24results about How to "Reduce error propagation" patented technology

The invention relates to the technical field of gene analysis bioinformatics, and particularly discloses a gene data analysis method based on a large language model.The method comprises the steps that the quality of gene data to be analyzed is analyzed, whether the gene data to be analyzed is available or not is judged, and a gene sequence segmentation scheme of the available gene data is determined; inputting the gene data after gene sequence segmentation into the large language model to obtain a gene data association feature data set and a gene data analysis result, determining a gene data analysis accuracy feature value, and judging whether the gene data analysis result conforms to expectation or not; the problems that quality and integrity problems of traditional biological data challenge the accuracy of analysis results and effective integration and analysis of multi-modal data still need to be improved are solved, the data processing efficiency is optimized in the whole process, repeated calculation, misjudgment and invalid results are reduced, the analysis efficiency is improved, and resource waste is reduced.

The invention relates to the technical field of meteorological monitoring and lightning protection, and discloses a radio signal-based lightning protection detection method for a meteorological station, which comprises the following steps of: receiving electromagnetic wave signals generated by lightning discharge in real time through radio receiving equipment of a plurality of meteorological stations, and acquiring characteristic data of time, frequency and intensity of the signals; an effective light velocity model is adopted to correct the influence of the propagation velocity of the lightning signal in the atmosphere, and the effective propagation velocity of the lightning signal is calculated by considering the ionospheric and atmospheric refraction effects; the position of the lightning power supply is estimated by measuring the arrival time difference and the time difference of lightning signals received by different sites and utilizing the geometric positions of the multiple sites. The positioning precision of the lightning power supply is greatly improved by adopting the cooperative work of the radio receiving equipment of a plurality of meteorological stations and combining the technologies of accurate lightning signal propagation speed correction, time synchronization error correction, geometric precision optimization and the like.

The invention provides a spoken language understanding method and device, electronic equipment, a storage medium and a product. The spoken language understanding method comprises the steps of obtaining a to-be-understood voice signal; inputting the voice signal into a trained spoken language understanding model to obtain semantic information of the voice signal; a spoken language understanding model comprising: an automatic speech recognition encoder for converting an acoustic feature of the speech signal into an acoustic representation; the automatic speech recognition decoder is used for performing text transcription on the acoustic representation to obtain a speech transcription text representation; the fusion module is used for fusing the acoustic representation and the voice transcription text representation to obtain a joint representation; the spoken language understanding decoder is used for performing semantic prediction on the joint representation to obtain a semantic representation; and the output layer is used for obtaining the semantic information of the voice signal based on the semantic representation, so that the semantic output can be directly predicted from the voice input through the end-to-end spoken language understanding model, the error propagation problem is reduced, and the accuracy of the semantic information of the voice signal output by the model is improved.

The invention relates to a joint entity and relation extraction method and system based on deep learning and a medium, and the method comprises the steps: S1, employing a BERT model as a sentence encoder, and obtaining a context sensing embedded vector of each mark in a natural statement X inputted by a user; s2, taking joint entity and relation extraction as a set prediction problem, and directly generating a prediction triple set by using a non-autoregression decoder based on Transform; and S3, setting a set prediction loss function to carry out model training, and generating an optimal matching relationship between the prediction triple set and the real triple set according to the set prediction loss function. According to the method, a non-autoregressive decoder is adopted, and a joint entity and relation extraction task is regarded as a set prediction problem, so that the sequence of triples does not need to be predicted, and the accuracy is remarkably improved; since the non-autoregressive decoder does not depend on the previously predicted output, error propagation due to step dependency is reduced.

The invention relates to the technical field of synthetic aperture radar imaging detection, in particular to a moving target imaging and repositioning method based on a bistatic synthetic aperture radar, and the method comprises the steps: obtaining an original echo signal of a moving target through the bistatic synthetic aperture radar, and correcting the original echo signal based on two-step transformation; extracting a first-order Doppler parameter estimation value of the echo signal after the two-step transformation and moving target distance information under different azimuth blocks; constructing a parameter estimation and relocation model, and solving a second-order Doppler parameter estimation value and relocation information according to the first-order Doppler parameter estimation value and the moving target distance information under different azimuth blocks; and according to the second-order Doppler parameter estimation value and the repositioning information, constructing a matching function to perform focusing imaging on the moving target. The objective of the invention is to synchronously realize parameter estimation and relocation of a moving target and improve the calculation efficiency.

The invention relates to a lightweight radiation source positioning method. The method comprises the following steps: acquiring a mask map containing building distribution information in a target area; drawing a grey-scale map containing the position information of each sampling point in the target area and the received signal intensity of each sampling point as an electromagnetic sampling map; an input image sample obtained after the mask map and the electromagnetic sampling map are fused is put into a lightweight radiation source positioning model, and radiation source position coordinates are obtained through analysis; the lightweight radiation source positioning model comprises a plurality of feature extraction modules used for performing multi-scale feature extraction based on residual connection and an attention mechanism on an input image sample; the feature fusion module is used for converting the multi-scale feature maps output by the different feature extraction modules into set sizes and then carrying out feature fusion; and the space mapping module is used for predicting the position coordinates of the radiation source according to the fusion features generated by the feature fusion module. The position of the radiation source can be effectively and accurately obtained.

This invention discloses an efficient inversion method, medium, and equipment for material parameters of rockfill dams that considers target uncertainty and prior knowledge constraints. The method includes steps such as determining inversion parameters and calculating measurement points, constructing a high-precision multi-output surrogate model, constructing a multi-objective function expression, performing multi-objective parameter inversion calculations, and optimizing the solution set. By constructing a multi-objective function, the solution set space of target settlement under different spatiotemporal importance trade-offs is obtained. When generating parameters, the prior relationships between material parameters are used as constraints to avoid the generation of unreasonable results. At the same time, the uncertainty of target settlement is considered and reduced by introducing an error allowable interval and a high-precision surrogate model. Finally, a relaxed and improved Bayesian multi-objective optimization algorithm is used to achieve efficient and high-quality calculation of parameter inversion.

The invention relates to the technical field of protocol analysis, in particular to a network protocol collaborative combination analysis method based on a tree hierarchical structure, which comprises the following steps of: acquiring different network protocol analysis requirements, training a basic network protocol analysis model and a coarse granularity model, and acquiring a hierarchical structure by analyzing the coarse granularity and the fine granularity of application requirements. The method comprises the following steps: identifying and combining network protocol analysis models of the same hierarchy, constructing a tree-shaped layered structure, extracting a model identification result and an operation relationship in the tree-shaped layered structure, designing a two-dimensional mapping table, guiding input traffic through the two-dimensional mapping table, analyzing the input traffic layer by layer according to the tree-shaped structure, and positioning the next step through an operation result of each step until requirements are met. And an efficient collaborative analysis process is obtained. According to the method, the problems of high model training difficulty, high resource consumption and accumulation of errors along with layer propagation existing in a single network protocol analysis algorithm for multi-granularity and multi-layer hybrid application requirements are solved.

The invention belongs to the technical field of infrared imaging system testing, and discloses an MRTD intelligent testing method for an infrared imaging system, and the method comprises the steps: collecting the original four-rod target image data of thermal infrared imagers of the same model at different temperatures under the spatial frequency under the condition that the spatial frequency is not changed; cutting the collected original image data, and only obtaining an important area of a four-bar target image; carrying out expansion processing on the image; annotating the images, and dividing the images into a training set, a test set and a verification set; an MRTD intelligent detection model is constructed; training and accuracy testing are carried out on the MRTD intelligent detection model, label value prediction is carried out on the image data of the verification set, and the temperature difference value corresponding to the predicted label value N is the MRTD value of the thermal infrared imager under the spatial frequency. According to the method, the real-time performance of detection is ensured, and the interference of repetitive work and subjective factors in a traditional test method is reduced.

The application discloses a kind of phase correction method and system for structured light stripe profile projection, when boundary is identified, including the process of twice boundary identification, edge extraction is carried out by gray scale chart in initial boundary identification, in the second boundary identification, edge information extraction is carried out by phase calculation, boundary is identified by gray scale chart and phase diagram two directions, improve the accuracy of boundary identification, further the result of twice boundary identification is fused, so that the phase of stable area is not disturbed by jump pixel, ensure the accuracy of boundary division, based on the region after division, filtering processing is carried out to stable area, reduce the loss of feature information, provide reference information for the correction of unstable area, reduce error transmission, when the unstable area is compensated for error with stable area as reference, phase noise can be reduced, so as to reduce point cloud noise and improve precision.

Embodiments of the present disclosure relate to adaptive predictive coding. A method performed in an encoder (300) for selecting a codec mode for a current frame includes obtaining (725) a bitrate for absolute codec and predictive codec. The method comprises calculating (730) a bit rate difference based on the acquired bit rate. The method includes low-pass filtering (735) of the bit rate difference. The method comprises selecting (740) a codec mode based on the bit rate difference, the low-pass filtered bit rate difference, and a prediction mode counter.

The invention provides a swing mirror type laser terminal advance angle light path fast return mirror compensation method and system, and the method comprises the steps: detecting an advance azimuth compensation angle and a pitch compensation angle through an angle sensor disposed on a reflector of a swing mirror type laser terminal, and obtaining the azimuth angle and the pitch angle of a motor; the method comprises the following steps: establishing a nonlinear compensation model based on an advanced azimuth compensation angle, a pitching compensation angle, and an azimuth angle and a pitching angle of a motor, and compensating an advanced angle of the oscillating mirror type laser terminal based on a calculation result of the nonlinear compensation model. Coaxiality errors caused by mechanical errors, environmental disturbance and system dynamic characteristics can be effectively corrected, advanced aiming of relative circumferential speed light propagation delay between two satellites is compensated, it is ensured that the laser terminal always points to a target in the rotating process, and high-precision positioning is achieved.

The invention relates to a microcosmic electron microscope image processing method and system based on multi-scale zooming, and the method comprises the steps: obtaining an image of a target object through a microcosmic electron microscope, storing the obtained microcosmic electron microscope image in a file, carrying out the bilinear difference value down-sampling processing of the image, and obtaining a microcosmic electron microscope image; storing the image subjected to bilinear difference downsampling processing as a low-resolution image LR; the low-resolution image LR is rotated, scaled and overturned, brightness and contrast are adjusted, random noise is added, and therefore additional variants are generated; obtaining a microscopic image data set; and a multi-scale scaling microscopic electron microscope image super-resolution MS-MESR network model is constructed. According to the method, flexible zoom multiple control can be realized, more selective results of zoom multiple are realized, and the resolution and quality of the microscopic electron microscope image and the applicability under different zoom factors are remarkably improved.

The invention discloses a medical image tool enhancement processing method based on auto-reflection reinforcement learning. The method comprises the following steps: constructing a multi-modal reasoning model for vision-language joint reasoning; and performing supervision and fine tuning on the model by utilizing a cold start data set with tool calling track marks, so that the model has basic structured reasoning capability and normative tool use behaviors. An early-stage error reasoning track and a later-stage correct reasoning track which are generated by the same input at different training check points are collected, an auto-reflection training sample is constructed, and auto-reflection fine tuning is performed on the model. And based on the reinforcement learning data set containing the real tool calling scene, constructing a reward function and executing tool enhanced reinforcement learning training. After a medical image to be processed and a corresponding natural language problem are received, the trained model can generate gradual natural language thinking, a tool calling instruction and an external tool observation result, and outputs a medical image analysis conclusion, so that refined and explainable medical visual reasoning is realized.

This invention discloses a method and apparatus for detecting lane-level turning information at at-grade intersections, comprising: preprocessing the road network, cleaning trajectory data and matching maps to connect trajectory points with their respective road segments; an intersection guidance zone identification step, considering the intersection trajectory distribution characteristics and vehicle straight-ahead stopping information at stop lines, to detect the intersection guidance zone; an intersection lane identification step, based on the intersection guidance zone identification results, extracting trajectories related to the guidance zone and performing multi-angle denoising, thereby inferring the guidance zone lanes; and an intersection lane-level turning information detection step, considering the intersection lane identification results, performing lateral and longitudinal statistics on the turning trajectories of each lane, and employing a supervised learning method to detect the intersection lane-level turning information. This invention is the first attempt to identify intersection lane-level turning information based on crowdsourced trajectories, making up for the lack of systematic research in lane-level turning information mining in existing works.

The application discloses a non-invasive AC voltage measurement system based on capacitive coupling and reference signal injection, belongs to the field of electric power measurement, and comprises a detection end, a measurement unit and a shielded cable. The detection end comprises two half-circular conductive electrodes which can be opened and closed and can be clamped on an insulated wire. The measurement unit comprises an instrument amplifier, a reference signal source, an analog-to-digital converter, a processor and an isolation power supply. The shielded cable is used for connecting the detection end and the measurement unit, and the shielding layer of the shielded cable is connected with the injected reference signal. The application adopts double-frequency injection and separation calculation, utilizes frequency separation of the reference signal and the power frequency signal, avoids similar numerical interference, reduces error propagation, and the frequency of the reference signal avoids multiples of the power frequency signal, so that fast Fourier transform extraction is facilitated.

This invention provides a method and system for correcting the modeling of lightning incidence angle distribution, comprising: constructing a lightning incidence angle probability density model based on the lightning incidence angle; constructing multiple candidate sets of model parameters based on the parameters of the lightning incidence angle probability density model; calculating the normalization constant of each candidate set of model parameters; calculating the goodness of fit and binning error corresponding to each candidate set of model parameters based on the lightning incidence angle probability density model; and determining the optimal parameters by comprehensively comparing the goodness of fit and binning error. This improves the fitting accuracy and stability of the incident angle statistical distribution to measured data.

This invention belongs to the field of radar information processing technology and discloses a pre-training method for a multimodal model of radar target detection, tracking, and recognition. The invention includes collecting model pre-training data; converting the collected radar information processing data from various models into a unified data format through data preprocessing to serve as input data for the pre-trained model; achieving simultaneous input of multimodal data, including images, tracks, and category text, through a unified pre-training framework, and completing joint representation learning of the input data by the pre-training framework; and designing a multi-task joint learning optimization algorithm to achieve parallel training of radar target detection, tracking, and recognition tasks. This invention achieves integrated processing of radar target detection, tracking, and recognition tasks by establishing a multi-task learning mechanism under a unified model training framework, effectively improving the overall performance of radar information processing. Simultaneously, by enhancing versatility and generalization, the model is applicable to intelligent processing tasks of various radar models.

The invention belongs to the technical field of atmospheric detection and remote sensing, particularly relates to an aerosol complex refractive index self-adaptive adjustment method and system based on a star-shaped structure, and provides a star-shaped structure function model and a comprehensive weight factor to fully consider the complexity of coastline space change. And when the aerosol lookup table is constructed, the imaginary part of the aerosol complex refractive index is adaptively adjusted according to the weight factor, so that the precision of optical thickness inversion of the atmospheric aerosol in the offshore region is improved.

The invention relates to a radiation source positioning method. The method comprises the following steps: acquiring a mask map containing building distribution information in a target area; drawing a grey-scale map containing the position information and the received signal intensity of each sampling point in the target area as an electromagnetic sampling map; fusing the mask map and the electromagnetic sampling map to obtain an input image sample; analyzing the input image sample by using a radiation source positioning model to obtain a radiation source position coordinate; the radiation source positioning model comprises an encoder and a decoder which are used for extracting multi-scale fusion features of the input image samples; and the space mapping layer is used for predicting the position coordinates of the radiation source according to the multi-scale fusion features. The position of the radiation source can be effectively and accurately obtained.

This invention belongs to the field of underwater acoustic communication technology. Specifically, it relates to a time-frequency joint equalization method for GMSK underwater acoustic communication. The method includes: converting and mapping the transmitted signal sequence, then modulating it according to the GMSK signal modulation method based on Laurent decomposition to obtain a GMSK complex baseband signal, and modulating it onto a carrier as the final transmitted signal, which is then transmitted into the underwater acoustic channel; the underwater acoustic channel receives the final transmitted signal, uses it as the received signal, and performs bandpass filtering, two-dimensional time-frequency synchronization, matched filtering, time-frequency equalization, demapping, and decision processing on the received signal to obtain the decided symbol, thus completing the time-frequency joint equalization.

The invention relates to a text recognition method and device, computer equipment, a storage medium and a computer program product, and the method comprises the steps: inputting a to-be-recognized text into a word segmentation network for word segmentation processing, obtaining a to-be-recognized word segmentation sequence, and then inputting the to-be-recognized word segmentation sequence into a recognition network for tag recognition. Obtaining a plurality of label links corresponding to the to-be-recognized word segmentation sequence, and finally recognizing the to-be-recognized text according to the plurality of label links to obtain a recognition result. Wherein each node on each label link represents a label of each word segmentation vector in the word segmentation sequence to be identified. According to the method, the to-be-identified text is generated into a plurality of label links, different label prediction results can be comprehensively considered, information in the text can be captured more comprehensively, limitation possibly brought by single prediction is overcome, and therefore the accuracy and robustness of overall identification are improved.

The application discloses a USBL system installation deviation angle calibration method based on an iterative reweighted two-step method, and on the basis of a traditional two-step method, estimates a transponder position and an installation deviation angle in the same closed loop iteration process, so that position error and installation deviation angle error are mutually corrected, thereby weakening the error transmission problem in the traditional two-step method; in view of the problem that different measurement points have different qualities, a comprehensive weight is constructed for each measurement point to suppress the influence of abnormal values, gross error points and measurement points with poor geometric conditions; through the comprehensive weight, the contribution of different measurement points to position updating and installation deviation angle updating is no longer equal, thereby improving the overall estimation stability and robustness; in each iteration, the comprehensive weight is used to construct a weighted least square updating model of the transponder position and a weighted nonlinear least square updating model of the installation deviation angle, so that the alternating feedback structure and the reweighting mechanism are truly coupled, and a closed loop calibration method capable of stable convergence is formed.

The application provides a blind carrier phase recovery method and system for optical fiber communication. At the sending end, first, the QAM symbol stream is grouped, and a reference pilot with a specific fixed amplitude is inserted according to the quadrant in which the first symbol of each symbol group is located. At the receiving end, after the preliminary blind carrier phase estimation is completed, the original quadrant corresponding to the reference pilot is identified by using the minimum Euclidean distance detection mechanism, and then the actual quadrant position of the pilot in the received signal is compared, so as to analyze the phase ambiguity offset with 90 degrees as the base, and accordingly, the phase correction is implemented on the corresponding symbol group. The application effectively breaks through the four-fold rotational symmetry limitation of the traditional constellation diagram by means of amplitude modulation coding, and can accurately identify and correct the fixed-angle phase deviation occurring in the blind phase recovery process. The method can significantly suppress the error code diffusion caused by the phase ambiguity, improve the transmission performance of the system in terms of the bit error rate and the generalized mutual information, and has good algorithm compatibility and engineering implementation convenience, and the system overhead is effectively controlled.