88 results about "Data generator" patented technology

The invention discloses a test data intelligent generation and privacy compliance desensitization method, which belongs to the technical field of electric digital data processing, and comprises the following steps: monitoring the structure change of a preset regulation update source and a target application system, and obtaining compliance dynamic information; analyzing the compliance dynamic information to generate a dynamic compliance rule base; based on the dynamic compliance rule base, generating an initial test data set through a preset intelligent data generator; desensitizing the initial test data set, and outputting a desensitized test data set; scanning the desensitized test data set in real time for compliance verification, and generating a compliance verification report; and dynamically optimizing the generation strategies of the dynamic compliance rule base and the intelligent data generator based on the compliance verification report. Dynamic rule base construction, intelligent separated data generation, closed-loop verification optimization and desensitization audit tracing technologies are adopted, laws and regulations and system changes can be automatically adapted, and the service validity, privacy security and law compliance of test data are synchronously guaranteed.

The invention relates to the technical field of wireless communication protection, in particular to a wireless communication self-adaptive safety protection method and system based on edge computing. Capturing physical layer channel state information when the Internet of Things terminal equipment accesses the network gateway for the first time, a functional area semantic identifier of the network gateway and a current timestamp, and generating an anonymous session voucher; tracking operation track data of the anonymous session voucher, and constructing a track operation fragment; obtaining a privacy risk assessment value based on trajectory uniqueness, spatial information sensitivity and behavior association analysis of the trajectory operation fragment; when the privacy risk assessment value exceeds a preset risk threshold value, triggering a track link breaking mechanism; obtaining privacy enhanced trajectory data based on privacy protection processing and a trajectory broken link mechanism; training a virtual data generation model through the privacy enhancement trajectory data; and after the virtual data generation model converges, driving a virtual data generator in the virtual data generation model to generate a virtual operation and maintenance data stream so as to realize self-adaptive protection of wireless communication data.

The invention discloses a photovoltaic-energy storage-electric vehicle collaborative V2G optimization scheduling method adaptive to a building scene, and relates to the technical field of building energy management and vehicle network interaction, and the method comprises the steps: training to obtain a data generator; outputting sample data of new energy output and building load scenes through a data generator, and generating a plurality of typical scenes by using a clustering algorithm and a Nash equilibrium model; in each typical scene, planning vehicle network interaction facility configuration, and carrying out cooperative scheduling on a source, a load, a storage and a vehicle; selecting participating automobiles, and scheduling the participating automobiles in combination with battery loss and a compensation mechanism; and comprehensive evaluation of economy and operation quality is carried out on the collaborative scheduling result of the source, the load, the storage and the vehicle, the planning result of vehicle network interaction facility configuration and the scheduling result of the participating vehicle. According to the method, the uncertainty scene serves as input, cost minimization is achieved through the two-stage model, and the economical efficiency and feasibility of the scheme are verified through evaluation.

The invention belongs to the technical field of artificial intelligence and distributed machine learning, and discloses a personalized federal learning data generation method and a related device. The personalized federal learning data generation method comprises the following steps: acquiring a local data similarity distribution vector of a selected client; acquiring a trained local model of the selected client and sending the trained local model to a server in federal learning; and obtaining a global data generator issued by the server, training based on the local sample similarity distribution vector to obtain a local data generator, and finally generating personalized federal learning data of the selected client by using the local data generator. The technical scheme disclosed by the invention specifically relates to a personalized federated learning data generation scheme based on adversarial training and similarity evaluation, and can effectively obtain a generator and generate personalized federated learning data with relatively high reliability.

The invention discloses a digital annotation quality automatic verification method and system, and relates to the technical field of data processing. The method comprises the following steps: identifying and classifying original annotation data to obtain an identification task level; task matching is carried out according to the identified task level and the multi-level processing path, and task import connection is established; obtaining a distribution verification target through task import connection, calling an adversarial data generator to create semantic and quality dimension variants, comparing the annotation consistency of original data and variant data, and obtaining annotation vulnerabilities and potential error information; and configuring an identification tolerance threshold according to the identification task level, configuring an identification constraint parameter based on the vulnerability and the potential error information, executing annotation verification in the matching processing path based on the identification constraint parameter, and outputting a verification result. The technical problems of low efficiency and insufficient accuracy caused by the fact that digital labeling quality verification depends on manual sampling inspection in the prior art are solved, and the technical effect of improving verification efficiency and accuracy is achieved.

The application provides a visual language large model adaptive bias evaluation method for evaluating bias performance of a target model on a specific task, and the method comprises the following steps: S1, obtaining an initial evaluation dataset, wherein the initial evaluation dataset comprises a plurality of evaluation data; S2, fine-tuning a preset large language model using the initial evaluation dataset in a preset fine-tuning manner to obtain an initial data generator; and S3, generating evaluation data based on the initial data generator to perform multiple rounds of bias evaluation on the target model, wherein in each round of bias evaluation, based on all evaluation data corresponding to the last round of bias evaluation, a data generator corresponding to the last round of bias evaluation is iteratively updated according to a preset objective function to obtain a current round data generator, and the current round data generator is used to generate a current round evaluation dataset to perform bias evaluation on the target model.

A virtual reality (VR)-based tour route providing apparatus, which creates a tour route of a corresponding tourism area through a VR experience in which a tourism area of users is implemented according to one embodiment of the present invention, includes an area information receiver, a zone information receiver, a VR data generator, a tracking data generator, a zone visit information generator, a zone-of-interest selection unit, and a tour route information generator.

The invention relates to a system (100) for optimizing the configuration of crash tests and simulations for vehicle development. The system (100) comprises a data preparation module (200) that splits a data set (210) containing crash test and simulation data into a first training data set (210), a validation data set (230), and a test data set (240); a training module (300) with a first AI model (350) for creating configurations of physical crash tests and virtual simulations, wherein the first AI model (350) is trained with the first training data set (220); and a data value module (400) that calculates data values ​​(φ(i)) for the data points (i) of the crash test and simulation data of the training data set (220) with respect to the validation data set (230).a model optimization module (500) comprising a second AI model (550) and a training data generator (570), wherein the training data generator (570) creates a second training data set (580) which is used to train the second AI model (550) to predict the relationship between input parameters (p) and the computed data values ​​(φ(i)); a parameter module (700) which uses the trained second AI model (550) to find crash test parameters with a maximum data value (φ(i)).

The invention discloses a wireless data transmission system and method using a compression technology, and the system comprises a data generator which is used for generating to-be-transmitted wireless data and transmitting the to-be-transmitted wireless data to a compression module, the compression module is used for compressing the to-be-transmitted wireless data into compressed data by using a self-adaptive differential pulse code modulation algorithm, and the compressed data is transmitted to the compression module through the self-adaptive differential pulse code modulation algorithm; the gateway end comprises a gateway protocol end and a transmission module, and the gateway protocol end is used for receiving the compressed data packet and sending the compressed data packet to the transmission module for transmission; and the server is used for receiving and decompressing the compressed data packet transmitted by the transmission module so as to realize the transmission of the wireless data to be transmitted. According to the invention, low-cost, high-efficiency and high-real-time wireless data transmission under a large field domain can be realized.

A modular adversarial training data generator (“generator”) generates adversarial training data for a machine learning (ML) model to detect malicious protocol data units (PDUs). The adversarial training data mimics high volume cyberattacks by perturbing PDUs to bypass malicious detection systems. For Hypertext Transfer Protocol (HTTP) PDUs, the generator rearranged, replaces, and grid searches values of HTTP header fields to generate the adversarial training data. The generator further biases grid search based on metrics for values of HTTP header fields that quantify impact of replacing the values on malicious verdicts by ML models trained on the adversarial training data.

The application discloses an electric energy meter informationization evaluation calibration model measurement and evaluation verification system and method, and the system comprises an electric energy meter informationization evaluation calibration model measurement uncertainty evaluation and control module, a load data generator based on a digital twin mixed reality technology and an electric energy meter informationization evaluation calibration model result measurement and evaluation verification method research and measurement and evaluation verification platform. The application evaluates and verifies the electric energy meter informationization calibration model in theory and practice, improves the informationization measurement accuracy and reliability to a new height, and also solves the problems of low work efficiency, high operation and electric energy meter replacement cost, high manual labor intensity, incapability of real-time monitoring of electric energy meters and timely fault reporting and early warning, incapability of effectively controlling the measurement faults and error power and the like existing in the currently generally adopted traditional manual field calibration or disassembly laboratory calibration mode.

The invention provides a large-model multi-modal and multi-dimensional data augmentation method and device based on a man-machine interaction mode, and the method comprises the steps: collecting multi-modal and multi-dimensional agricultural data, carrying out the preprocessing and formatting unification, and obtaining the processed agricultural data; obtaining a multi-mode demand instruction input by a user through the man-machine interaction interface; calling a data augmentation model based on the multi-modal demand instruction and the processed agricultural data to obtain augmented agricultural data; performing cross-modal consistency verification on the augmented agricultural data, and fusing the augmented agricultural data passing the verification with the processed agricultural data to obtain an augmented agricultural data set; the data augmentation model comprises a large language model in the agricultural field, a CLIP model in the agricultural field, an agricultural knowledge graph and a data generator corresponding to multiple modal types associated through an attention mechanism of a shared semantic space. And therefore, high-quality multi-modal augmented data which is consistent in semantics and reasonable in agriculture can be generated.

A setting system includes a setting input section, an image data generator, and a control data generator. The setting input section is configured to receive an input of setting data on a setting of operation of a plurality of facilities. The plurality of facilities includes a lighting fixture and a shielding apparatus. The image data generator is configured to render, based on model data, image data representing a light environment corresponding to the setting data of the building. The control data generator is configured to generate control data for executing control corresponding to the setting data over the plurality of facilities.

The application relates to the technical field of image processing, and discloses a light-weight single-pixel imaging method based on frequency domain-space domain modulation attention, which comprises a single-pixel image data generator and a calculation reconstruction generator; an initial inversion estimation is obtained through inverse mapping of a measurement matrix; then, the initial inversion estimation is input into a dual-domain collaborative attention deep reconstruction network; the network is composed of a plurality of dual-domain collaborative attention module cascades connected in series, a multi-head spectrum calibration module in the module realizes long-distance information reconstruction through dual-domain feature information fusion interaction; a multi-scale-depth dynamic convolution fusion module extracts features in a space domain, accurately processes high-frequency edges, and significantly enhances the reconstruction capability of texture details; and the features are corrected in combination with a residual error guiding update module. Through the dual-domain collaborative mechanism, the reconstruction precision and the detail recovery capability of single-pixel imaging are effectively improved while the network parameter quantity and the calculation complexity are greatly reduced, and the application is suitable for edge deployment.

The present disclosure pertains to a power amplifier system that promotes enhanced signal linearity and overall system efficiency. The system includes a power amplifier with an amplification path for a radio frequency (RF) signal, and detector circuitry operationally linked to sample locations along this path. The detector circuitry captures and transmits signal characteristics of the RF signal. A voltage standing wave ratio (VSWR) quadrant data generator in communication with the detector circuitry generates VSWR quadrant data based on the detected signal characteristics. The baseband circuitry, comprised of a memory unit preconfigured with digital pre-distortion (DPD) coefficients and a DPD processor, controls the shaping of pre-distortion applied to the RF signal based on the VSWR data, thereby enhancing signal linearity. The components of the system interconnect and collaboratively function to optimize the performance and efficiency of the power amplifier system.

The invention relates to the technical field of data enhancement, and discloses an infrastructure resource library data enhancement method, system and device based on artificial intelligence and a medium, comprising: constructing an infrastructure contract format and a noise knowledge engine, obtaining contract data, initiating through an adversarial noise generation agent, and based on the infrastructure contract format and the noise knowledge engine, obtaining the contract data; generating antagonism data under the fact constraint; and constructing a target contract key clause identification model to obtain a contract identification result and a loss value, generating an instant reward signal or a delay penalty signal of the adversarial noise generation agent, and performing strategy updating on the adversarial noise generation agent. According to the method, a closed-loop online data enhancement architecture based on fact constraint and antagonistic learning is introduced, so that the fidelity and rationality of generated data are remarkably improved, the efficiency and pertinence of data enhancement are greatly improved, and co-evolution of a data generator and a data learner is realized.

Embodiments of the present disclosure disclose an image-to-hyperspectral data conversion method, device, electronic equipment and medium. A specific implementation of the method comprises: obtaining image data from a pre-constructed image set; preprocessing the obtained image data to obtain pre-processed image data; generating hyperspectral data according to a hyperspectral data generator and the pre-processed image data; generating a performance analysis result corresponding to the hyperspectral data according to a performance analyzer and the hyperspectral data; generating optimization parameter data according to the performance analysis result; performing multi-stage optimization on the hyperspectral generator and the performance analyzer according to the optimization parameter data to obtain an optimized hyperspectral generator and an optimized performance analyzer; and generating target hyperspectral data according to the optimized hyperspectral generator and the optimized performance analyzer. The implementation generates hyperspectral data, performs multi-stage optimization and performance evaluation, and obtains target hyperspectral data of higher quality.

Systems and methods for super sampling and viewport shifting of non-real time 3D applications are disclosed. In one embodiment, a graphics processing unit includes a processing resource to execute graphics commands to provide graphics for an application, a capture tool to capture the graphics commands, and a data generator to generate a dataset including at least one frame based on the captured graphics commands and to modify viewport settings for each frame of interest to generate a conditioned dataset.

A brain function improvement facilitating device includes: a storage; a display screen; a timer; a heart rate measurement unit; a target exercise intensity calculator; a setting unit; a first determiner that determines whether a heart rate of a user is greater than or equal to a target heart rate set by the setting unit; a duration time measurement controller that performs control to measure a duration time in which the heart rate of the user is greater than or equal to the target heart rate if the heart rate of the user is determined to be greater than or equal to the target heart rate; a second determiner that determines whether a total of the duration time measured within a predetermined time limit has reached the target time set by the setting unit; a storage controller that controls the storage to store information indicating that the duration time has reached the target time if it is determined that the total of the duration time has reached the target time; a display data generator that generates display data to be displayed on a display screen based on reaching the target time if it is determined that the duration time has reached the target time; and a display controller that controls the display screen to display at least one of the calculated target heart rate, actual related data and time, or the display data.

In an image data control apparatus, an image data generator generates image data using external light received thereby. A processing unit is configured to process the image data. A determiner determines whether to request transmission of the image data to the processing unit. A transmitter transmits the image data to the processing unit upon the determiner determining to request transmission of the image data to the processing unit, and transmits alternative image data in place of the image data to the processing unit upon the determiner determining not to request transmission of the image data to the processing unit.

The present disclosure provides a device, a method and a semiconductor device for reducing current change rate, relates to the technical field of semiconductors, and particularly relates to the fields of semiconductor chips, artificial intelligence chips, power control and the like. The implementation scheme is: a virtual data generator configured to generate virtual data that changes with clock cycles; a signal injection controller configured to activate a power maintenance flag signal for providing to a dynamic power logic circuit in a first period and a second period, and control the virtual data to be provided to the dynamic power logic circuit, deactivate the power maintenance flag signal in a third period, and control the data load to be provided to the dynamic power logic circuit; and a multiplexer configured to provide the virtual data to the dynamic power logic circuit in the first period and the second period based on the control of the signal injection controller, and provide the data load to the dynamic power logic circuit in the third period.

Provided is an edge module, a control system, a remote control system, a controller, and a communication method that enable fixed-cycle processing. An edge module includes: a data generator configured to generate first data including a serial number; a transmitter configured to assign a first timestamp to the first data and transmit the first data to a controller; a receiver configured to receive second data generated by the controller based on the first data, the second data including the serial number and the first timestamp, and to obtain a second timestamp corresponding to a reception time of the second data; and an abnormality detector configured to determine whether communication with the controller was successfully performed based on at least one of: (i) the serial number included in the second data, and (ii) a difference between the first timestamp included in the second data and the second timestamp.

Examples of the present disclosure relate to a data generator for generating training data or monitoring data for use in a communication network. An apparatus is configured to receive one or more tuning parameters, where the tuning parameters are based, at least in part, on data obtained from measurements of physical conditions such that the tuning parameters represent, at least in part, the physical conditions. The tuning parameters may be used to adjust a data generator, where the data generator includes a model representing a physical condition, and the adjusted data generator may be used to generate data for the machine learning model.

The present disclosure relates to an alternative protein material prediction device. The device includes a protein feature extractor configured to composite features of protein as sequence features, structural features, and physicochemical features; a protein graph data generator configured to generate nodes based on the sequence features and physicochemical features of the protein and generate edges between the nodes based on the structural features of the protein, thereby generating protein graph network data, and an alternative material protein predictor configured to generate an alternative protein material prediction model for predicting an alternative protein material by learning the protein graph data that reflects the composite features of the protein.

Method and apparatus for improving synthetic ground truth data by means of a data generator and for training a target machine learning model. The method includes: providing ground truth data samples which relate to ground truth source data and are synthetically generated by the data generator; comparing a performance of the data generator for the provided ground truth data samples with a performance threshold value; generating anew ground truth data samples for the same ground truth source data by means of the data generator if the performance threshold value for the provided ground truth data samples is not achieved; replacing the ground truth data samples for which the performance threshold value is not achieved with the newly generated ground truth data samples; and training the target machine learning model on the basis of the replaced and provided ground truth data samples.

The embodiment of the invention provides an image transmission device, electronic equipment and a transmission control method, relates to the technical field of chips, and improves the transmission efficiency of data. According to the specific scheme, the image transmission device comprises a transmitter and a data generator. The transmitter is used for transmitting first data to a first device coupled with the image transmission device, and transmitting third data to the first device when second data transmitted by the first device is not received within a preset duration after the first data is transmitted. Wherein the preset duration is a time window for the first device to send the second data. The data generator is used for generating first data and third data. The embodiment of the invention is used for the process of data transmission between the image transmission device and the first device.

The present application relates to a kind of privacy big model self-adapting adjustment method and system based on size model and LoRA, specifically, server side utilizes data generator, and combining differential privacy technology, based on public data set, generate synthetic data;The data distribution of adjacent two times generated on server is calculated KL divergence, if the data distribution KL divergence change is less than divergence threshold value, then the generated data is input to user side model, and user sends output representation to server;If the representation change of adjacent two times is greater than representation threshold value, each user calculates the model prototype corresponding to each category of data owned by it, after server receives the model prototype sent by user, it is projected on the base space that existing LoRA is made up of, if projection residual is greater than or equal to projection threshold value, then add and train LoRA module.

The invention relates to the technical field of image processing, and discloses a lightweight single-pixel imaging method based on frequency domain-space domain modulation attention, which comprises the following steps: constructing a single-pixel image data generator and calculating a reconstruction generator; initial inversion estimation is obtained through reverse mapping of the measurement matrix; sending the data to a double-domain collaborative attention depth reconstruction network; the network is formed by cascading a plurality of stages of double-domain collaborative attention modules which are connected in series, and multi-head spectrum calibration modules in the modules realize remote information reconstruction through double-domain feature information fusion interaction; the multi-scale-depth dynamic convolution fusion module performs feature extraction in a spatial domain, accurately processes a high-frequency edge and remarkably enhances the reconstruction capability of texture details; and a residual error is combined to guide the updating module to correct the features. Through a double-domain cooperation mechanism, the reconstruction precision and the detail recovery capability of single-pixel imaging are effectively improved while the network parameter quantity and the calculation complexity are greatly reduced, and the method is suitable for edge deployment.

To collect point group data capable of measuring an absolute distance only by a camera.SOLUTION: The point-cloud-data generating device 10 includes an input device 11 that inputs a plurality of images simultaneously captured from different viewpoints using a plurality of cameras 200A and 200B in a known positional relationship, a point cloud generator 12 that generates point cloud from the plurality of images and estimates positions of the plurality of cameras 200A and 200B on the point cloud, and a corrector 13 that calculates a ratio of an interval between the plurality of cameras 200A and 200B in a real space to an interval between the plurality of cameras 200A and 200B on the point cloud and corrects a scale of the point cloud using the calculated ratio.SELECTED DRAWING: Figure 3