418 results about "Model validation" patented technology

A data collection system that provides the means to collect, store and make data available for a location analytics inference pipeline. The system incorporates a feedback mechanism enabling algorithms produced from the inference pipeline to drive the collection strategy to yield higher data quality and to produce reference data for model validation.

The invention discloses a multi-scale target detection method based on a deep convolutional neural network. The method comprises the steps of (1) data acquisition; (2) data processing; (3) model construction; (4) loss function definition; (5) model training; and (6) model verification. The method combines the ability of extracting image high-level semantic information of the deep convolutional neural network, the ability of generating candidate regions of region generation networks, the repair and mapping abilities of a content-aware region-of-interest pooling layer and the precise classification ability of multi-task classification networks, and therefore multi-scale target detection is completed more accurately and efficiently.

Computer-implemented systems and methods are provided for model validation of a model for compliance and credit risk. Model input, output, and processing validation areas are contained on a computer system. A handle data structure connects the model validation areas with handles that comprise a unified metric. A handle represents combinations of covariate patterns and describes the joint distribution of risk characteristics.

A system and method for verifying a memory consistency model for a shared memory multiprocessor computer systems generates random instructions to run on the processors, saves the results of the running of the instructions, and analyzes the results to detect a memory subsystem error if the results fall outside of the space of possible outcomes consistent with the memory consistency model. A precedence relationship of the results is determined by uniquely identifying results of a store location with each result distinct to allow association of a read result value to the instruction that created the read result value. A precedence graph with static, direct and derived edges identifies errors when a cycle is detected that indicates results that are inconsistent with memory consistency model rules.

The invention discloses a digital twinning evolution mechanism and method for an intelligent mine scene, the digital twinning is formed by mutual coupling and evolution integration of a physical model, a logic model, a simulation model and a data model, and the method comprises the following steps: (1) defining the physical model; (2) representing the logic model; (3) establishing a simulation model; (4) optimizing the simulation model; (5) verifying the simulation model; (6) constructing a data model; and (7) performing digital twinning integration. According to the invention, by constructinga digital twinning model, data mirroring and information interaction between a digital twinning body and a physical entity are realized, and object twinning, process twinning and performance twinningof the physical entity and a virtual space digital twinning body are realized; according to the invention, through a digital twinning evolution mechanism and method, remote visual monitoring of a physical space intelligent mine scene is realized in a virtual space intelligent mine scene, and intelligent perception, real-time monitoring, accurate positioning and health prediction are carried out on mine application scene equipment.

Computer-implemented systems and methods are provided for model validation of a model for compliance and credit risk. Model input, output, and processing validation areas are contained on a computer system. A handle data structure connects the model validation areas with handles that comprise a unified metric. A handle represents combinations of covariate patterns and describes the joint distribution of risk characteristics.

The invention discloses an image reflection removing method based on a depth convolution generative adversarial network. The method comprises the following steps: 1) data obtaining; 2) data processing; 3) model structuring; 4) loss defining; 5) model training; and 6) model verifying. In combination with the capability of a depth convolution neural network to extract high-level image semantic information and the flexible capability of the generative adversarial network to define the loss function, the method overcomes the limitations of a traditional method using low level pixel information, therefore, making the method with a better adaptive ability to the reflection removal of a generalized image.

Models are generated using a variety of tools and features of a model generation platform. For example, in connection with a project in which a user generates a predictive model based on historical data about a system being modeled, the user is provided through a graphical user interface a structured sequence of model generation activities to be followed, the sequence including dimension reduction, model generation, model process validation, and model re-generation.In connection with a project in which a user generates a predictive model based on historical data about a system being modeled, and in which the project includes a series of user choice points and actions or parameter settings that govern the generation of the model based on rules, which direct the user to select and apply an optimal model.

In one embodiment of the invention, a portable device with multiple integrated sensors for vital signs scanning and method of using said device is disclosed. The portable personal scanning device includes multiple sensors such as a plurality of ECG, thermometer, PPG, accelerometer, and microphone for determining a user's vital signs. The method includes concurrently scanning with one or more sensors, validating and enhancing the results of each sensor scan with other concurrent sensor scan and patient interaction models, processing the sensor scans separately or in combination to extract user's vital signs, validating the vital signs extracted by comparison to physiological models, and fusing the similar vital signs extracted from more than one process according to a determination of the measure of quality of the process that produced the vital sign.

The invention discloses a tumor malignant risk stratification auxiliary diagnosis system of an artificial intelligence medical image. The system comprises a data acquisition module, a data preprocessing module, a model establishment module, a model verification and optimization module, a stratification diagnosis module and a database platform. The tumor malignant risk stratification auxiliary diagnosis system of the invention is based on artificial intelligence technology, successive stratification of the malignant risk of the tumor can be achieved, the clinical diagnosis thinking is simulated, based on the high-precision ability of detecting benign lesions and malignant tumors of artificial intelligence model, and the space-occupying lesions with definite imaging features are diagnosed automatically. As a result, the system can substantially assist the clinical management decision of space-occupying lesions, improve the existing work flow of clinical diagnosis, increase the confidenceof doctors in diagnosis, reduce the work pressure, reduce the anxiety of patients with low-risk malignant lesions, greatly improve the diagnostic rate of benign lesions and malignant tumors, and hopefully realize the landing implementation of artificial intelligence clinical auxiliary diagnosis.

A photomask dataset corresponding to a target-pattern is verified by simulating a resist-pattern that will be formed in a resist layer by a lithography process, simulating an etched-pattern that will be etched in a layer by a plasma process wherein said simulation comprises calculating a flux of particles impacting a feature, and determining whether the etched-pattern substantially conforms to the target-pattern.

The invention relates to a device and a method for detecting hatching egg incubation quality based on computer vision, belonging to the technical field of agricultural product detection. The device comprises a CCD camera, an optical chamber, an object stage, a light source, an image acquisition card and a computer. The method comprises the following steps: adopting the CCD camera to acquire a perspective image of an egg, and then transmitting the image into the computer through the image acquisition card; processing the image to extract a color feature parameter; and discriminating whether theegg is fertilized or not according to a Bayes discrimination model for egg incubation quality. The invention detects the hatching egg incubation quality by applying a computer vision method and can reduce the labor intensity and the interferences of artificial subjective factors and obtain a more objective and accurate result. Proved by experiments, when the method is used for detecting the hatching egg incubation quality, the detection accuracy on the sixth day of a white-shell hatching egg reaches 100 percent, and the accuracy for verifying the model reaches 100 percent; the detection accuracy on the sixth day of a brown-shell hatching egg reaches 97.1 percent, and the accuracy for verifying the model reaches 100 percent.

The invention provides a vehicle road load simulation method which comprises the following steps of: 1) complete vehicle modular modeling; 2) complete vehicle modular model validation; 3) complete vehicle model assembly; 4) complete vehicle model validation; 5) establishment of finite element models of enhanced durable roads; 6) model initialization; 7) obtaining of virtual road load result; and 8) comparative validation with physical road load. The invention has the advantages of shortening development time and reducing costs, which is of great significance to vehicle development during the conceptual design phase. During the conceptual design phase of new models, fatigue analysis can be carried out before the design and manufacture process by using a virtual prototype to predict road load so as to truly predict the life of the product, thus greatly reducing great expenses caused by producing prototype machines and testing the fatigue life.

The invention provides an electric power credit evaluation method based on big data, and relates to an evaluation method. An existing credit evaluation system only based on a traditional quantitative analysis method cannot meet requirements of an electric power company to dig data value, undertake a customer credit rating system, and realize responsibility of enterprise derivative value at present. The electric power credit evaluation method comprises a data preparation step, an index system construction step, a model construction step, a model verification step and a model output step. According to the technical scheme, systematic cleaning and analysis are performed on a large number of historical data such as customer basic information, payment behavior, electricity using behavior and social credit information, an implicit behavior model and credit characteristics in the data are dug by applying a mathematical statistics method and a BP neural network model improved by genetic algorithm, a relationship between historical information and credit performance is captured, the credit performance is synthetically evaluated by a credit level, classified credit management of electric power customers is advanced, and differentiated arrears collection measure research work is supported and carried out to form a diversified collection strategy.

A computer method and apparatus of online automated model identification of multivariable processes is disclosed. The method and apparatus carries out automatically all the four basic steps of industrial process identification: 1) identification test signal design and generation, 2) identification plant test, 3) model identification and 4) model validation. During the automated plant test, process models will be automatically generated at a given time interval, for example, every hour, or on demand; the ongoing test can be automatically adjusted to meet the process constraints and to improve the data quality. Plant test can be in open loop operation, closed-loop operations or partly open loop and partly closed-loop. In a (partial) closed-loop plant test, any type of controller can be used which include proportional-integral-derivative (PID) controllers and any industrial model predictive controller (MPC). The obtained process models can be used as the model in advanced process controllers such as model predictive control (MPC) and linear robust control; they can also be used as inferential models or soft sensors in prediction product qualities. The apparatus can be used in new MPC controller commissioning as well as in MPC controller maintenance.

The invention discloses an MPI parallel programming system and an automatic generation method for framework code, which are based on visual modeling. A model is provided by a user interface layer based on a visual operating interface; a code generating layer is used for analyzing the model, wherein, the user interface layer consists of a graphic editor used for graphical modeling and a code editor used as the modifying environment for generating a code framework; the code generating layer comprises a arithmetic framework module, a model verifying module and a code generating module, wherein, the arithmetic framework module is used for describing features and actions of a mode, encapsulating the design mode once again and providing a program framework; the model verifying module is used for verifying the logical correction of the model before the code is generated; and the code generating module is used for invoking different parallel standard lib according to users' different demands. Compared with the prior art, the system and the method have good universality, so that even the professional staff without professional knowledge can have the code framework generated automatically via the method of visual modeling.

The invention discloses an establishing method for an energy consumption model of a base station. The establishing method comprises the following steps of: a, presetting an independent variable and obtaining historical energy consumption data and base station service and configuration data corresponding to the variable; b, fitting a straight line for the obtained data by using multivariate linear regression analysis, returning regression analysis parameters that describe the fitted straight line and establishing the energy consumption model; c, carrying out model validation; and d, carrying out model modification according to a model validation result until the model validation result indicates that the obtained energy consumption model achieves the expected result. The invention further discloses an energy consumption predicating method of the base station, a corresponding establishing device for the energy consumption model of the base station, and an energy consumption predicating device of the base station. Compared with the conventional energy consumption management system, the energy consumption model obtained by the establishing method provided by the invention has an accurate and effective energy consumption predicating function.

An aerial remote sensing platform remotely collects information including environmental monitoring data. The aerial remote sensing platform includes a camera, a microcontroller, and sensors. A ground base station communicates with the aerial remote sensing platform. The microcontroller monitors a pitch, a yaw, and a roll of the aerial remote sensing platform and automatically adjusts a pan and a tilt of a camera accordingly, thereby locking on a region of interest for the capture and processing of visible light and thermographic images of mixing from point source pollutant discharges. The aerial remote sensing platform auto adjustment occurs responsive to the microcontroller, which processes a variety of feedback loop information gathered by the sensors. A viewer is configured to display an analysis of the collected information including visible light and thermographic images. The analysis provided is used to validate a CORMIX simulation model for point source mixing.

A system and method for profiling subnet-level aggregate network data traffic is disclosed. The system allows a user to define a collection of features that combined characterize the subnet-level aggregate traffic behavior. Preferably, the features include daily traffic volume, time-of-day behavior, spatial traffic distribution, traffic balance in flow direction, and traffic distribution in type of application. The system then applies machine learning techniques to classify the subnets into a number of clusters on each of the features, by assigning a membership probability vector to each network thus allowing panoramic traffic profiles to be created for each network on all features combined. These membership probability vectors may optionally be used to detect network anomalies, or to predict future network traffic.

A metadata validation process that allows for deferring object model validation until after the objects are created. The process also allows for multi-threaded processing of the validation rules, thus increasing overall performance. Validation is performed by enforcing a series of validation rules on an appropriate subject. Rules are specified according to the subject that they are validating (i.e., attribute level, association level, object level or collection level). The metadata driven validation process implements several validation types on different validation units. Correctness validation rule types ensure that a validation unit satisfies all semantic rules defined for it. Completeness validation rule types ensure that a validation unit contains all the necessary data and is ready for further use. At design time, only correctness type validation is performed. Thus, the present invention advantageously allows for incomplete objects to be created at design time. The developer, however, in this case may opt to perform completeness validation at any time. In general, a developer may opt to perform completeness and / or correctness validation at any time independent of deployment processing. In another aspect, full validation (e.g., completeness and correctness) is automatically performed on the objects during the process of creating a configuration prior to deployment.

The invention discloses a multiaxis differential spectrum absorbing method and device for vertically distribution detection of atmospheric components. Wherein, a plurality of telescopes are arranged to observe the space from the ground in different angles of pitch, so as to detect and collect sun lights scattered by the atmosphere; the surfaces of the telescopes' lens cones are blackened to reduce stray light, the scattered sun light is converged at the ends of optic fibers after passing the lens and light filters in the telescopes; the optic fibers at the tail ends of the telescopes are combined and guide the scattered sun lights into the incident slit of an imaging spectrometer, which first disperses the scattered sun lights, and then concentrate them on a photosensitive surface of a face array imaging CCD detector; the spectral signals go through data analysis and treatment after A / D conversion and inputting into a computer. The invention is applicable for chemical research of atmosphere, provides input data and model verification for chemical model of atmosphere; and is applicable for verifying the results of atmospheric components detected by a satellite.

The invention discloses a quantitative adulteration detection method for peanut oil based on multiple-source spectroscopic data fusion. The quantitative adulteration detection method comprises the following steps: oil sample preparation; spectrum acquisition: respectively acquiring Raman spectrograms and near-infrared spectrograms of all adulterated oil samples; spectroscopic data fusion: performing data level fusion on the preprocessed Raman spectrograms and the preprocessed near-infrared spectrograms to obtain a fusion spectrogram; quantitative adulteration model establishment: extracting characteristic wavelengths of the fusion spectrogram, and establishing a quantitative peanut oil sample adulteration model through a multivariate quantitative calibration method; model verification: analyzing samples to be detected. The detection method performs data fusion on the edible oil spectrograms of two spectrums, has good complementarity, can reflect the inner characteristic information of edible oil more comprehensively, and is quick, convenient, efficient, non-destructive, free from preprocessing, high in accuracy, and strong in applicability.

The invention discloses a perceptual loss and deep convolutional neural network-based sketch simplifying method. The method comprises the steps of: 1) obtaining data: obtain the data by using a sketchmap and corresponding label data; 2) processing data: transforming the sketch map and corresponding line sketch map of an image data set into a format required for training a deep convolutional neural network through preprocessing; 3) constructing a model: constructing a deep convolutional neural network suitable for sketch simplification according to a training objective; 4) defining a loss function; 5) training the model: calculating the loss value of the network according to the loss function, calculating the gradient of each network layer parameter through back propagation and updating parameters of each layer of the network through a random gradient descent method; and 6) verifying the model: verifying the trained model by using a verification data set and testing the generalizationperformance. The method disclosed by the invention has the advantages that sketches with rougher miscellaneous lines and unclear main structure lines can be processed by a sketch simplifying network;and the network has strong robustness to the influence of illumination.

A data collection system that provides the means to collect, store and make data available for a location analytics inference pipeline. The system incorporates a feedback mechanism enabling algorithms produced from the inference pipeline to drive the collection strategy to yield higher data quality and to produce reference data for model validation.

The invention discloses a body-based verification tool of a power grid public information model and a method of the body-based verification tool, wherein a model verification system which is provided for analyzing metadata information from a CIM / XML (Common Information Model / Extensive Makeup Language) power grid model, comparing the metadata information with a uniform standard information model and analyzing compatibility and semantic consistency of the model; and aiming at semantic requirements of model verification, an OWL (Web Ontology Language) main body language is adopted as the description language for integrating the body model, and a body integration verification method based on CIM Schema and CIM Profile is provided. According to the invention, the defect of power grid enterprise information integration bus on semantic specification and the problem of mismatching models occurring during mutual operation process are solved, thereby being beneficial to the specification in semantics of the information bus and convenient to the management of assembly models and the upgrading maintenance of CIM.

The invention relates to a modeling method of a medium-temperature coal tar total-fraction hydrogen cracking lumping kinetic model. The method includes the steps of a, dividing hydrogen cracking reaction virtual lump components; b, fundamentally assuming the hydrogen cracking lump kinetic model; c, building a hydrogen cracking overall reaction network; d, building a hydrogen cracking lump reaction kinetic model; e, determining the kinetic parameters of each reaction, and selecting a target function; f, verifying the extrapolation performance and predication capacity of the built medium-temperature coal tar total-fraction hydrogen cracking lump kinetic model according to experimental comparison. The method has the advantages that six lump kinetic models are built for the medium-temperature coal tar total-fraction hydrogen cracking reaction process, experimental verification shows that relative error of model prediction results is smaller than 3%, so that the model is good in explanation and prediction capacity on the medium-temperature coal tar total-fraction hydrogen cracking reaction process.

The invention relates to the technology of the near space aerocraft. In order to provide a system-level consistency check novel model verification system output on the basis of the dynamic time sequence of the near space aerocraft model, give qualitative model verification based on the monte carlo method, carry out risk analysis to the model, comprehensively carry out consistency check to a model to be verified and give a visual verification criterion, the invention adopts the technical scheme that the verification and simulation realization method for a near space aerocraft model comprises the following steps of: (1) on the basis of the monte carlo simulation method, carrying out qualitative verification; (2) carrying out the qualitative model verification based on a time domain analysis method; and (3) carrying out qualitative model verification based on risk analysis, taking a model parameter as a risk to give a visual qualitative model verification criterion, giving a model comparison standard when a plurality of models to be verified exist, verifying whether the model can be safely used, and giving a safe-use range. The verification and simulation realization method for a near space aerocraft model, which is disclosed by the invention, is mainly applied to the design of the near space aerocraft.

A pulverized coal fineness soft-sensing method in a powder-making process based on a least squares support vector machine depends on a hardware platform, measuring instruments and a computer system used for performing a software calculation, wherein said software performs a soft-sensing through real-time process data obtained by a controlling computer of a data collector. Said method comprises the steps of selecting assistant variables, determining the structure of sample data, determining sample components, building a soft-sensing model based on the least squares support vector machine and performing the model training, parameter selection as well as the model validation and the model parameter emendation. The present invention is applied simply and easily, has a lower cost, has more accurate soft-sensing results, and helps to achieve an optimal control and an optimal operation of a pulverizing process.

The invention discloses distributed packet-switching chip model verification system and method and relates to the field of designing packet processing chips. The system comprises a core control module, a command line interface module, an SDK (software development kit) proxy interface module, a virtual network test meter module, a virtual chip configuration module and a C model package module. The core control module is a server program module. The command line interface module, the SDK proxy interface module, the virtual network test meter instrument, the virtual chip configuration module and the C model package module are client program modules. The system is a C / S / C (client / server / client) structured distributed system supporting three application scenarios, namely, system function model verification, software-hardware cooperated simulation verification and prototype verification. Modules required for different application scenarios are different. Verification closure time of the functions of the system at different levels can be shortened greatly, development efficiency is increased, and debugging cost is lowered.