based on data security, which comprises a data import module, a detection and analysis module and a risk research module.The data import module collects multi-source heterogeneous data streams of a campus system, performs scene recognition operation and anonymization preprocessing on the multi-source heterogeneous data streams, and obtains to-be-analyzed data streams.The detection and analysis module performs multi-dimensional context correlation operation on the to-be-analyzed data streams, generates guiding learning data correlation graph, constructs an abnormal guiding learning detection model based on an isolated forest based on the guiding learning data correlation graph, performs unsupervised anomaly detection based on the abnormal guiding learning detection model, and generates fine-grained anomaly detection results.The risk research module constructs a threat intelligence database, performs root cause tracing operation on the fine-grained anomaly detection results to obtain abnormal tracing detection data, performs real-time risk quantification evaluation on the abnormal tracing detection data, and generates comprehensive risk event data including risk level and attack path research.

This invention discloses an automated operation monitoring method for an aerial work platform process system, belonging to the field of automated data monitoring technology. The method includes: collecting historical multi-source heterogeneous data and its actual operating status; using a machine learning model to learn data relationships and obtain a system operation monitoring model; collecting real-time multi-source heterogeneous data and inputting it into the model to obtain output; and generating an automated operation monitoring report. The model training employs a hybrid optimization algorithm: based on interactive information, a fractional-order exploration strategy, a social learning exploration strategy, and a spatial tumbling exploration strategy are sequentially used for optimization, iterating until a termination condition is met. The parameters in the final target parameter combination are used as the final parameters of the machine learning model. This invention, by integrating optimization algorithms with multiple exploration strategies, can adaptively search for the optimal model configuration, improve model accuracy and generalization ability, and achieve precise monitoring and early warning of the operating status of the aerial work platform process system.

Disclosed are methods, systems, and computer programs for generating petrophysical analysis data. The methods include: accessing first porosity data and first well log data associated with an NMR well; randomly selecting data elements comprised in the first porosity data and the first well log data; developing, based on the randomly selected first porosity data and first well log data, a learning data structure; generating, based on the learning data structure, a machine learning model; receiving second well log data using sensor(s) deployed about a non-NMR well; adapting, based on the second well log data, variable input parameters of the machine learning model to generate second porosity data; and formatting the second porosity data to generate petrophysical analysis data. The petrophysical analysis data indicates one or more of: a petrophysical characterization of a rock type; flow dynamic data of a carbonate reservoir; well test or perforation zone data; and pore data.

The application belongs to the technical field of water quality safety intelligent monitoring, and discloses a single smell substance intelligent detection and self-checking system based on an electronic nose sensor array and multi-source information cooperation, which comprises an electronic nose sensor array module (containing a geosmin specificity sensor, a dimethyl isochroman specificity sensor and a dimethyl trisulfide specificity sensor), a multi-parameter monitoring module, a machine learning data processing module and a result output module; the multi-parameter monitoring module collects the dissolved oxygen consumption rate, pH value, conductivity and water temperature of water bodies in real time; and the machine learning data processing module establishes a dynamic correlation model of the single smell substance concentration and characteristic parameters through a machine learning algorithm. The application effectively solves the false positive misjudgment problem under the interference of a complex matrix, improves the detection accuracy to more than 95%, and has a concentration prediction error of less than or equal to 5%, thereby providing reliable technical support for the prevention and control of single smell pollution in drinking water, surface water and other water bodies.

Provide a system. 【Solution means】 Means for receiving students' learning data in real time and analyzing the data using a machine learning model for evaluating the learning progress, Means for automatically generating an optimized learning plan based on the students' understanding level, Means for providing an interactive teaching material according to the generated learning plan and giving feedback, Means for visualizing the learning progress to the instructor via a dashboard, Means for providing an online community function for promoting communication between educators and students, Means for providing individualized teaching materials to students in real time using a learning support robot, Means for recognizing students' expressions and voices to infer the understanding level and presenting appropriate teaching materials, A system including the above.

This invention relates to the field of knowledge recommendation technology and provides a personalized medical knowledge recommendation method and system for medical students. The invention acquires periodic test data; performs phased progress learning planning; compares the overlap of knowledge digestion; acquires reference learning data from other medical students; analyzes knowledge learning gaps; selects multiple gap-filling medical knowledge areas; and recommends relevant knowledge to the target medical student. Based on the target test data, the invention compares the periodic test data for phased progress learning planning and knowledge digestion overlap, selects reference medical students, acquires reference learning data, analyzes knowledge learning gaps between the reference and target learning data, selects multiple gap-filling medical knowledge areas, and recommends relevant knowledge to the target medical student. This allows the invention to uncover the commonalities and differences in learning among different medical students, providing them with forward-looking and targeted learning improvement directions for rapid academic improvement, effectively enhancing the learning efficiency of the target medical students.

The application discloses a kind of based on longitudinal federal learning's client data classification method, device and equipment, including obtaining to be detected client dataset, and to be detected client dataset is input to preset longitudinal federal learning classification model, preset longitudinal federal learning classification model includes feature coding module, feature purification module and server classification module;Data padding is carried out to to be detected client dataset using feature coding module, and output feature embedding dataset;Tensor decomposition is carried out to feature embedding dataset by feature purification module, and low-rank recovery tensor matrix is generated;Low-rank recovery tensor matrix is input to server classification module and is aggregated classification, and output target classification prediction result;The technical problem that the existing longitudinal federal learning data classification method can cause the situation that client data is missing in federal learning, thereby leading to the performance of longitudinal federal learning model is greatly reduced is solved.

This imaging method includes capturing a third image using an imaging device. The imaging device includes an imaging element, an optical system, and an optical element. The optical system forms an image in a light-receiving area from incident first light. The optical element guides second light to the light-receiving area. An angle between an optical axis of the optical system and a principle ray that enters the optical system is different for the first light and the second light. The third image is an image obtained when the first image and the second image are simultaneously displayed on a display. The first image is formed by an object point that radiates the first light. The second image is formed by an object point that radiates the second light.

The application discloses a health service production and education integration big data interaction method and system, comprising constructing a digital public information platform, building a production and education cooperation and human resource scheduling full-chain management module; embedding a user service feedback port in the platform, setting a credit level public display module to display to users, colleges and enterprises in real time, forming a service digital trust mechanism, synchronously feeding back credit data to a grading evaluation system, carrying out full-occupation-cycle production and education integration based on the platform, collecting learning data in real time in the training process, updating the skill archives of employees and grading evaluation results in combination with enterprise practice examination results. The application solves the limitations of single subject local data processing, improves the pertinence of talent training and the matching degree of student probation and employment, solves the problems of difficult enterprise talent screening and employee skill lag, and improves the trust degree of users to health service providers.

This invention relates to a method for determining timing chain elongation, a method for assessing wear, and an apparatus. The method for determining timing chain elongation, applied to a vehicle, includes: acquiring initial phase self-learning data obtained by self-learning the initial phase of the engine; determining initial phase deviation data between the actual and theoretical phases of the engine based on the initial phase self-learning data; determining actual phase deviation data caused by timing chain elongation based on the initial phase self-learning data and the initial phase deviation data; and determining the elongation rate of the timing chain in the engine based on the actual phase deviation data. This embodiment eliminates the need for additional chain position sensors or tensioner stroke sensors, thereby accurately quantifying the timing chain elongation rate without increasing hardware costs. This enables timely detection of faults in the engine timing chain drive system and improves the fault detection efficiency of the engine timing chain drive system.

A model learning unit learns a prediction model on the basis of learning data, a target setting unit sets a target output parameter value by interpolating between a goal output parameter value and an output parameter value which is the closest to the goal output parameter value in output parameter values in the learning data, a processing condition search unit estimates input parameter values which corresponds to the goal output parameter value and the target output parameter value, a model learning unit updates the prediction model by using a set of the estimated input parameter value and an output parameter value which is a result of processing that a processing device performs as additional learning data.

This application provides a learning companion system, relating to the field of smart education technology. The system includes multiple first terminals, at least one second terminal, and a processing layer. The first terminals receive pre-class teaching information and / or in-class teaching information and transmit it to the processing layer. The processing layer generates target learning data and a target visualized classroom learning report based on pre-class learning feedback and / or in-class learning feedback, and sends them to the second terminals. The second terminal generates post-class teaching instructions based on the learning data and the visualized classroom learning report and transmits them to the first terminals. The first terminals also collect post-class learning feedback information from target students in response to the post-class teaching instructions and send it to the processing layer. The processing layer generates target feedback results and sends them to the first terminals. This application provides a technical solution with a high degree of correlation between the generation of teaching instructions and learning data.

The application provides an enterprise load-oriented incremental multi-step time series prediction method, comprising reading historical data from a relational database and performing cleaning and preprocessing; jointly normalizing features and target variables to construct a supervised learning data set; constructing a deep learning model comprising a shared feature encoder, an attention fusion layer and a multi-task output layer, and training to obtain a model that can directly output multi-step prediction results; if there is a saved model, then performing incremental training based on the latest data; using the updated model to perform prediction and restoring physical quantity values through inverse normalization; and finally writing the prediction results and time stamps into a specified table in the database. The application can realize full-process automation and lightweight database integration, has engineering practicability and scalability, and improves enterprise load prediction accuracy and stability.