12results about How to "Improve data processing accuracy" patented technology

The invention provides an intelligent guard system for large-scale micro-nano processing laboratory equipment, and the system comprises a data collection module, a risk early warning module, a fault positioning module and an operation and maintenance management module, the fault is traced by using the collected real-time environment data, real-time state data, experimental data, real-time behavior data and identity information, so that the details of the fault can be known in time, and fault maintenance, responsibility investigation and the like can be carried out; the operation and maintenance management module can carry out multi-dimensional operation and maintenance management on the laboratory according to the collected real-time environment data, real-time state data, experimental data, real-time behavior data and identity information, so that the efficiency, the performance and the service life of equipment are improved, the maintenance cost of the equipment is reduced, and effective utilization of social resources is facilitated.

The invention discloses an intelligent control system for a driving motor of a new energy automobile based on energy balance, and relates to the technical field of new energy automobiles, the intelligent control system for the driving motor of the new energy automobile based on energy balance realizes comprehensive monitoring, accurate analysis and dynamic regulation and control of the running state of the motor through organic combination of a plurality of modules; compared with a traditional control system, the system has the advantages that the data processing precision, the real-time response capability and the intelligent level are remarkably improved, complex and changeable operation working conditions can be better dealt with through a double comparison mechanism and adaptive regulation and control, efficient and safe operation of the motor is ensured, and therefore the overall performance and reliability of the new energy automobile are improved, and the new energy automobile is suitable for popularization and application. Through comparative analysis, the system can identify the optimal control strategy, and dynamically adjust in the real-time driving process to ensure that the motor is always in the optimal state.

The invention discloses a time delay estimation model optimization method and system based on kernel density, a medium and equipment, and the method comprises the steps: obtaining time delay related data of a power distribution network communication network, and the time delay related data comprises the number of access subjects of the power distribution network communication network, the kernel density function bandwidth and a time delay measurement value; constructing a time delay estimation model of the power distribution network communication network by adopting a non-parametric kernel density estimation method according to the time delay related data; and optimizing the fixed bandwidth of the time delay estimation model by adopting a Gaussian kernel function and a geometric distance to obtain a time delay estimation optimization model. Under the condition of multivariate main body access, the adaptive nonparametric kernel density estimation method is utilized to adjust the time delay estimation model, the probability density function is estimated according to the characteristics of the data, different characteristics in the data are conveniently captured, and the obtained time delay estimation model of the communication network of the power distribution network is high in data processing precision and high in reliability. The method has good engineering application value.

The invention relates to a question and answer model training method and device, computer equipment and a readable storage medium. The method comprises the steps of obtaining a training question corresponding to a target field and a training answer corresponding to the training question; inputting the training question into an initial question-answering model, wherein the initial question-answering model comprises a pre-training question-answering network and an auxiliary network; extracting a first question feature of the training question through a pre-training question and answer network, and extracting a second question feature of the training question through an auxiliary network; determining a predicted answer to the training question based on the first question feature and the second question feature; the auxiliary network comprises a feature extraction unit and a dimension control unit; generating a model loss based on the predicted answer and the training answer; adjusting a feature extraction unit based on the model loss; and based on the model loss and a reference threshold, adjusting a dimension control unit until a convergence condition is met, and obtaining a target question and answer model corresponding to the target field. By adopting the method, the answer prediction accuracy of the question and answer model can be improved.

The embodiment of the invention discloses a collaborative training method and device of a model, equipment, a chip and a program product. The method comprises the following steps: deploying a target network model into a storage and calculation integrated chip; obtaining sample data in a training data set and a true value corresponding to the sample data; according to the sample data and the target network model deployed in the storage and calculation integrated chip, obtaining a predicted value corresponding to the sample data; and updating the target network model deployed in the storage and calculation integrated chip according to the difference between the true value corresponding to the sample data and the predicted value corresponding to the sample data.

This application discloses an information recommendation method, which may be applied to the field of artificial intelligence. The method includes: obtaining a target feature vector; and processing the target feature vector by using a recommendation model, to obtain recommendation information, where the recommendation model includes a cross network, a deep network, and a target network; the target network is used to perform fusion processing on a first intermediate output that is output by the first cross layer and a second intermediate output that is output by the first deep layer, to obtain a first fusion result, and the target network is further used to: process the first fusion result to obtain a first weight corresponding to the first cross layer and a second weight corresponding to the first deep layer, and weight the first fusion result with the first weight and the second weight separately, to obtain a first intermediate input and a second intermediate input. In this application, the target network learns heterogeneous parameter distribution for different interaction networks, to avoid excessive sharing, and introduces interaction signals between different interaction networks. This improves prediction accuracy of the model.

The invention relates to the field of data processing, in particular to a three-dimensional scene reconstruction method based on point cloud data feature extraction, and the method comprises the steps: determining the type of a target region according to the region complexity of the target region and the collection dynamic influence degree, and determining the actual region state according to the proportion of a first type region and the concentration degree of the first type region, determining a task construction mode according to the actual region state; under the first task construction condition, determining the representation state of the aggregation region according to the area trend degree and the area reference value of the aggregation region, and performing secondary judgment on the task construction mode according to the representation state of the aggregation region; determining a task distribution mode according to a comparison result of the resource margin richness and a preset resource margin richness, wherein the task distribution mode is task distribution based on the resource margin or task distribution based on the resource margin and a conflict influence value; according to the invention, the model construction data management efficiency is improved.

The invention belongs to the technical field of ground plane laser measurement, and particularly relates to a stratum deformation early warning system in the parallel stacking construction process of an overpass existing pipeline, which comprises an acquisition module, a cloud processing center and an early warning response module, the collection module completes data collection through a laser monitoring device arranged at the construction position of the existing pipeline parallel stacking construction tunnel, and the collection module sends collected monitoring data to the cloud processing center; the cloud processing center calculates the feedback monitoring data to judge whether formation deformation occurs, if the formation deformation occurs, whether an early warning threshold value is reached is judged, and if the early warning threshold value is reached, an early warning notification is sent to an early warning response module; and after the early warning prompt module obtains the early warning notification, emergency early warning is carried out through the cloud processing center and the laser monitoring device. Whether the stratum deforms or not is measured through laser, the measurement result is accurate, and early warning prompt is safer.

This invention discloses a β-γ-ray dose equivalence test method for the irradiation qualification of nuclear power cables. Using the polyolefin sheath material of nuclear power cables as the test object, standardized Type II dumbbell test pieces with multiple thickness gradients are prepared. Gradient dose irradiation is carried out using β-ray and γ-ray devices matched to the design reference accident parameters. Through closed-loop environmental control, precise detection of elongation at break, least squares linear fitting, quantitative thickness correction, and equivalence verification, the β-γ-ray dose equivalence conversion relationship is accurately obtained. This invention solves the technical problems of existing technologies, such as the difficulty in implementing β-ray irradiation tests, the lack of a standard β-γ equivalence relationship, the significant influence of environmental / thickness interference on test data, and the lack of verification steps. It adds technical features such as closed-loop environmental control, thickness correction models, equivalence verification, and intelligent data processing, resulting in accurate and reliable equivalence ratios. It is applicable to the irradiation qualification of third-generation nuclear power cables, significantly improving test accuracy and licensing prospects.

The invention relates to the technical field of satellite control systems, provides a miniature electromagnetic valve propulsion system, and solves the problem that normal operation of a propeller control system is affected by inaccurate data processing due to influence of factors such as noise and temperature in a signal conversion process. According to the method, historical data of an electromagnetic valve are processed in different periods, a control model with a specific period is constructed, prediction of a real-time flight control command and engine performance simulation are achieved, and through comparison of real-time data and prediction data, the performance of a propeller is influenced, and driving power cannot be provided for a satellite. And the system can find and process abnormal data in time, so that the engine control system can operate normally. According to the invention, the accuracy of data processing is improved, the real-time monitoring and optimization of the performance of the propeller are realized, and the stability of satellite flight is improved.

The embodiment of the invention provides a multi-modal data processing method and device, equipment and a storage medium. Obtaining multi-modal data; wherein the multi-modal data comprises visual data and text data; inputting the visual data into a plurality of visual encoders for feature extraction to obtain a plurality of visual features; wherein the plurality of visual encoders are different from one another; and inputting the plurality of visual features and the text data into a multi-modal large language model, and outputting a target text. According to the multi-modal data processing method provided by the embodiment of the invention, after feature extraction is performed on the visual data through a plurality of different visual encoders, the extracted visual features are input into the multi-modal large language model for processing, so that the data processing precision of the multi-modal language model can be improved.

The present invention relates to the field of data processing technology, specifically to a processing method and device based on sewage drug detection data, the method comprising obtaining sampling data of each sewage sampling point in a target area, coupling each sampling data with the historical sampling data of its corresponding sampling point, obtaining abnormal data, and determining the drug content in the sewage as a first drug concentration; determining a pipe network segment to be checked; obtaining sewage sampling data in a previous pipe network segment connected to the pipe network segment to be checked, and determining the drug content in the sewage as a second drug concentration; obtaining sewage sampling data at any position upstream of the sampling point position corresponding to the abnormal data in the pipe network segment to be checked, and determining the drug content in the sewage as a third drug concentration; and determining the source of drug discharge according to the comparison results of the first drug concentration, the second drug concentration and the third drug concentration. The present invention solves the problems of low efficiency and easy data error in traditional manual statistical analysis of sewage drug detection data.