Patents
Literature
Patsnap Eureka AI that helps you search prior art, draft patents, and assess FTO risks, powered by patent and scientific literature data.

2577results about How to "Effective filtering" patented technology

Text mining data query method and system based on cross-modal similarity

The invention provides a cross-modal similarity-based text mining data query method and system, relates to the technical field of data mining, comprises a cross-modal similarity-based text mining data query method and system, and belongs to the technical field of natural language processing. The method comprises the following steps: constructing a multi-level feature extraction network to respectively obtain text, image and time sequence feature vectors; feature fusion is realized by using an adaptive cross-modal contrast learning network, and a cross-modal similarity matrix is constructed; performing hierarchical screening on the target text by adopting a multi-level screening model to obtain a candidate text set; a text relation reasoning module is constructed based on a graph neural network, and text clustering and relation prediction are completed through a multi-task learning framework. According to the method, the accuracy of cross-modal text data query is improved, the adaptivity of feature fusion is enhanced, and efficient text relationship mining is realized.
Owner:JIANGXI SHUYI TECH CO LTD

Electricity stealing prevention method and system for intelligent electric meter box

The invention relates to the field of electric meter box monitoring, and discloses an anti-electricity-stealing method and system for an intelligent electric meter box, which is used for solving the problem of false alarm caused by harmonic interference when anti-electricity-stealing early warning is carried out, and comprises the following steps: collecting time domain signals of current and voltage in the intelligent electric meter box in a historical detection time period; converting the time domain signal into a frequency domain signal, extracting harmonic information of each harmonic, and establishing a normal harmonic feature model to obtain a target harmonic range; the method comprises the following steps: acquiring time domain signals of current and voltage in an intelligent electric meter box in real time, extracting harmonic information, performing filtering processing according to a target harmonic range, performing first abnormality judgment by using an abnormality detection algorithm, performing multi-parameter joint evaluation if abnormality is judged for the first time, and performing second abnormality judgment according to a multi-parameter joint evaluation result. And electricity larceny early warning is carried out according to the second abnormal judgment result, so that false alarm and missing alarm are effectively reduced, and the adaptability of the system to a complex electricity larceny mode is remarkably enhanced.
Owner:ZHEJIANG YUNTAI ELECTRIC POWER TECHNOLOGY CO LTD

Hydrogen valve comprehensive test system based on vacuum leak detection method

The invention relates to the technical field of leakage detection, and further relates to a hydrogen valve comprehensive test system based on a vacuum leakage detection method. The system comprises a modeling part, an analysis processing part, a real-time control part and a prediction control part. The modeling part is used for establishing a gas state equation in a vacuum chamber and a thermodynamic equilibrium equation of the vacuum chamber according to initial parameters of a test environment; the analysis processing part is used for calculating leakage flux by considering the influence of temperature change on the vacuum degree; the real-time control part is used for adjusting the hydrogen flow; and the predictive control part is used for performing Kalman filtering on the system temperature, the pressure and the hydrogen flow so as to ensure the tracking of the set hydrogen flow. The invention provides an efficient, stable and reliable technical solution for performance evaluation and flow regulation of the hydrogen valve.
Owner:HANFU IND EQUIP (JINAN) CO LTD

Rotary machine fault diagnosis method based on residual contraction convolution and attention mechanism

The invention discloses a rotary machine fault diagnosis method based on residual contraction convolution and an attention mechanism, and belongs to the field of rotary machine fault diagnosis. According to the method, firstly, a soft threshold filtering activation function is designed based on Meta-ACON, a soft threshold filtering module is developed around the activation function, and the module can adaptively learn a filtering threshold and a forward dynamic weighting coefficient, so that irrelevant features are effectively filtered, and the expression ability of effective features is enhanced. And then, in combination with multi-channel and multi-scale residual convolution, a multi-channel residual shrinkage convolution unit is constructed to realize multi-scale and multi-channel effective feature extraction. Besides, according to the method, a high-order statistical modeling idea is utilized, a GSoP attention mechanism is introduced behind a multi-channel residual shrinkage convolution unit, and the ability of extracting discriminative features is enhanced through second-order statistical information of a high-level channel feature map. Test results show that the method has good fault identification capability and generalization capability.
Owner:LUOYANG INST OF SCI & TECH

Gene detection data cleaning method and system based on artificial intelligence

The invention relates to the technical field of genomics, in particular to a gene detection data cleaning method and system based on artificial intelligence. The method comprises the following steps: acquiring gene detection data, and performing preliminary filtering on the gene detection data based on removal of low-quality reads to obtain preliminary filtering data; performing linker sequence identification and removal on the preliminarily filtered data, and performing sequencing error correction to generate gene sequence data; performing GC content analysis and repetitive sequence identification on the gene sequence data to obtain sequence feature data; performing dynamic quality threshold adjustment according to the sequence feature data, and performing secondary filtering on the gene sequence data to generate high-quality sequence data; and performing sequence pattern recognition on the high-quality sequence data to obtain sequence pattern feature data. According to the method, deep analysis is carried out by utilizing sequence characteristics such as GC content and nucleotide distribution, and the recognition capability on sequence complexity and variation modes is improved.
Owner:SICHUAN BAIYUE BIOTECHNOLOGY CO LTD

Building facade structure line accurate extraction method based on laser point cloud feature depth

The invention discloses a building facade structure line accurate extraction method based on laser point cloud feature depth, and is applied to the technical field of point cloud feature extraction. The method comprises the following steps: acquiring building point cloud data, preprocessing the acquired building point cloud data, and dividing the preprocessed building point cloud data into a training set and a test set; inputting the training set into a PCPNET network for training to obtain a PCPNET network training model, and extracting normal vector features of the test set through the PCPNET network training model; adopting a Gaussian weighting function to realize quadratic optimization of the normal vector features, and identifying edge feature points according to the difference of the normal vector features after quadratic optimization; and optimizing the edge feature points through an improved filtering algorithm, calculating normal vectors of the optimized edge feature points, and connecting the normal vectors into feature lines based on preset angle and distance thresholds. According to the method, the extraction precision and robustness of the normal vector features are remarkably improved, and a reliable data basis is provided for building facade structure analysis.
Owner:SUZHOU UNIV OF SCI & TECH

Proposition method and system combining large model and inference engine, medium and product

The invention discloses a proposition method and system combining a large model and an inference engine, a medium and a product, and relates to the field of artificial intelligence. Matching knowledge point rules from a preset reasoning rule base; combining the knowledge point rules to obtain a knowledge point combination sample; reasoning the knowledge point combination topics through a reasoning engine to generate a formal language expression topic set; based on a preset translation template, converting the formal language expression question set into a natural language expression question set; rewriting the natural language expression question set through the large model to generate an optimized question set; judging whether each optimization question in the optimization question set meets a preset output condition or not; and outputting the optimization questions meeting a preset output condition in the optimization question set. By implementing the method, the logic accuracy of question generation is improved.
Owner:可之(宁波)人工智能科技有限公司

Unmanned aerial vehicle self-noise filtering method and system based on multi-channel time-frequency spatial filtering

The invention relates to the technical field of signal processing, in particular to an unmanned aerial vehicle self-noise filtering method and system based on multichannel time-frequency spatial filtering. The method comprises the following steps: acquiring a multi-channel noise-containing sound signal by using a microphone array carried by an unmanned aerial vehicle, and performing short-time Fourier transform to obtain a frequency domain signal; estimating a ratio mask by using a deep neural network, and performing logarithm normalization processing on the frequency domain signal; remodeling the sound signal to match the FFT length of the time-frequency spatial filter, and constructing the time-frequency spatial filter in combination with the multi-channel ratio mask; removing the self-noise of the unmanned aerial vehicle by using the filter and obtaining a target sound signal; and finally, converting the frequency domain signal into a time domain signal through inverse short-time Fourier transform. Through innovative fusion of multi-channel acquisition, deep learning and time-frequency space filtering, the self-noise of the unmanned aerial vehicle in a complex environment is accurately filtered, and the accuracy and efficiency of transformer substation noise monitoring are remarkably improved. Reliable acoustic data are provided for state monitoring of the power equipment, and the method has important engineering application value.
Owner:ELECTRIC POWER RES INST OF GUANGXI POWER GRID CO LTD

Robust image matching system and method based on multi-scale attention feature fusion

The invention relates to the technical field of computer vision and image processing, in particular to a robust image matching system and method based on multi-scale attention feature fusion, and the system comprises a feature extraction module which extracts a multi-resolution feature map containing image features from a to-be-matched image; the position coding module performs spatial position coding on the image features to improve the spatial perception capability of the model; the multi-scale attention module performs multi-scale enhancement processing on the image features, so that the model is more effectively focused on the significant part of the image; the attention module establishes a dependency relationship between image features; the coarse-level matching layer module generates a coarse-level matching result and screens out a matching pair with relatively high confidence coefficient; and the coarse-to-fine module carries out fine processing on the coarse matching result, so that more accurate feature matching is realized. According to the method, the effectiveness and reliability of the method are verified by testing the image matching performance of the to-be-tested picture under the conditions that the to-be-tested picture contains different illumination changes, extreme visual angle changes, repeated textures and the like.
Owner:CHANGZHOU UNIV

Transformer partial discharge monitoring method based on ultrahigh frequency technology

The invention provides a transformer partial discharge monitoring method based on an ultrahigh frequency technology, and relates to the technical field of partial discharge monitoring, and the method comprises the steps: presetting a monitoring point array, laying an ultrahigh frequency sensor group for dynamic discharge monitoring, and obtaining an ultrahigh frequency signal time sequence array; a multi-dimensional environment signal time sequence array is obtained, filtering compensation is carried out, and a partial discharge signal time sequence array is obtained; carrying out distributed signal analysis to obtain a partial discharge characteristic parameter array; constructing a partial discharge feature tree, and performing correlation matching to obtain a partial discharge reason feature parameter array; calling a partial discharge operation and maintenance strategy library; and traversing and matching to obtain a target partial discharge operation and maintenance strategy set, and carrying out partial discharge early warning operation and maintenance processing. According to the invention, the technical problem of low accuracy and efficiency of partial discharge processing caused by the fact that transformer partial discharge monitoring mostly depends on empirical judgment and manual analysis and lacks systematic diagnosis and automatic processing decision support in the prior art is solved.
Owner:JIANGSU HAOHAN INFORMATION TECH

Self-adaptive multi-hop retrieval knowledge graph construction method for large language model questions and answers

The invention discloses a self-adaptive multi-hop retrieval knowledge graph construction method for large language model questions and answers. The method comprises the following steps: extracting triples of an input question and a corpus; self-adaptive multi-hop retrieval matching is carried out, and a specific triple related to the question in the corpus is retrieved and recognized; constructing a knowledge graph by using the question in the adaptive multi-hop retrieval matching process and a corpus triple; the answer predictor model comprises node embedding, message passing and selective filtering, and the node embedding is used for encoding questions and a knowledge graph; carrying out message transmission based on the coded problem and the knowledge graph based on a graph convolutional network; selective filtering is used for filtering redundant information of the knowledge graph; according to the method, the problem that the large language model generates illusion can be relieved, the complex reasoning process can be understood, and therefore the question answering and logical reasoning performance of the large language model is improved.
Owner:CHANGAN UNIV

Method for evaluating damage grade of reinforced concrete building structure

ActiveCN119046814AHigh precisionEfficiently handle nonlinear relationshipsData processing applicationsEnsemble learningModal testingEvaluation result
The invention belongs to the field of building damage detection, and provides a reinforced concrete building structure damage grade evaluation method, and the method comprises the steps: carrying out the modal test of a to-be-evaluated reinforced concrete building structure, obtaining the modal parameters of the to-be-evaluated reinforced concrete building structure, and enabling the modal parameters to at least comprise the inherent frequency, the damping ratio and the vibration mode; a dynamic load is applied to a reinforced concrete building structure to be evaluated, the energy response of the reinforced concrete building structure is measured, the energy response at least comprises input energy and dissipated energy, and the energy loss rate is calculated; calculating variation of modal parameters before and after damage, wherein the variation at least comprises inherent frequency variation, damping ratio variation and modal confidence factors; and inputting the variable quantity of the modal parameter and the energy loss rate as input features into a random forest classifier to generate an evaluation result of the structural damage level. According to the scheme, the accuracy of damage grade evaluation can be improved.
Owner:CHENGDU THIRD ARCHITECTURAL ENG CO

Method and device for detecting and identifying mixed text image

The invention provides a detection and identification method and device for a mixed text image, and the method comprises the steps: collecting an original image set, constructing a data set, and carrying out the preprocessing of the data set; based on a convolutional network classifier, a CRAFT detection model and a TROCR recognition model, performing model training by using the data set to generate an optical character recognition model; acquiring a test set in the data set, and performing text detection and type recognition on the test set by using the optical character recognition model; performing picture lifting processing on the test set; based on the type recognition result, performing text recognition on the processed test set by using an optical character recognition model to generate a recognition result; fusing the identification results to generate a fused identification result; using a multiple screening method to optimize the fused recognition result, and generating a unique text recognition result; and comparing the unique text recognition result with the corresponding text picture content, and calculating and counting the recognition accuracy of the optical character recognition model.
Owner:SHANDONG LUNENG SOFTWARE TECH

Method for calculating bending resistance of reinforcing steel bar lap joint of underground diaphragm wall based on big data

The invention discloses an underground diaphragm wall steel bar lap joint bending resistance calculation method based on big data, and belongs to the technical field of underground engineering design. In order to solve the problems that the condition change of a construction site is difficult to respond in real time, the construction parameters are difficult to dynamically adjust and the complex construction environment is difficult to adapt, the data set quality and the bending resistance prediction precision are effectively improved through multi-source data acquisition and key feature screening, and a physical prediction model of a steel bar lap joint is established. According to the method, factors such as stress transmission and binding power are comprehensively considered, so that the model has good adaptability under different working conditions, model parameters are dynamically adjusted, the model responds to changes of environments and construction errors in real time, the weight is automatically adjusted when monitoring data fluctuates, accurate prediction of the bending resistance is achieved, potential failure risks are effectively prevented, and the construction quality is improved. And a failure mode pre-judgment and multi-level risk management mechanism helps a construction team to quickly identify a high-risk area and flexibly adjust a construction scheme.
Owner:SHENGZHOU WANGXIN JINSHUI CONSTR INVESTMENT CO LTD

Visual inspection method and system for appearance defects of IC package carrier

The invention relates to the technical field of image processing, in particular to a visual inspection method and system for appearance defects of an IC package carrier, and the method comprises the steps: according to the regional concentration ratio of each super-pixel block in each convex hull, the color brightness difference value and the area of all connected domains in each super-pixel block in each convex hull, carrying out visual inspection on the appearance defects of each super-pixel block in each convex hull; and determining a region feature value of each superpixel block in each convex hull, determining a filtering window of each pixel point in the package carrier gray-scale map, and detecting the filtered package carrier. According to the method and the device, the window size of each pixel point on the IC package substrate in a filtering algorithm is determined by analyzing the normal texture region and the oxidation defect region in the gold surface region on the IC package substrate and the characteristics of the noise appearing in the acquired image, so that the accuracy of detecting the oxidation defect in the gold surface region on the IC package substrate is improved.
Owner:SUPER TECH IND (GUANGDONG) CO LTD +1

Impeller assembly, pre-filter and water system

The invention discloses an impeller assembly, a pre-filter applying the impeller assembly and a water consuming system, and relates to the technical field of pre-filters, the impeller assembly comprises an impeller body and a mounting seat, the impeller body is provided with a connecting part, and the connecting part is used for being in transmission connection with an outer framework of the pre-filter; the mounting seat is provided with an impeller cavity, the impeller body is rotatably mounted in the impeller cavity, and one axial side of the mounting seat is provided with a first water inlet communicated with the impeller cavity and the outside of the mounting seat; a second water inlet communicated with the impeller cavity and the outside of the mounting seat is formed in the peripheral side of the mounting seat; a water outlet communicated with the impeller cavity and the outside of the mounting seat is formed in the other axial side of the mounting seat; and the water outlet is used for being communicated with a drain outlet of the pre-filter. According to the technical scheme provided by the invention, the flushing efficiency and the drainage efficiency of the pre-filter can be improved; the water flow drives the outer framework to rotate, so that the self-cleaning capability and the filtering effect of the equipment are further improved.
Owner:FOSHAN SHUNDE MIDEA WATER DISPENSER MFG +1

Preparation method and application of high-water-solubility dihydroquercetin for cosmetics

The invention discloses a preparation method and application of high-water-solubility dihydroquercetin for cosmetics, and belongs to the technical field of compound extraction.The preparation method comprises the steps that larch roots are dried and smashed, soaked in cold water and filtered, filter residues are extracted through plate-frame pressure filtration, an extracting solution is concentrated, alcohol precipitation is conducted, an alcohol solution is concentrated and refrigerated till crystals are separated out, and the high-water-solubility dihydroquercetin is obtained. Adding the crystal into an inclusion solution, filtering after reaction, adding a composite decolorizing agent into the filtrate, filtering again, adjusting the pH value to be neutral, diluting until the solid content of the filtrate is 3 + / -0.5%, passing through a column, concentrating, and spray-drying to obtain the high-water-solubility dihydroquercetin for cosmetics, which is in a powder state and light in color. The composition has a good anti-wrinkle effect, a tightening effect and a soothing effect, and can be used for preparing cosmetics.
Owner:西安绿天生物技术有限公司

Substation safety control method and system based on three-dimensional reconstruction model

The invention provides a substation safety management and control method and system based on a three-dimensional reconstruction model, and relates to the technical field of safety management and control, and the method comprises the steps: obtaining real-time point cloud data inputted by a laser point cloud sensor and a real-time image collected by a camera; comparing the real-time point cloud data with a background point cloud model, and extracting a changed dynamic region by using a background difference algorithm; the dynamic reconstruction module is used for identifying a first position of a first dynamic target in the dynamic region based on the real-time point cloud data; calculating a second position of a second dynamic target contained in the real-time image; if the first position is successfully matched with the second position and the sparse structure of the real-time point cloud data is successfully matched with dense pixels of the real-time image, determining that the first dynamic target and the second dynamic target are both dynamic flight targets; and determining the position of the dynamic flight target according to the first position and the second position, and marking in the three-dimensional model. The safety management and control precision of the transformer substation can be improved.
Owner:WUHAN HUITEST POWER TECH CO LTD

AWG chip automatic test system based on photodiode array

The invention discloses an AWG chip automatic test system based on a photodiode array, and the system comprises a light source transmitting module which is used for transmitting laser and finally coupling the laser to an input waveguide of an AWG chip; the six-axis control platform controls the single-mode optical fiber FA to be accurately coupled to the AWG chip by adjusting the position of the single-mode optical fiber FA; the light source detection module comprises a PD array composed of a plurality of photodiodes, and the PD array obtains information of an output waveguide of the AWG chip; data acquired by the upper computer and the data acquisition system are transmitted to the upper computer to be processed and displayed, and by taking a CWDM chip as an example, insertion loss, central wavelength, bandwidth, crosstalk and the like of the chip can be calculated by respectively acquiring optical power of four channels in the chip, so that whether the chip reaches the standard or not is judged. In the process, as the AWG chip is a passive device, input and output waveguides at the two ends of the chip carry out optical transmission in an alignment mode, and if alignment is wrong, large loss is likely to be generated.
Owner:SHAOXING RES INST OF ZHEJIANG UNIV

Low-cost intelligent fire monitoring system and method based on multi-modal data fusion

The invention discloses a low-cost intelligent fire monitoring system and method based on multi-modal data fusion. The system comprises a data acquisition module, a data preprocessing module, a feature extraction module, a data processing module, a dynamic threshold generation module and a fire alarm module. The data acquisition module is used for acquiring data; the data preprocessing module is used for carrying out standardization, denoising and time synchronization on the acquired data of each sensor; the feature extraction module extracts features corresponding to the flame data; the data processing module obtains a classification result whether there is fire or not; the dynamic threshold generation module generates a fire alarm dynamic threshold; and the fire alarm module is used for comparing the final fire probability value with a fire alarm dynamic threshold value and sending out an alarm signal when the probability value is greater than the fire alarm dynamic threshold value, and manual fire extinguishing needs to be carried out in time. According to the method, effective fusion of multi-sensor data is realized, corresponding threshold values can be adaptively adjusted, and the fire prevention capability is improved.
Owner:XI'AN UNIVERSITY OF ARCHITECTURE AND TECHNOLOGY

Touch response method and device based on vehicle-mounted display screen and electronic equipment

The invention discloses a touch response method and device based on a vehicle-mounted display screen and electronic equipment, and relates to the field of data processing. In the method, a touch signal on the surface of a vehicle-mounted display screen is obtained, and the touch signal comprises an initial touch position, a touch track and touch duration; based on the initial touch position and the touch duration, judging whether the touch signal is a mistaken touch operation or not; if it is determined that the touch signal is not the mistaken touch operation, feature extraction is conducted on the touch track to obtain a track feature vector, and the track feature vector comprises the track length, the track curvature and the track direction; inputting the track feature vector into a preset gesture recognition model, and recognizing a corresponding touch gesture; according to the touch gesture, determining a touch response area in combination with interface display information of the vehicle-mounted display screen; and in the touch response area, generating a touch response event according to a preset interaction rule, and displaying the touch response event on the vehicle-mounted display screen. By implementing the technical scheme provided by the invention, the user experience is improved.
Owner:ALTRON OPTOELECTRONICS (SHENZHEN) CO LTD

Surface treatment equipment for automobile frame

The invention belongs to the technical field of automobile frames, and particularly relates to automobile frame surface treatment equipment which comprises a box body, the top of the box body is fixedly connected with a first supporting frame, the top of the first supporting frame is fixedly connected with a first motor, and the output end of the first motor penetrates through the first supporting frame and is fixedly connected with a threaded lead screw; according to the surface treatment equipment for the automobile frame, the automobile frame is vibrated for multiple times in the cleaning agent through the multiple sets of clamping blocks, so that the cleaning agent powerfully washes the automobile frame, bubbles on the surface of the automobile frame can be eliminated while the automobile frame is vibrated in the cleaning agent, the cleaning agent makes full contact with the automobile frame, and the cleaning effect is improved. And when the automobile frame rotates backwards by 180 degrees and faces upwards, the supporting table continues to be controlled to move upwards, through the arranged push blocks, the automobile frame is vibrated repeatedly when moving upwards, so that a cleaning agent attached to the surface is shaken off, and the step of manual wiping is omitted.
Owner:HEFEI CHANGQING MACHINERY

Control method of waterproof touch type LED driving controller

The invention relates to the field of LED driving controllers, in particular to a control method of a waterproof touch type LED driving controller, which automatically identifies environmental conditions and user preferences through an AI algorithm, dynamically adjusts sensor parameters, realizes self-adaptive configuration, intelligently distinguishes real touch control and false triggering through charge detection, capacitive sensing and weak magnetic field change, and improves the control accuracy of the waterproof touch type LED driving controller. A user operation intention such as light touch, long press and sliding is deeply analyzed in combination with machine learning, after a control strategy is dynamically generated, LED brightness and color temperature are automatically adjusted and a command is executed according to the user intention and environmental parameters, the control strategy is optimized through a reinforcement learning algorithm, meanwhile, a health monitoring function is integrated, a user operation mode and environmental data are analyzed, and the user experience is improved. The health state of the equipment is intelligently judged, potential faults are warned in advance, and stable operation of the equipment and user experience are ensured.
Owner:DONGGUAN CHENGLIANG LNTELLIGENT TECH CO LTD

Circulating cooling immersed liquid cooling server

The invention provides a circulating cooling immersed liquid cooling server, which comprises a main body box and a server body, and is characterized in that the server body is arranged in a mounting through groove in a penetrating manner and is clamped in a mounting frame; multiple sets of liquid outlet pipes and liquid inlet pipes are arranged on the two sides of the main box respectively, multiple sets of copper pipes and a cooling assembly are arranged on the back face of the main box, and the multiple sets of copper pipes are arranged in the cooling assembly in a penetrating mode. One end of the copper pipe is connected with the liquid outlet pipe through a piezoelectric ceramic pump, a filter box is arranged on one side of the main body box, and the other end of the copper pipe is connected with the filter box which is connected with the liquid inlet pipes through a conveying assembly; and a controller and a liquid level sensor are arranged on the main body box. According to the server, through mutual cooperation of the piezoelectric ceramic pump, the copper pipe, the filter box and the conveying assembly, cooling liquid can circularly flow, cooling operation on the cooling liquid is achieved through the cooling assembly, the cooling liquid can continuously take away heat generated by the server body, and the phenomenon that the server body is damaged due to continuous accumulation of the heat is avoided. The temperature of the cooling liquid is too high, so that the cooling capacity is lost.
Owner:SUGON DATAENERGYBEIJING CO LTD

Multi-level retrieval enhancement method for large language model

The invention discloses a multi-level retrieval enhancement method for a large language model, particularly relates to the technical field of multi-level retrieval, and aims to more accurately understand a query intention of a user by performing initial analysis on a query input by the user and generating a core query statement and a primary retrieval document set. Performing clustering analysis on the primary retrieval document set to identify a redundant risk document set, a reserved document set and an information loss risk document set, obtaining redundant detection information and credibility evaluation information of the redundant risk document set, performing primary screening on the redundant risk document set, effectively filtering redundant retrieval documents with relatively low credibility, and improving the reliability of the retrieval documents. A supplementary query statement is generated according to an information loss risk document set, secondary retrieval is performed in a knowledge base, multi-angle requirements and potential fuzzy intentions of a user are more comprehensively covered, a document priority recommendation index is generated through comprehensive calculation of a document quality coefficient and a document freshness coefficient, and the user experience is improved. And the documents with high quality and high timeliness are preferentially displayed to the user, so that the user experience is improved.
Owner:无锡锡商银行股份有限公司

Water supply network pipe burst alarm positioning system based on real-time pressure monitoring data

The invention relates to the technical field of pipe network monitoring, in particular to a water supply pipe network pipe burst alarm positioning system based on real-time pressure monitoring data, and the system comprises a pressure data collection module which collects pressure data from a plurality of pressure monitoring points of a water supply pipe network; the pressure drop threshold calculation module is used for determining an alarm pressure drop threshold of each pressure monitoring point based on historical pressure data; the voltage drop alarm module is used for calculating a current time window voltage drop value and comparing the current time window voltage drop value with a voltage drop threshold value; the association alarm module is used for clustering alarm information of adjacent pressure monitoring points and combining a plurality of alarm points which are close in space into a single alarm event; the alarm positioning module is used for matching a pipe explosion node corresponding to the current alarm information by utilizing a mode recognition algorithm and quickly positioning the pipe explosion node; according to the method, the system calculates the aggregation characteristics, the description capability and the credibility of the alarm event are enhanced, and the identification and positioning efficiency of the system on the pipe explosion event in the complex pipe network environment is improved.
Owner:SHENZHEN WATER GRP CO LTD

Product recommendation method and device, computer equipment and storage medium

The embodiment of the invention belongs to the technical field of artificial intelligence, and relates to a product recommendation method and device, computer equipment and a storage medium, and the method comprises the steps: constructing a user portrait corresponding to a requesting user according to dialogue text data; performing semantic analysis operation on the dialogue text data according to a large language model to obtain a real intention of a user and a product demand index; constructing a dynamic knowledge graph according to the user portrait, the real intention of the user and the product demand index; generating a product comparison set according to the dynamic knowledge graph; and performing a comparison operation on the to-be-selected products in the product comparison set according to an iterative comparison process to obtain a target recommended product. According to the method, through deep user portrait and multi-dimensional similarity calculation, dynamic comparison set generation and progressive accurate comparison, irrelevant information is effectively filtered, the cognitive burden of the user is reduced, and the accuracy of personalized recommendation is effectively improved.
Owner:CHINA PING AN PROPERTY INSURANCE CO LTD

Multi-modal federated learning method with robustness

The invention belongs to the field of artificial intelligence and machine learning, and discloses a multi-modal federated learning framework with robustness, which comprises the following steps: S1, acquiring missing modals, screening out modals with strong correlation with the missing modals, and performing prediction by adopting multi-layer perceptron training to complete modal alignment; s2, labeling the feature data in the modality, and combining the feature data to synthesize a feature vector; s3, fusing the feature data based on an attention mechanism to form a modal vector; s4, performing local training on the aligned and fused modal data; and S5, uploading the local training model to a server, carrying out aggregation operation by adopting a federal learning framework, and distributing the model to a client to carry out next iteration until the model is converged. According to the method, through prediction of missing mode and multi-mode data fusion, the problem that the training model precision is reduced when the mode is missing is effectively solved, the influence of the mode missing on the training precision is reduced, and the reliability of federal learning is improved.
Owner:NANJING UNIV OF POSTS & TELECOMM

Transformer state monitoring method and system based on voiceprint perception, and medium

The invention discloses a transformer state monitoring method and system based on voiceprint perception and a medium, and the method comprises the following steps: collecting a sound wave signal and a vibration signal of a transformer, carrying out the preprocessing of the collected signals, and carrying out the feature extraction; converting the obtained signal into frequency domain information, and detecting and marking an abnormal value in the signal; the method comprises the following steps: constructing a T-Manba model, and training the Mamba model by using known signal data of a transformer in a normal operation state; and inputting the obtained data into the trained Mamba model, and identifying the abnormal state of the transformer. According to the invention, real-time monitoring and diagnosis of the operation state of the transformer are realized through sound analysis, automation and intelligentization of state monitoring of the transformer are realized, and real-time and accurate monitoring and diagnosis of the operation state of the transformer are provided.
Owner:STATE GRID JIANGSU ELECTRIC POWER CO LTD NANJING POWER SUPPLY COMPANY +2

CNC machine tool with anti-jamming servo tailstock

The invention discloses a CNC machine tool with an anti-jamming servo tailstock, and relates to the technical field of machine tool equipment.The CNC machine tool comprises a workbench, a machine tool body is arranged at the upper end of the workbench, a spindle box is arranged at the upper end of the machine tool body, a chuck is arranged on one side of the spindle box, a fixing plate is arranged in the machine tool body, and a first servo motor is arranged on one side of the machine tool body; the output end of the first servo motor penetrates through the lathe bed and is provided with a lead screw, two supporting guide rods are symmetrically arranged between the fixing plate and the inner surface of one side of the lathe bed, a moving block is arranged on the outer surface of the lead screw, a tailstock sliding saddle is arranged at the upper end of the moving block, and a mounting table is arranged at the upper end of the tailstock sliding saddle. According to the CNC machine tool with the anti-jamming servo tailstock, the rear center can be stably driven to transversely move, impurities on the surfaces of the lead screw and the supporting guide rod can be efficiently cleaned and automatically lubricated during use, the jamming condition is prevented, and meanwhile, the center base body and the rear center can be conveniently disassembled, assembled and replaced.
Owner:CHONGQING TIEDINGNIU MASCH TOOL CO LTD