135results about How to "Lower build costs" patented technology

The invention discloses a transformer noise detection method based on season trend decomposition and time sequence prediction, and belongs to the technical field of transformer noise anomaly detection. Comprising the following steps: S1, acquiring noise data of a transformer, and preprocessing the noise data; s2, adopting an STL decomposition model to obtain a seasonal component, a trend component and a residual component; s3, performing Transform time sequence prediction on the seasonal components and the trend components to obtain prediction results of the seasonal components and the trend components; s4, performing anomaly detection on prediction results of the seasonal components and the trend components, and outputting a detection result; and S5, carrying out weighted composition on the detection results of all the components, and outputting a diagnosis result. According to the method, the time sequence data of the noise of the transformer substation is decomposed, predicted and subjected to abnormal point detection, and reliable technical support is provided for state monitoring and fault prevention of transformer substation equipment.

The embodiment of the invention provides a data construction method, a code question-answering method, a task platform and a code question-answering system. The data construction method comprises the steps of obtaining a first code snippet and an initial sample question for the first code snippet; rewriting the initial sample problem by using a text processing model and taking directory structure information of a sample code warehouse in the code library as context information to obtain a sample problem; based on the first code snippets and the sample questions, recalling related second code snippets from a code library; the text processing model is utilized, the second code snippets serve as context information, sample answers corresponding to the sample questions are generated, the sample questions and the sample answers are sample data used for training a code question and answer model, and the code question and answer model is suitable for a code warehouse. The data construction efficiency is improved, the data construction cost is reduced, the model obtained through training is adaptive to the code warehouse, and the user experience is improved.

The invention discloses a Chinese named entity recognition method based on hierarchical label enhancement contrast learning. The method comprises the steps that 1, labels and characters are coded through a pre-training language model; 2, deep fusion is carried out on coding results through a cross-modal Transform network, and enhanced features are generated; 3, maximizing the similarity between the matched sentences and labels through a sentence level contrast learning method; 4, minimizing the difference between the character representation and the label enhancement representation through a lexical-level contrast learning method; 5, coding and decoding are carried out by using BiLSTM and CRF, and a final entity category prediction sequence is generated; and 6, constructing a total loss function and training the network. The Chinese named entity recognition method effectively improves the accuracy and generalization ability of Chinese named entity recognition, especially has excellent performance when processing complex scenes such as polysemy words and homosynonyms, does not need to depend on a dictionary, is high in reasoning speed, and is suitable for practical application scenes of large-scale text processing.

The invention provides an automatic quality improvement method for a subject big question bank and a related device, and the method comprises the steps: obtaining a unit knowledge point list and a question sample of a specified unit, and constructing a cue word according to the question sample; taking the unit knowledge point list of the specified unit and the question sample as the context of a preset basic model, repeatedly calling a single question generation instruction according to the cue word, and generating a first question bank containing at least two questions; calling a question duplicate removal instruction to perform question duplicate removal on all questions in the first question bank, if repeated questions exist, calling a rewriting instruction to rewrite the repeated questions, and generating a second question bank without the repeated questions; and finally, calling a question quality inspection instruction to inspect and inspect the knowledge points of all questions in the first question bank or the second question bank and the answer accuracy so as to improve the question bank quality. According to the technical scheme, full-chain intellectualization from question construction to quality inspection can be realized, the construction cost is effectively reduced, and the question bank quality is improved.

The invention provides an interactive task knowledge base construction method and device based on a large language model, and belongs to the technical field of semantic communication, and the method comprises the steps: generating a task corpus request instruction based on a first keyword prompt template, inputting the task corpus request instruction to the large language model, and obtaining a task corpus of a target task; matching a target standard task corpus according to the task corpus request instruction, performing text comparison analysis on the task corpus by utilizing the target standard task corpus, and judging whether the task corpus meets a preset task corpus requirement or not; and under the condition that the requirement is met, generating an entity extraction instruction based on the second keyword prompt template, inputting the entity extraction instruction into the large language model, and obtaining key semantic information of the task corpus so as to generate a task knowledge base of the task target. Compared with a traditional manual or semi-automatic method, large-scale data can be processed more quickly, and the fast-paced knowledge updating requirement is met.

The embodiment of the invention provides a diffusion model construction method and device, equipment and a storage medium. The method comprises the steps that a pre-training model is acquired, a target model structure of a target diffusion model is constructed, and the target model structure comprises a basic network and a target up-sampling stack network; performing initialization operation on the target model structure through the pre-training model to obtain an initial diffusion model; sequentially carrying out time sequence modeling processing and text embedding processing on the initial diffusion model to obtain a processed initial diffusion model; and training the modeled initial diffusion model according to a training set to obtain the target diffusion model, the training set comprising a plurality of pieces of text information, and the target diffusion model being used for analyzing the plurality of pieces of text information to obtain corresponding video information. The method aims at generating the diffusion model which is excellent in performance and universal, reduces the construction cost of the diffusion model, and can meet the requirements of scenes such as the financial field.

The invention provides a semantic navigation map construction method and device based on a depth camera and a storage medium. The method comprises the following steps: acquiring a depth image and image data of an environment by using a depth camera, and generating point cloud data; processing the point cloud data by using a three-dimensional object detection model so as to identify different types of objects in the environment and generate point cloud data with semantic tags; converting the point cloud data with the semantic tag into single-line laser data with the semantic tag, and inserting the single-line laser data with the semantic tag into a local semantic sub-map; when the local semantic sub-maps meet preset conditions, the local semantic sub-maps are inserted into a sub-map queue, and the local semantic sub-maps are spliced into a global map; in the map generation process, the loop detection technology is used for recognizing the previous scene passed by the robot to correct the pose of the robot. According to the method, the map construction cost and computing resources are reduced, and the real-time performance of map construction and the adaptability of a dynamic environment are improved.

The invention relates to the field of liquid on-line monitoring, and provides an electrochemical sensor for measuring conductivity and dielectric constant of liquid, comprising: a probe comprising a first electrode plate, a second electrode plate and a third electrode plate which are successively arranged; the measuring circuit comprises an RC circuit, an amplifying circuit, a comparison circuit and a NOT gate circuit, the NOT gate circuit is used for obtaining a fourth voltage u4, the duration time of a high level in the fourth voltage u4 is first time t1, the duration time of a low level in the fourth voltage u4 is second time t2, and the first time t1 is a second time t2; the first time t1 and the second time t2 are used for determining the conductivity and the dielectric constant. The dielectric property is determined through accurate time measurement, the measurement technology and high time / bandwidth are integrated together, and very high precision and repeatability are achieved. Wherein the accuracy of the conductivity can reach piezemens / meter (pS / m), and the accuracy of the dielectric constant can reach 0.0001.

The invention discloses a cable fault abnormity monitoring method and system, and relates to the technical field of cable monitoring, and the method comprises the steps: obtaining the historical fault data of a to-be-monitored cable, and carrying out the analysis of the historical fault data according to the fault type, and obtaining the fault electric data of the corresponding fault type; analyzing the fault electric data to screen monitoring data items of the monitoring equipment; acquiring attenuation properties of the corresponding data items in the cable according to the monitoring data items, determining data propagation distances of the monitoring data items according to the attenuation properties, and recording the data propagation distances as effective propagation data; setting monitoring points for the cable based on the effective propagation data to obtain monitoring point data; and detecting the cable based on the monitoring point data to obtain monitoring data, determining whether the cable has a fault according to the monitoring data, and outputting a corresponding judgment result. The effect of reducing the construction cost of the monitoring system is achieved through reasonable setting of the monitoring points.

A microfluidic chip and detection system are disclosed. The microfluidic chip includes: a sample storage chamber for storing the sample to be dispersed; a continuous phase storage chamber for storing the continuous phase; a droplet generation array for generating two-phase microdroplets from the sample and the continuous phase; and a droplet spreading chamber for spreading the generated two-phase microdroplets. By integrating the sample storage chamber, continuous phase storage chamber, droplet generation array, and droplet spreading chamber into a single integrated structure, the enclosed space effectively reduces the risk of droplet breakage and reagent contamination during droplet transfer to the droplet spreading chamber. The detection system includes: a centrifugal drive shaft and a centrifugal loading disk; several detection units fixed on the centrifugal loading disk; and a microfluidic chip fixed on each detection unit. The detection unit integrates a temperature control module, an excitation light module, and an optical detection module, enabling the entire process from droplet generation to detection, reducing the overall construction cost of the detection system and achieving device miniaturization.

The invention discloses a corpus standardized template construction and file automatic generation method and system, and the system comprises a template construction module, a file generation module, a verification module, a user interface module, a storage and management module, an extension and optimization module, and a quality evaluation module. The standardized management of the corpus is realized, the problem that the structure and format of the corpus lacks unified standards in the traditional method is solved, and the compatibility and sharing of the corpus are improved. Meanwhile, data cleaning, format conversion and multi-modal data integration technologies are combined, corpus files meeting template requirements are automatically generated, the data processing flow is simplified, and the risk of data inconsistency or loss is reduced.

This invention discloses an optimization method for a dual-source FSI ranging system based on wavelet synchronous compression transform, mainly addressing the problems of high system complexity and poor performance in existing dual-source FSI ranging systems when separating mixed interference signals. The solution includes: constructing a dual-source FSI ranging optical path to generate mixed interference signals; building a signal processing module based on the wavelet synchronous compression transform algorithm; sending the mixed interference signals to the signal processing module through a data acquisition device in the ranging optical path; using wavelet synchronous compression transform to separate the interference signals generated by the two lasers in the frequency domain; then achieving separation of the individual interference signals by reconstructing the frequency bands of the two interference signals; and finally using the separated two interference signals to calculate the distance and obtain the final ranging result. This invention can guarantee ranging accuracy without increasing the complexity and design cost of the dual-source FSI ranging system.

The invention discloses a drug for constructing a venous inflammation model in the field of phlebitis model technology, comprising the following preparation steps: S1: preparing a carbomer hydrogel, comprising the following steps: step 1: putting 2.5g of carbomer and 19.5ml of pure water into a beaker and stirring to form a mixture; step 2: letting the mixture stand and stirring several times at intervals of 5 to 10 minutes; step 3: adding several drops of triethanolamine into the beaker; step 4: adding 5ml of pure water into the beaker, then adding 0.5ml of antibacterial agent, and then stirring; S2: injecting 25ml of 20% mannitol into the carbomer hydrogel and then standing; S3: injecting 1.5μl of TBHP stock solution into the carbomer hydrogel, and uniformly mixing the carbomer hydrogel with TBHP. The present invention uses carbomer hydrogel to uniformly mix mannitol and TBHP; and compared with injecting drugs into blood vessels, injecting into the perivascular area is not only easier to operate and has a high success rate, but also can avoid the drug from spreading to the whole body with the blood, causing systemic toxicity.

The invention relates to the technical field of virtual games, and discloses a self-learning virtual display method and system based on external cause stimulus input, and the method comprises the steps: monitoring and obtaining a game process state and external cause stimulus input data after normalization quantization processing in real time; generating corresponding original picture data according to the game process state and the exogenous stimulus input data; configuring a choice box to carry out frame selection on the original picture data to obtain sampling picture data; inputting the sampled picture data into the trained mode recommendation model, and obtaining a classification result output by the mode recommendation model; configuring a game virtual mode according to a classification result; a large number of training samples with flexible and variable specifications can be provided for an original mode recommendation model in a training stage, sampling picture data of a large number of players is obtained in combination with a cloud end, classification accuracy is stably improved, individual requirements of the players are considered during classification, game experience of the players is comprehensively improved, and the user experience of the players is improved. And a player can obtain better game experience in the game.

The present application provides a kind of topography probability distribution intelligent identification method based on lightweight convolutional neural network, constructs special terrain identification network by fusing the advantages of MobileNetV3 and ShuffleNetV2, improves model accuracy by combining channel attention mechanism and knowledge distillation technology, realizes the real-time inference of embedded platform using structured pruning and INT8 quantization.At the same time, this method not only outputs the probability distribution of five types of terrain and the confidence evaluation based on information entropy, but also realizes the unsupervised ROI automatic positioning through the improved Grad-CAM++ algorithm, without additional labeling of boundary box data.The present application can realize high-precision, high-robustness terrain identification with very low hardware cost and computing overhead, and provide reliable environment perception basis for exoskeleton proactive adaptive control.

The embodiments of this specification provide data construction, code question and answer methods, task platforms, and code question and answer systems, wherein the data construction method includes: obtaining a first code snippet and an initial sample question for the first code snippet; using a text processing model, using the directory structure information of a sample code warehouse in a code library as context information, rewriting the initial sample question to obtain a sample question; based on the first code snippet and the sample question, recalling a related second code snippet from the code library; using a text processing model, using the second code snippet as context information, generating a sample answer corresponding to the sample question, wherein the sample question and the sample answer are sample data for training a code question and answer model, and the code question and answer model is applicable to a code warehouse. The efficiency of data construction is improved, the cost of data construction is reduced, and the trained model is adapted to the code warehouse, which improves the user experience.

The present application relates to the field of artificial intelligence technology, and provides a method, device, electronic device and storage medium for automatically constructing a knowledge base. The method includes: in a large language model, obtaining an initial text through a dialogue system of the large language model; when the initial text is not in the form of a question-answer pair, using a pre-constructed first prompt template to clean the initial text to obtain a first usable text; using a pre-constructed second prompt template to extract structured knowledge from the first usable text to obtain the initial information required to construct a knowledge base; using a pre-constructed third prompt template to perform knowledge fusion and verification on the initial information to obtain target information, and construct a knowledge base based on the target information. The present application solves the problem of high cost of constructing a knowledge base in the prior art.

The invention discloses a big data platform construction method and device, computer equipment and a storage medium. The method comprises the steps that a first server cluster and a second server cluster are configured; configuring corresponding software for the first server cluster and the second server cluster respectively, and obtaining first software corresponding to the first server cluster and second software corresponding to the second server cluster; if the first software and the second software have the same software type and different versions, the first server cluster, the first software and the first operating system form a first cluster, and the second server cluster, the second software and the second operating system form a second cluster; therefore, a credential big data platform can be constructed on a hybrid cluster composed of the first cluster and the second cluster. According to the method and the device, the non-creative server cluster can be added on the basis of the creative server cluster, so that the creative big data platform can be constructed on the hybrid cluster formed by the creative servers and the non-creative servers, and the flexibility of server configuration is improved.

The application discloses a digital twin system construction method and device, electronic equipment and a storage medium, and relates to the technical field of electronic information. The method comprises the following steps: acquiring system description data corresponding to a target system; the target system comprises a plurality of nodes which are connected to each other in communication; performing analysis and processing on the node behaviors based on the system description data by using a twin system analysis model to obtain a node behavior model of each node; the node behavior model is used for indicating the node configuration and the node behavior of the corresponding node; instantiating a virtual node corresponding to each node in a virtual digital system; and configuring the virtual node corresponding to each node based on the node behavior model of each node to obtain a target digital twin system corresponding to the target system. According to the method, the cost and difficulty of constructing the target digital twin system are reduced.

The embodiment of the invention provides a prediction model generation method and device for power grid dispatching future state early warning, and the method comprises the steps: collecting historical data which may face scenes in the dispatching process of a power grid enterprise and cost data of investment future state early warning; processing the historical data and the cost data through a z-score standardization method, and scaling the data of different orders of magnitude to the same order of magnitude; performing feature selection on different prediction targets by adopting Elastic Net to obtain respective high-correlation feature data sets, splitting the high-correlation feature data sets into training sets and test sets, and training a BP neural network model by using the training sets to obtain prediction models of different prediction targets, and performing performance test on the prediction model through the test set to obtain a final prediction model. The calculation speed and the decision-making efficiency can be improved.

The invention provides a multi-camera real-time mapping method and device, electronic equipment and a storage medium, and the method comprises the steps: obtaining sensor data and multi-camera data; performing pre-integration processing on the sensor data to obtain pre-integration data; performing feature extraction on the multi-camera data to obtain feature point data; initializing the pre-integral data and the feature point data to obtain initialized data; fusing the pre-integral data and the feature point data to obtain first fused data; and obtaining a mapping result according to the initialized data and the first fusion data. By implementing the application, real-time mapping in the automatic driving process can be fully covered through multi-dimensional processing of the sensor data and the multi-camera data, the construction cost of the automatic driving system is reduced on the premise of reducing the use of the sensors, meanwhile, the multiple cameras can be coordinated, the influence of the external environment on camera acquisition is avoided, and the user experience is improved. The problems of divergence, fuzziness and the like of acquired data are avoided, and the measurement precision and the mapping precision are improved.

The invention relates to an intelligent decision-making method for a block chain application design mode based on text classification, and the method comprises the steps: constructing an LDA-BERT semantic analysis model which combines the topic extraction capability of LDA and the deep semantic understanding capability of BERT, carrying out the improved fuzzy C-means clustering according to the analysis result of the model, and carrying out the fuzzy C-means clustering. And then, a model for intelligent decision making of a block chain application design mode is carried out, and before the model is constructed, data about the block chain design mode is collected and sorted, including problems and block chain design modes corresponding to the problems and the like. The constructed intelligent classification model of the block chain application design mode based on text classification can analyze a new problem and output an analysis result. According to the method, the model for intelligent decision making of the block chain application design mode based on text classification is realized, data information and results in the model operation process are collected, block chain design mode decision making can be better carried out, and therefore more references are provided during block chain architecture design.

The application belongs to the technical field of databases, and specifically discloses an efficient GPU near neighbor graph index construction method based on data locality, which comprises the following steps: dividing index original data into multiple subsets, and performing data processing on the multiple subsets to obtain rearranged data; constructing subgraphs corresponding to each subset in the rearranged data, and constructing a graph index based on the subgraphs; and refining the structure of the graph index to obtain a final graph index. The method can effectively alleviate the problems of insufficient memory resources and excessive calculation cost in the subgraph merging stage, significantly reduce the construction cost of the graph index, and effectively extend to large-scale data sets to improve scalability.

The invention discloses a construction method and system of a medical dialogue thinking chain data set. The method comprises the following steps: firstly, preprocessing original medical session data, and screening a target session round to obtain structured data; then, in combination with downstream tasks such as medical dialogue generation and disease diagnosis assistance, a forward reasoning Prompt template is constructed, and the problems that an existing method depends on the response of a real doctor, and a thinking chain lacks clinical authenticity and depth are solved. And through a forward reasoning-review-reflection mechanism, on the basis of structured data and a Prompt template, generating a thinking chain and reaction data which conform to real clinical reasoning logic, and finally forming a medical dialogue thinking chain data set. The data set is consistent with clinical diagnosis logic, and does not need to depend on the reaction of an original doctor. The diagnosis accuracy, the response pertinence and the interpretability of the medical dialogue system can be effectively improved, and powerful support is provided for constructing the reliable and interpretable medical dialogue system.

The invention relates to a method for dual-mode detection of fluoroquinolone antibiotics by using an aptamer-based test strip. The preparation method comprises the following steps: uniformly mixing AuNRs / PDA with a streptavidin solution, carrying out oscillation incubation, adding a biotin-modified aptamer Bio-Apt, and carrying out incubation again, so as to obtain an AuNRs / PDA-Apt signal probe; assembling a test strip comprising a sample pad, a nitrocellulose membrane, a water absorption pad and a polyvinyl chloride bottom plate; the nitrocellulose membrane is coated with a bovine serum albumin-target conjugate and SA-Bio-cDNA in advance to respectively form a test line T line and a control line C line; the method comprises the following steps: mixing a solution containing fluoroquinolone antibiotics with an Au NRs / PDA-Apt signal probe, incubating, dropping onto a sample pad of a test strip, reacting, detecting and observing the colorimetric intensity and the photo-thermal intensity, and carrying out qualitative or quantitative analysis on the fluoroquinolone antibiotics. According to the Au NRs / PDA mediated dual-mode detection constructed for the first time, the sensitivity and the practicability of antibiotic analysis are remarkably improved, and the flexibility, the reliability and the accuracy of the lateral flow chromatography test paper are remarkably improved.

This invention discloses a method and system for detecting driver visual distraction by integrating perception of the external traffic environment. The method includes the following steps: constructing a driving scene dataset based on a panoramic map for feature extraction; designing a lightweight road semantic segmentation network; and introducing a visual large language model with deep logical reasoning capabilities to construct a visual distraction detection system that integrates perception of the external traffic environment. This invention uses an online visual large language model to analyze the semantic segmentation results of the external traffic environment and outputs in real time the target areas that the driver should focus on in the current scene. By verifying the consistency between this result and the driver's actual gaze point, accurate detection of the driver's visual distraction state is achieved.

The invention relates to a subminiature high-bandwidth single-ended fiber-outgoing film lithium niobate electro-optical modulator, which comprises an SSMP radio frequency assembly, a high-frequency transition film substrate, an electro-optical modulator chip, a dual-core FA, a terminal matching film substrate, an input optical fiber, an output optical fiber and a packaging shell. According to the thin-film lithium niobate electrooptical modulator chip, the waveguide is of a U-shaped structure, the waveguide input port and the waveguide output port are located on the same side of the chip, the dual-core FA is adopted to align the waveguide input port and the waveguide output port for coupling, the space needed by coupling is reduced, and the coupling convenience is improved; a mode field conversion structure is designed on the double-core FA, and the size of an optical fiber spot size is reduced; an SSC spot size converter structure is designed on a chip waveguide, the size of a waveguide spot size is amplified, the size of an optical fiber spot size and the size of a waveguide spot size are converted into the same size, and efficient and low-loss coupling is achieved. The radio frequency feed-in port, the optical input and output port and the direct current pin are located on the opposite side of the device, so that the user terminal array type dense arrangement is facilitated, and the small-sized integration of the system is realized.

The application discloses a continuous multi-direction human behavior recognition method and system and a storage medium, and is applied to the field of human behavior recognition in an indoor scene. In view of the problem that a current human behavior recognition method based on a millimeter wave radar can recognize only a single human behavior movement direction, and a recognition mechanism cannot support continuous operation, the application splits continuous human actions by using extracted movement parameters, obtains real-time split single actions, and then realizes continuous multi-direction human behavior recognition based on the millimeter wave radar by using a parameter estimation method and a mixed feature extraction model of a light-weight deep learning model. The application can realize continuous and stable multi-direction human behavior recognition in an indoor scene, and can be conveniently deployed on a small edge computing platform by using a light-weight recognition method, so that a human behavior recognition system constructed by the application has the advantages of convenient deployment, simple system structure, and easy productization, and can effectively control system cost.

The utility model provides a rural domestic sewage treatment system which comprises an adjusting tank, a bacteria culture tank, a bacteria-algae symbiotic tank, a benthic symbiotic tank and a water storage tank which are sequentially connected, an adjusting pump for adjusting the liquid level of the bacterium culture pond is arranged between the adjusting pond and the bacterium culture pond, a water valve is arranged on a pipeline for connecting the bacterium culture pond and the bacterium-algae symbiotic pond, suspended filler is arranged in the bacterium culture pond, the top of the bacterium-algae symbiotic pond is arranged in a light-transmitting mode, and the benthic symbiotic pond adopts an algae-grass-fish mode. Macromolecular substances in rural domestic sewage are decomposed into micromolecular substances, organic nitrogen and phosphorus are decomposed into inorganic nitrogen and phosphorus which can be absorbed by carp algae, the inorganic nitrogen and phosphorus can be supplied to growth and reproduction of anaerobic bacteria and can also be supplied to growth and reproduction of bacteria and algae in a subsequent bacteria-algae symbiotic pool, the construction cost is reduced, the economic value is improved, and the environment is beautified.

The application relates to a one-dimensional isometric adiabatic friction pipe flow parameter solving method. The method comprises the following steps: collecting the total pressure, total temperature, outlet back pressure and gas mass flow of a pipeline inlet by using a non-intrusive system; iteratively solving the inlet Mach number by using a compressible flow control equation, and calculating the inlet static pressure and static temperature; comparing the outlet critical pressure with the back pressure to judge the flow jam state and determine the outlet boundary condition; combining the boundary condition to solve the outlet Mach number, total pressure and static temperature; solving the total average friction coefficient according to the limit pipe length theorem; setting a step length, taking the inlet Mach number as an initial condition, iteratively solving the along-path Reynolds number by using a fourth-order Runge-Kutta method and correcting the along-path Reynolds number to obtain the along-path distribution of the Mach number; iteratively solving the along-path distribution of the static pressure and velocity based on a gas dynamics method; comparing the predicted value with the measured value of the outlet static pressure, and outputting the along-path curve and the average friction coefficient. The method can obtain the along-path distribution of the one-dimensional isometric adiabatic friction pipe flow parameter.