899results about How to "Improve acquisition efficiency" patented technology

The invention discloses a method and system for generating enhanced knowledge questions and answers for heterogeneous database mixed retrieval, and the method comprises the following steps: (1) processing text data, and constructing a heterogeneous database; (2) receiving user query, and performing mixed retrieval in the heterogeneous database; the retrieval results are merged, and documents and triple information related to user query are obtained; (3) reordering the retrieved documents and triple information by using a reordering model, and selecting K retrieval results with the highest relevancy with user query; and (4) combining user query and retrieval results, constructing cue words, inputting the cue words into the large language model, and generating coherent and accurate natural language answers. By means of the method and device, more comprehensive and accurate information retrieval can be achieved, and the accuracy and reliability of knowledge question and answer results are improved.

The invention relates to a storage battery quick charging safety control method, device and equipment and a storage medium. The method comprises the following steps: modeling an equivalent circuit and a thermal equivalent circuit of the storage battery to obtain a dynamic decoupling matrix; battery safety state data are obtained through parallel calculation; performing abnormal mode recognition to obtain an abnormal mode recognition result; inputting an abnormal mode recognition result into a heterogeneous graph gate loop unit, and extracting spatial-temporal features through heterogeneous graph convolution and a GRU structure to obtain deep fault features; calculating the charging safety capacity, and carrying out multiple constraint control on the thermal management and the charge state of the battery to obtain a charging power distribution scheme; the charging power distribution scheme is input into the two-stage model prediction controller for distributed power redistribution, a charging control instruction is obtained, the battery state can be monitored in real time, the abnormal mode can be accurately recognized, self-adaptive control is carried out based on multiple constraints, and the safe and efficient fast charging process is achieved.

The invention discloses a cross-modal image text retrieval method based on deep learning, and provides a novel cross-modal feature extraction and alignment framework, efficient cross-modal retrieval is performed by learning semantic representation of an image and a text in a common feature space, the framework comprises a plurality of modules, and in a feature extraction stage, the cross-modal feature extraction and alignment framework is used for extracting and aligning the images and the texts. BERT Tokenizer and BERT Embedding are adopted for text coding, and the image feature extraction is combined with a Faster R-CNN (Convolutional Neural Network) model and a ResNet-101 model, so that the dimensional consistency of image features and text features is ensured. In the feature alignment stage, fine-grained semantic alignment of image and text features is realized through an I2T Attention module and a T2I Attention module, and the cross-modal matching precision is remarkably improved. In the correlation scoring stage, a correlation scoring matrix is generated by calculating the similarity score between the image features and the text features, so that the accuracy of cross-modal matching is improved. And further optimizing the feature alignment effect through normalization and attention redistribution. In an optimization stage, the system innovatively adopts a shared semantic and sorting loss strategy, integrates an OpenCLIP model framework, efficiently mines a semantic relationship between an image and a text from unlabeled large-scale data through comparative learning, and shows strong transfer learning ability and inference precision.

The invention discloses an intelligent filling system and method based on a large model, and the method comprises the steps: collecting the current call content, converting the call content into a text through a voice recognition large model, and obtaining a call text; carrying out preprocessing operation on the call text, and carrying out data analysis on the preprocessed call text by utilizing a text analysis large model to obtain an analysis result; making a plurality of work order templates, forming a work order database, and matching the analysis result with the work order database to obtain a current work order template; automatically filling an analysis result into a current work order template; and obtaining the emergency degree of the current service according to the analysis result, updating the emergency degree to the current work order template, and obtaining the service work order of the current call. The invention relates to the technical field of customer service, and solves the technical problems of single application scene and low intelligent degree in the existing intelligent form filling method.

The invention relates to an electrochemical impedance acquisition method based on a time sequence model and an electrochemical impedance model. The method comprises the following steps: constructing an experimental test matrix taking a battery material system, temperature, SOC and the like as variables; performing charging and discharging test on the battery according to the experimental test matrix; after the battery stands, current excitation is applied through an electrochemical workstation, real electrochemical impedance spectroscopy data are obtained, parameter identification is carried out, and a model training label is extracted; extracting features of the process data of the charging and discharging test, and constructing an electrochemical impedance data set; based on a time sequence pre-training basic model, training by utilizing the electrochemical impedance data set to obtain an electrochemical impedance model; and predicting the electrochemical impedance spectrum of the battery to be detected by adopting the trained electrochemical impedance model. Compared with the prior art, the method has the advantages that the time sequence large model and the electrochemical impedance model are combined, efficient acquisition and accurate analysis of the electrochemical impedance spectrum are realized, and the method has wide application potential in real-time monitoring and dynamic application environments.

The invention discloses a roadbed compaction degree monitoring method and system based on airborne laser radar, and belongs to the technical field of roadbed compaction degree intelligent monitoring, the monitoring method comprises the following steps: using an unmanned aerial vehicle to collect roadbed data, and constructing a historical database and a navigation database; according to the timestamps of the data in the historical database and the navigation database, all the timestamps in the two databases are combined and subjected to duplicate removal, and a complete time axis is constructed; according to the three-dimensional point cloud data subjected to navigation deviation parameter correction, relevant parameters of the roadbed compactness are obtained through calculation, and a relevant parameter database is constructed; fitting the relevant parameter data in the relevant parameter database and the three-dimensional point cloud data in the historical database to obtain a roadbed compactness characterization parameter model; and finally, combined with two-dimensional image data in a historical database, generating change images of the roadbed compactness in different time periods. According to the invention, intelligent monitoring of the roadbed compactness can be effectively realized.

The embodiment of the invention discloses a network security alarm processing method and device, electronic equipment, a storage medium and a computer program product. The method comprises the following steps: acquiring a network security alarm; retrieving a target text vector associated with the network security alarm in a network security domain knowledge base; the network security domain knowledge base is a knowledge base constructed after vectorization processing is carried out on network security corpora; based on the target text vector, an alarm vector corresponding to the network security alarm, a preset tool library and a large language model, obtaining a research and judgment processing flow of the network security alarm; executing a research and judgment processing flow; and obtaining a research and judgment result of the network security alarm based on an execution result of the research and judgment processing flow.

The invention discloses an equipment maintenance knowledge representation method combining knowledge distillation and semantic analysis. The method comprises the steps of collecting structured data, unstructured text data and multi-modal data, constructing a semantic knowledge graph by using a natural language processing technology, extracting visual knowledge by using a computer vision technology, and fusing the visual knowledge with semantic knowledge to form a multi-modal knowledge base. And performing interactive knowledge distillation on the text data by using a large language model, generating optimized knowledge key points, and fusing the optimized knowledge key points with the multi-modal knowledge base to generate a comprehensive knowledge base. Knowledge is hierarchically organized according to the domain ontology and stored in a graph database, a natural language query interface is constructed, intelligent answering is performed by integrating technologies, question and answer corpus training is automatically generated, and automatic question and answer services are provided for users. According to the method, heterogeneous data integration, multi-modal knowledge fusion, hierarchical knowledge organization, natural language interaction and man-machine collaborative optimization are realized, and an intelligent equipment maintenance knowledge base can be efficiently constructed.

The invention provides a method for dynamically constructing a coal mine geology digital twinborn body based on multi-source heterogeneous information. Comprising the steps of AIGC-assisted geological result reading, end-to-end structured information self-adaptive generation, primary geologic body model construction through a generalized triangular prism algorithm, geometric smooth subdivision based on Kriging interpolation, special geological element dynamic updating and coal mine digital twinborn matrix generation. According to the method, the AIGC is introduced to assist in reading the geological result innovatively from the remarkable characteristics of the data intensive normal form of the geological information, and intelligent generation of structured information matched with an automatic modeling updating algorithm is realized through big data assimilation of multi-source heterogeneous information; a quick response process of information processing, primary geologic body generation, geometric smoothing, abnormal geologic body encryption and global geologic body construction is formed, and a high-precision coal mine geologic digital model is dynamically constructed by utilizing geologic information of different geologic survey stages. And the requirements of high precision, renewability and automation of the coal mine digital twinborn body are met.

The invention provides a multimedia resource playing method and device, equipment and a storage medium, and belongs to the technical field of multimedia. The method comprises the steps that in the process of playing a multimedia resource, an interaction object is displayed, the multimedia resource is used for introducing a target object, and the multimedia resource is associated with a first virtual resource related to the target object; in response to an interaction operation on the interaction object, playing the multimedia resources at multiple speeds; in response to the multimedia resource playing to the preset progress, receiving success information is displayed, and the receiving success information is used for indicating that the first virtual resource is successfully received. According to the technical scheme, a scheme for playing the multimedia resources at multiple speeds through the interactive operation can be provided, and the information acquisition efficiency and the man-machine interaction efficiency are improved.

The invention provides a resource access control method and device, electronic equipment, a computer readable storage medium and a computer program product. The method comprises the steps that under the condition that a resource access request initiated by a target object is received, a target expression matched with the target object is acquired from a permission language library; obtaining metadata information associated with the target object, and generating a query statement based on the metadata information and the target expression; and accessing a resource library based on the query statement to obtain target resource information used for responding to the resource access request. According to the method and the device, the calculation amount for matching and filtering the data can be greatly saved, so that the resource access efficiency can be effectively improved.

According to the online enrichment and automatic sorting device and method for deep-sea microorganisms in the high-pressure environment, multi-stage enrichment is carried out by simulating the high-pressure and special-temperature environment of deep sea, and meanwhile, based on environment index changes observed online in the enrichment device, the deep-sea microorganisms in the high-pressure environment can be automatically sorted; the optimal time point of microbial passage is automatically judged, and a culture is transferred to a next-stage enrichment device, so that the problems that in a high-pressure environment, microbial enrichment is difficult to transfer at the optimal time point, manual operation is complicated and the like are effectively solved, and the early-stage enrichment efficiency of obtaining a single species of deep-sea microorganisms is improved; besides, automatic streak separation and automatic secondary sorting of the enriched substances are carried out in a culture medium in a high-pressure environment, and specific single colonies are automatically transferred to a pure culture kettle, so that the problems of overlapping of single colonies, low separation degree, complicated manual operation, difficulty in automatic secondary sorting in the high-pressure environment and the like in an existing high-pressure sorting technology are solved; and the high-pressure environment deep-sea microorganism sorting efficiency is greatly improved.

The invention belongs to the technical field of autonomous vehicles, and particularly relates to a vehicle end visual angle accident scene restoration method based on an automobile data recorder video. Comprising the following steps: step 1, preprocessing an accident video; 2, performing semantic understanding on the scene; step 3, extracting pixel tracks of the traffic participants; step 4, carrying out joint calibration on the inverse perspective mapping parameters; 5, carrying out inverse perspective transformation; step 6, carrying out post-processing on the road contour and track data; for accident video data with unknown and different camera parameters and lack of multi-source synchronization information, the invention provides a vehicle-end visual angle accident scene restoration framework taking a camera inverse perspective mapping parameter joint calibration strategy as a core. Scene space-time information including a road contour and a traffic participant track can be extracted from an automobile data recorder video without any prior information support, accident scene data is established for testing an automatic driving automobile, and the testing process of the automatic driving automobile can be effectively improved.

The invention provides an on-orbit relative radiometric calibration coefficient obtaining method and device of a spaceborne camera, and the method comprises the steps: obtaining an on-orbit relative radiometric calibration coefficient relation model which is used for representing the relation between on-orbit relative radiometric calibration coefficients under different imaging parameters; determining a bright field uniform area and a dark field uniform area of the spaceborne camera under the first imaging parameter; on the basis of the bright field uniform region and the dark field uniform region, obtaining an on-orbit relative radiation calibration coefficient under the first imaging parameter; on the basis of the on-orbit relative radiometric calibration coefficient relation model and the on-orbit relative radiometric calibration coefficient under the first imaging parameter, on-orbit relative radiometric calibration coefficients under other imaging parameters are obtained, and the technical problem that in the prior art, the on-orbit relative radiometric calibration coefficient obtaining efficiency is low is solved.

The invention discloses an intelligent reply method and device based on natural language processing, equipment and a medium. The method comprises the steps that an initial knowledge graph corresponding to historical reply information is constructed through a knowledge graph construction rule, and a corresponding common reply link is generated; and judging whether the initial knowledge graph contains target reply information matched with the question information input by the user or not according to a natural language matching model, if so, obtaining the target reply information and outputting the target reply information, and if not, sending the question information to a service window so as to transfer service personnel to answer. According to the intelligent reply method, the historical reply information can be analyzed based on the natural language, the initial knowledge graph is constructed, and the common reply link corresponding to the initial knowledge graph is generated, so that a user can click and quickly obtain the reply content; the target reply information matched with the question information input by the user is accurately acquired based on the natural voice matching model and the initial knowledge graph, and the efficiency and accuracy of acquiring the reply information are improved.

The invention relates to the technical field of artificial intelligence data processing, in particular to an optimization method for agricultural intelligent questions and answers based on LLM. By combining natural language processing, deep learning and agricultural knowledge, the invention aims to improve the performance of the agricultural intelligent question-answering system and provide more accurate, efficient and personalized services for users. Farmers can quickly obtain required agricultural knowledge and technology through the intelligent question-answering system without looking up a large number of literatures or asking for experts.

The embodiment of the invention relates to the technical field of information, in particular to an AI-based clothing commodity display method and system. The method comprises the following steps: receiving a three-dimensional model of a target garment, and generating a virtual model; receiving commodity pictures of a plurality of other garments in a display page and current login user information, generating a target model posture according to the plurality of commodity pictures and the user information, and adjusting the posture of the virtual model according to the target model posture; fusing the three-dimensional model of the target garment with the virtual model to obtain a fused model; generating a scene according to the plurality of commodity pictures and the user information, and placing the fusion model in the scene; and creating a view-finding camera in the scene according to the target model posture, and obtaining a target display picture according to the view-finding camera.

The invention relates to the field of content generation, and particularly discloses an automatic content generation system and method based on user data, which comprises the following steps: firstly, acquiring keyword text data input by a user and a history browsing news manuscript, then performing feature extraction and correlation analysis on the keyword text data and the history browsing news manuscript by using a deep learning technology, and finally, generating a content through a generator. According to the method, the user personalized news manuscript recommendation is generated, so that the requirements and interests of the user are more accurately understood, the problem of information overload is effectively reduced, and the user experience and the information acquisition efficiency are improved.

The invention relates to the field of resource prediction, in particular to a data transmission resource prediction method based on historical data, and the method comprises the steps: obtaining the historical data transmission data when a satellite runs to different transit time windows, thereby analyzing the characteristics of data transmission behaviors, calculating the fuzzy representation parameters of the data transmission behaviors, and setting labels for sites; and determining the dimension of historical data transmission data needing to be called based on the fuzzy representation parameter, calling the historical data transmission data of the corresponding dimension, adopting a pre-trained prediction model to predict the data transmission resource demand quantity in each historical period, and verifying the validity of the prediction model for the site based on the data transmission resource receiving quantity in the corresponding historical period. And determining whether a prediction result of the prediction model is output or not, determining the demand quantity of the data transmission resources, or determining the demand quantity of the data transmission resources of the station based on a fluctuation interval of the historical data transmission resource receiving quantity. According to the method, the resource prediction method is determined by analyzing the historical data of the site, the prediction accuracy is improved, and the efficiency and the utilization rate are improved.

The invention provides a fracture morphology automatic identification and learning method based on multi-modal fusion learning, and relates to the field of metal fracture morphology analysis, and the method comprises the steps: S1, constructing a data set; s2, carrying out fracture image shooting and image annotation, and after annotation is completed, dividing a data set into a training set, a verification set and a test set; step S3, carrying out multi-modal fusion learning; and S4, feature importance analysis and SHAP analysis are carried out. According to the method, image recognition and machine learning are combined, accurate prediction of dimple fracture and brittle fracture areas of the aluminum alloy is achieved, meanwhile, key factors influencing the metal fracture morphology can be effectively obtained, and therefore an important reference basis is provided for effective optimization of material performance.

The invention relates to the technical field of equipment fault detection, and discloses an equipment fault prediction and diagnosis method which is applied to an ATE system, and the method comprises the steps: collecting the real-time operation data of equipment through a plurality of sensors, and dynamically scheduling the sensors through an adaptive algorithm, so as to adjust the data collection frequency and range according to the state of the equipment; the collected data is preprocessed; training a plurality of machine learning models based on the historical fault data and the operation data, and deploying the models to an ATE system for real-time analysis; predicting an equipment fault according to the model, generating alarm information when a potential fault is detected, and sending the alarm information to an operator through a user interface or a notification system; generating a fault diagnosis report according to an analysis result and providing a repair suggestion; recording a fault instance and data, and updating and retraining the machine learning model to improve prediction and diagnosis accuracy; and generating a digital twin model of the equipment based on real-time and historical data, simulating the running state of the equipment, predicting future faults, and evaluating the health condition of the equipment.

The invention provides a training fault tolerance method and device applied to a distributed training system and a chip product, and relates to the technical field of distributed training. The method comprises the following steps: for a distributed training system comprising a plurality of computing nodes and at least one standby node, under the condition that the standby node is set as a new computing node, determining a target computing node related to the new computing node from the plurality of computing nodes; the check point data stored in the memory of the target computing node is migrated to the new computing node, and the check point data is used for indicating the state when the computing node participates in executing the distributed training task; and controlling the new computing node to participate in execution of the distributed training task based on the check point data. According to the embodiment of the invention, the check point data stored in the memory of the target computing node is migrated to the new computing node, and the check point data does not need to be read from the persistent storage system, so that the reading efficiency of the check point data is improved.

The invention discloses a mushroom yield estimation method based on three-dimensional reconstruction lossless phenotype parameter extraction, belongs to the technical field of agricultural informatization and intelligentization, and solves the problem of low mushroom yield estimation accuracy. The method comprises the following steps: performing point cloud three-dimensional reconstruction on shiitake mushrooms by using a region of interest to generate dense point cloud data; preprocessing the dense point cloud data; based on an improved CP-PointNet + + model and a clustering algorithm, segmenting mushroom point cloud, extracting multi-dimensional key phenotypic parameters of pileus diameter, thickness and stipe height, and covering main features of mushroom phenotypic information; the improved PointNet + + network comprises the following steps: adding a CBAM module to an MLP layer of the PointNet + + network, and replacing common convolution with partial convolution PConv; and taking the phenotypic parameters of the shiitake mushrooms as input variables of the neural network model, and predicting the shiitake mushroom yield. The method is suitable for phenotypic character analysis and breeding research of edible mushrooms such as shiitake mushrooms and the like.

The invention discloses a seed information question-answering system and method. The invention provides a variety information question-answering system. The system comprises a computing resource; the big language model calls computing resources to perform seed information question answering, the big language model is configured to learn knowledge related to the seed from a seed field knowledge source, and the big language model comprises an input acquisition module configured to be used for acquiring text information related to seed information questions from user input; the semantic understanding module is coupled to the input acquisition module and is configured to be used for converting the text information into vector information; the information reasoning module is coupled to the semantic understanding module and is configured to be used for generating an answer information vector based on knowledge related to the seed industry in combination with the vector information and the context, and the answer information vector comprises breeding decision information; and the result output module is coupled to the information reasoning module and is configured to be used for converting the answer information vector into answer text information.

The present invention discloses a roadbed compaction monitoring method and system based on airborne laser radar, belonging to the technical field of intelligent monitoring of roadbed compaction. The monitoring method includes using a drone to collect roadbed data, and constructing a historical database and a navigation database; according to the timestamps of the data in the historical database and the navigation database, all timestamps in the two databases are merged, duplicated, and constructed to obtain a complete time axis; according to the three-dimensional point cloud data corrected by the navigation deviation parameters, the roadbed compaction related parameters are calculated, and a related parameter database is constructed; the related parameter data in the related parameter database and the three-dimensional point cloud data in the historical database are fitted to obtain a roadbed compaction characterization parameter model; finally, combined with the two-dimensional image data in the historical database, the roadbed compaction change image under different time periods is generated. The present invention can effectively realize the intelligent monitoring of roadbed compaction.

The invention discloses an image processing method and device, electronic equipment and a computer readable storage medium, and the method comprises the steps: obtaining a to-be-processed image which comprises a to-be-recognized first card; identifying an image area where each first card is located in the to-be-processed image; based on the image area, extracting a first card image corresponding to each first card from the to-be-processed image; performing image identification processing on the first card image of each first card to obtain an identification result corresponding to each first card; based on the identification result of each first card, card information of each first card is determined; and displaying card information of each first card through a graphical user interface provided by the terminal equipment. According to the method and the device, the step of acquiring the card information corresponding to the entity card is simplified, the acquisition time of the card information is saved, the real-time performance and the responsiveness of information acquisition are improved, and the card information acquisition efficiency is improved.

The invention relates to the technical field of grassland monitoring, in particular to an adjustable grassland remote sensing monitoring system and method, and the system comprises a data preprocessing module, a variation trend analysis module, a self-adaptive remote sensing parameter configuration module, a grassland monitoring network layout optimization module and a remote sensing observation and intelligent estimation module. The data preprocessing module integrates historical remote sensing and ground actual measurement data and provides a data basis for subsequent analysis; the change trend analysis module calculates the change rate and the standard deviation of the key features and identifies the change trend of the grassland ecological environment; the adaptive remote sensing parameter configuration module adaptively adjusts remote sensing monitoring parameters, and the monitoring efficiency is improved; the grassland monitoring network layout optimization module optimizes the layout of the remote sensing detection unit and improves the efficiency of the monitoring network; and remote sensing observation and intelligent estimation use a spatial interpolation method to estimate the value of a future observation point, so that the workload of field measurement is reduced. The method is used for solving the technical problem that the current grassland monitoring management effect is poor.

The invention relates to the technical field of intelligent voice in the artificial intelligence technology, can be applied to the field of voice interaction, and discloses a voice response method, device and equipment based on a knowledge base and a storage medium, and the method comprises the steps: obtaining target voice, and converting the target voice into text content; obtaining a question statement in the text content, and encoding the question statement to generate a sentence vector; selecting a knowledge base storing the target user group data as a target knowledge base, and obtaining answer information in the target knowledge base through the sentence vector; selecting the answer information as a cue word template, and obtaining emotion information of the target voice; inputting the prompt word template, the text content and the emotion information into a preset large language model, and obtaining a reply text generated by the large language model; and processing the sound sample and the reply text of the target speaker, generating a response voice of the target speaker, and playing the response voice. According to the invention, the obtaining efficiency of the response voice of the target speaker is improved.

The invention discloses an automobile open source boundary research method based on a large language model, a medium and terminal equipment, and the method comprises the steps: carrying out the static code analysis of an extracted program file, and obtaining first open source component information; performing large language model analysis on the extracted engineering construction file and configuration file to obtain second open source component information; the firmware of the automatic driving system is analyzed, and third open source component information is obtained by jointly using an SCA software component analysis technology; analyzing the configuration file of the Apk through a large language model to obtain fourth open source component information; and performing data merging and data deduplication on the first open source component information, the second open source component information, the third open source component information and the fourth open source component information. According to the method, the construction information of the source code project and the configuration information in the firmware are analyzed through the large model technology, and then the open source component information is obtained, so that the obtaining efficiency and effectiveness of the open source information are improved.

The embodiment of the invention provides a data acquisition method and device, electronic equipment and a storage medium, and belongs to the technical field of data processing. The method comprises the following steps: receiving a business data query request, wherein the business data query request comprises a screening condition and a target business name of a target business; matching the target service name with service names in a pre-configured dimension table to obtain a matching result, the dimension table comprising a plurality of service names and a service code corresponding to each service name; and under the condition that the matching result indicates that the target service name corresponds to the first service code, acquiring data related to the target service from a database according to the screening condition and the first service code. According to the embodiment of the invention, the service data acquisition efficiency can be effectively improved through the pre-configured dimension table, and the query difficulty is reduced.