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

50 results about "Probability propagation" patented technology

Boiler state abnormity monitoring system based on artificial intelligence

The invention discloses a boiler state abnormity monitoring system based on artificial intelligence, and particularly relates to the field of abnormity monitoring, comprising a multi-modal sensor data acquisition module, an edge computing node module, a migration enhanced abnormity detection module and a fault tracing module; according to the invention, a boiler multi-mode sensor network is constructed, vibration, temperature, pressure and other data acquisition modules are integrated, adaptive filtering and working condition classification are realized through edge computing nodes, and a data enhancement strategy is dynamically adjusted; the anomaly detection module fuses a physical mechanism model and a twin neural network, generates a multi-dimensional deviation index, and starts fault tracing after a dynamic threshold triggers an alarm; probability propagation is carried out based on a Bayesian knowledge graph, fault trajectories are simulated in combination with an adversarial generative network, root cause probability ranking and maintenance priority scores are generated, real-time support is provided for operation and maintenance decisions, and intelligent diagnosis and dynamic optimization of the boiler state are achieved.
Owner:DONGYING HAIXIN HEATING SUPPLY CO LTD

Visual analysis method and system for security and protection monitoring data

The invention relates to the technical field of data visualization, and discloses a visual analysis method and system for security and protection monitoring data, and the method comprises the steps: carrying out the asynchronous collection of a video monitoring data stream, an infrared sensor data stream, an access control system data stream and an audio collection data stream in a security and protection monitoring system, and obtaining a multi-source asynchronous data stream set; generating a potential causal event pair set based on the multi-source asynchronous data stream set, and creating a time window allocation table according to the potential causal event pair set; performing causal probability calculation on the multi-source asynchronous data stream set to obtain a multi-dimensional causal probability propagation matrix; and asynchronous time delay compensation and causal chain visualization are performed on the multi-dimensional causal probability propagation matrix based on the time window allocation table to obtain a dynamic causal relationship graph, so that visual dynamic display of the causal chain of the asynchronous data stream is realized, and brand new monitoring situation awareness experience is provided for security personnel.
Owner:SHENZHEN BINGYITONG INFORMATION TECHNOLOGY CO LTD

Map updating method, system and device based on visual SLAM system, medium and product

The invention discloses a map updating method, system and device based on a visual SLAM system, a medium and a product, and relates to the technical field of robot vision, and the method comprises the steps: recognizing a dynamic target in a scene image based on a target detection algorithm of a deep neural network, and employing a feature point motion probability propagation method to remove feature points of the dynamic target; feature point matching is carried out on the static feature points and static points in a local map, tracking is carried out on each frame of scene image, and the camera pose and the road sign point depth in the local map are estimated; optimizing road sign points and camera poses in the local map; detecting a loop based on a DBoW2 model, correcting a camera track accumulative error by executing the camera pose image, and determining an optimized camera pose image; the optimized local map and the optimized camera pose map are kept globally consistent, the optimized local map is updated, the updated map structure is determined, and the positioning and mapping tasks are reliably and efficiently completed in an unknown dynamic environment.
Owner:SHANGHAI UNIV

Robust noise reduction processing method and system for sound wave signal self-supervised learning enhancement

The invention provides a sound wave signal self-supervised learning enhanced robust noise reduction processing method and system, and relates to the technical field of signal processing, and the method comprises the steps: carrying out the feature enhancement of an initial time-frequency representation through a dynamic adaptive mask strategy, constructing a self-supervised reconstruction task based on the mask time-frequency representation, separating noise and signal subspaces in a semantic manifold space, and carrying out the self-supervised learning enhanced robust noise reduction. And establishing a probability propagation network in combination with time domain continuity characteristics to model a local dependency relationship, and finally generating a noise reduction weight and realizing semantic fidelity optimization. The method can effectively improve the noise reduction effect and semantic integrity of sound wave signals in a noise complex environment.
Owner:BEIJING GUANYU INFORMATION TECHNOLOGY CO LTD

Automatic fault tracing method, system and equipment based on topological coding and medium

The invention relates to the technical field of data processing, and particularly provides an automatic fault tracing method, system and device based on topological coding and a medium, and the method comprises the steps: distributing a unique topological code reflecting the hierarchical position of physical equipment for the physical equipment in a factory power distribution system; the method comprises the following steps: collecting real-time operation data through an edge collection terminal, and binding topological codes to form a real-time data stream with a label; constructing a topology dependency degree model, and defining the inherent failure rate of each node and the dependency degree of each node on a superior node; when the fault of the target equipment is detected, reversely traversing upstream nodes along a topological path based on the model, and calculating the fault suspicion probability of each upstream node by adopting a probability propagation algorithm; and carrying out normalization processing on the probabilities, fusing real-time state data, carrying out weighted correction, and finally generating a fault diagnosis report sorted according to the probabilities. According to the invention, by introducing a multi-level topological coding system and a probability propagation algorithm, rapid and accurate positioning and intelligent diagnosis of faults are realized.
Owner:INSPUR ARTIFICIAL INTELLIGENCE RES INST CO LTD SHANDONG CHINA

Fault root cause dynamic prediction method and system based on probability propagation reasoning

The invention provides a fault root cause dynamic prediction method and system based on probability propagation reasoning, and relates to the technical field of fault positioning prediction, and the method comprises the steps: constructing a directed graph structure, associating each node with a dynamic fault probability, and constructing a fault state expression system; based on the fault state expression system, calculating and dynamically updating an edge weight, and constructing a dynamic propagation graph model; generating fault occurrence probabilities of the monitoring data, and fusing the fault occurrence probabilities to obtain aligned unified probability data; on the basis of a dynamic propagation graph model, on-graph probability propagation reasoning is carried out by using unified probability data, a final root cause is determined, and rapid fault positioning is realized; an intelligent self-healing and strategy optimization algorithm based on reinforcement learning is introduced, a repair strategy is automatically generated and executed, a decision is continuously optimized through online learning, and a self-evolution safety operation and maintenance system is formed. According to the invention, accurate and rapid prediction and positioning of system faults are realized.
Owner:山东省大数据中心

Open scene target detection method and device

The embodiment of the invention provides an open scene target detection method and device, and the method comprises the steps: inputting an input image into a backbone network of a preset convolutional neural network, extracting the local granularity feature of the input image, carrying out the local sensing of the local granularity feature through deep convolution, extracting the corresponding local space information, and carrying out the recognition of the local space information; capturing a global context dependency relationship among the local space information through a multi-head self-attention mechanism, and performing multi-scale feature fusion on the global context dependency relationship and the local granularity features to obtain corresponding multi-level feature representation; inputting the multi-level feature representation into a set movement probability propagation model for target detection operation, and outputting a final position coordinate and a category label corresponding to an object identified as a dynamic target as a target detection result through the movement probability propagation model; according to the invention, the precision and real-time performance of target detection in an open scene can be effectively improved.
Owner:NO 15 INST OF CHINA ELECTRONICS TECH GRP

House security level prediction method and system based on Bayesian network

The invention discloses a house safety level prediction method and system based on a Bayesian network, and relates to the technical field of house safety evaluation. The method comprises the following steps: acquiring structural characteristic parameters, environmental influence factors and historical state information of a target house; performing three-level quantization processing on the structural characteristic parameters to obtain a structural evaluation matrix, and constructing a dynamic attenuation function based on the environmental impact factors to calculate a dynamic attenuation weight; generating a historical state index based on the historical state information through a multi-dimensional feature fusion technology; based on the structure evaluation matrix, the dynamic attenuation weight and the historical state index, calculating the structure redundancy as a hidden variable; constructing a three-layer Bayesian network; and probability propagation calculation is carried out through a Bayesian network inference engine, and the house safety level of the target house is determined according to the probability distribution of the output layer. By implementing the technical scheme provided by the invention, the actual safety level of the house can be accurately evaluated and predicted.
Owner:SICHUAN WATER CONSERVANCY VOCATIONAL & TECH COLLEGE +1

Recommendation method and device based on multi-objective optimization and computer readable storage medium

This application provides a recommendation method, apparatus, and computer-readable storage medium based on multi-objective optimization. The method includes: acquiring user rating information, including ratings from multiple users for multiple items; clustering users into different clusters using K-means clustering based on the user rating information; predicting the predicted ratings of users for items in each cluster based on a probability propagation algorithm with improved resource allocation; generating an initial population in each cluster based on the predicted ratings of users for items; and solving a multi-objective optimization problem based on an accuracy objective function and a diversity objective function using an evolutionary strategy guided by accuracy-preference users, thereby obtaining the target recommendation result for each cluster. By using the above method to guide the evolutionary direction of the algorithm through user preferences, the target recommendation result can be made more biased towards the accuracy objective function while taking into account diversity.
Owner:CHINA UNIONPAY

Fault root cause dynamic prediction method and system based on probabilistic propagation reasoning

The disclosure provides a fault root cause dynamic prediction method and system based on probability propagation reasoning, relates to the technical field of fault positioning prediction, and comprises the following steps: constructing a directed graph structure, associating a dynamic fault probability with each node, and constructing a fault state expression system; based on the fault state expression system, calculating and dynamically updating the edge weight to construct a dynamic propagation graph model; generating the fault occurrence probability of the monitoring data, and fusing the fault occurrence probability to obtain aligned unified probability data; based on the dynamic propagation graph model, performing graph probability propagation reasoning on the unified probability data to determine the final root cause, thereby realizing rapid fault positioning; introducing an intelligent self-healing and strategy optimization algorithm based on reinforcement learning to automatically generate and execute a repair strategy, and continuously optimizing the decision through online learning to form a self-evolving safe operation and maintenance system. The disclosure realizes accurate and rapid prediction and positioning of system faults.
Owner:山东省大数据中心

Product data multi-dimensional full-period management method, system, equipment and medium

The invention provides a product data multi-dimensional full-period management method, system and device and a medium, and relates to the technical field of product data management, and the method comprises the steps: constructing a preset module; receiving project configuration information input by a user, calling and executing a corresponding clipping rule from a preset module based on the project configuration information, and dynamically generating a project baseline; on the basis of the project base line, constructing an incidence relation graph between the deliverables, and on the basis of the incidence relation graph and the dynamic Bayesian network, establishing a change influence probability propagation model between the deliverables; when it is monitored that the target deliverable is changed, based on a change influence probability propagation model, quantitative influence values of the change source deliverable on the associated deliverable in quality, progress and cost dimensions are calculated, and a change influence analysis report is generated; and outputting visual results of the project base line, the incidence relation graph and the change influence analysis report. Product data management is realized or assisted through an automatic system, and the management efficiency and effect of the product data are improved.
Owner:AVIC CIVIL AIRCRAFT AIRBORNE SYSTEM ENGINEERING CENTER CO LTD

Base station fault root cause positioning method and system based on knowledge graph

The invention belongs to the technical field of communication network operation and maintenance and artificial intelligence crossing, and particularly relates to a base station fault root cause positioning method and system based on a knowledge graph, and the method comprises the steps: constructing a communication network knowledge graph carrying causal direction attributes and propagation weights, and carrying out the adaptive spatial-temporal clustering convergence of alarm data, mapping the alarm event to a knowledge graph to activate a causal propagation sub-graph, calculating a root cause probability score through multi-hop message transmission perceived along a causal relationship edge execution direction by a graph attention network, outputting a node which is ranked in the front as a root cause, and continuously updating a graph weight through an expert feedback closed loop, according to the method, root cause positioning is reconstructed from a feature classification normal form to a probability propagation normal form of a causal relationship graph, the accuracy rate reaches 85%, and the positioning duration is shortened to 3 min.
Owner:GUANGZHOU GUANGJIE NETWORK TECH CO LTD

A complex network-based system simulation system credibility prediction method

The application relates to a complex network-based system simulation system credibility prediction method, relates to the technical field of simulation, and aims to solve the technical problems that existing simulation credibility evaluation methods cannot effectively describe the complex coupling and interaction relationship between subsystems in a system simulation system, depend on system-level reference data, and have poor applicability. The technical means are as follows: S1: abstracting simulation entities of the system simulation system into network nodes with initial credibility, abstracting information interaction relationships between entities into directed edges, defining node weights and edge weights, and constructing a weighted directed network; S2: based on the network, using a credibility probability propagation algorithm composed of a neighbor node propagation function, a credibility reservation function and an iterative solving algorithm to iteratively calculate and obtain modified credibility of each node until convergence; and S3: according to the modified credibility of each node and the corresponding node weight, weightedly synthesizing the system-level credibility prediction result of the system simulation system. The application has a good application prospect in the technical field of simulation credibility evaluation.
Owner:HARBIN INST OF TECH

Space-time joint planning and decision-making method and system based on uncertainty perception

This application provides a spatiotemporal joint planning and decision-making method and system based on uncertainty perception. The method generates a spatiotemporal probability occupancy heatmap of the moving targets by acquiring detection results from moving targets within a preset range of the autonomous vehicle, and static map data within the same range. The static map data and the heatmap are then fused, and the occupancy probability distribution at the current moment is updated based on a probability propagation mechanism. A three-dimensional probability occupancy grid is constructed, and the values ​​in the grid and the collision time are mapped to a risk field. A chance-risk analysis of preset candidate strategies is performed to determine the output decision, the planning space, and the optimal perception trajectory within that space. This application avoids blindly proceeding under high uncertainty conditions, effectively transmitting perceived uncertainty at the planning layer, and significantly improving the robustness and safety of the autonomous driving system in long-tail scenarios.
Owner:HONEYCOMB (WUHAN) MICROSYSTEM TECH CO LTD

Unified probability propagation prediction method and system suitable for wui fire

The application discloses a unified probability propagation prediction method and system suitable for WUI fire, and the method comprises the following steps: constructing a unified state space and discretizing; constructing an external driving model of building layer fire propagation; constructing a building layer dynamic Bayesian network with each building as a node, parameterizing a conditional probability table of the building layer dynamic Bayesian network based on a physical driving mechanism; establishing a mapping relationship between an internal state of the building layer and an external observation category; realizing state correction and rolling prediction based on Bayesian updating; when external observation is obtained, the system state is updated based on Bayesian updating in combination with observation information, a posterior state distribution is obtained, the posterior state distribution is taken as an initial state of the next time step prediction, a rolling assimilation closed loop is formed, and finally a multi-state probability risk field of building scale evolution over time is output. The application can realize unified coupling of a landscape, a building and a flying fire, obtain state probability output over time, and can be assimilated and corrected.
Owner:ZHEJIANG UNIV

Defect hotspot detection method based on two-dimensional profile extraction

The application provides a defect hotspot detection method based on two-dimensional contour extraction, relates to the field of semiconductor processing technology, and combines a dynamic gradient tracking method (DGT) with a probability propagation network (PPN). The former extracts two-dimensional contours in a layout accurately through adaptive step length and turning point identification, and overcomes the limitation of poor adaptability of traditional static edge detection to complex patterns. The latter propagates probability through an adjacency graph, fuses local contour features and global layout information, and breaks through the limitation of traditional isolated analysis methods. The fusion can not only predict potential defect hotspots in the design stage, but also dynamically adapt to different process nodes and layout densities, significantly improves detection accuracy and robustness, provides unprecedented early optimization capability for photolithography research and development, and thus reduces the defect occurrence rate in the manufacturing process and accelerates the product launch time.
Owner:上海芯无双仿真科技有限公司

A hybrid uncertainty propagation method, system, device, and medium for complex equipment response

The application discloses a hybrid uncertainty propagation method, system, equipment and medium for complex equipment response, and the method comprises the following steps: finding samples determining the CDF boundary in an iterative manner through a sequence simulation strategy, so as to gradually approach the upper and lower boundaries of the structural response P-box. In the above sequence solving process, the CDF solving of each step is used to convert the high-dimensional probability propagation problem into the moment analysis problem of a single-variable sub-system through a single-variable dimension reduction integral method. After the interval of the statistical moment of the system response is obtained, the statistical moment of the response is fitted in the interval range to obtain the CDF through Johnson fitting distribution. The sequence simulation strategy of the application is more effective in the probability box analysis of complex equipment response than the direct simulation method, and more accurate system response is obtained only by using less sample point information under the condition of ensuring the same calculation accuracy. The application combines the advantages of interval sequence simulation and statistical moment analysis, and solves the sampling efficiency problem in a high-dimensional problem.
Owner:NANJING INST OF TECH

A social network friend prediction method integrating local information and community information

The present invention discloses a social network friend prediction method that integrates community information and local information, including the following steps: preprocessing the social network using a community detection algorithm to obtain the community structure of the network; weighting the edges between users using community information to more accurately reflect the relationship between users; k-layer probability propagation; calculating the user similarity matrix; and user node similarity sorting. The method predicts the similarity between users using a probability propagation method, and further improves the accuracy of the prediction through the community information between users. The present invention considers more local information by designing k-layer probability propagation, and considers the problem that different neighbors of users have different influences on them by using user community information, so that the similarity between users can be better predicted, and the prediction effect can be effectively improved.
Owner:ANHUI UNIV

Dynamic environment monocular multi-object slam method based on instance segmentation and three-dimensional reconstruction

The application relates to the technical field of automatic driving and computer vision, in particular to a dynamic environment monocular multi-object SLAM method based on instance segmentation and three-dimensional reconstruction, which comprises the following steps: acquiring an image frame sequence collected in the driving process of an automatic driving vehicle; performing feature extraction on each frame image; using an instance segmentation network to obtain dynamic objects and static objects in the image on the key frame image; marking and removing the influence of dynamic feature points; obtaining the current frame dynamic feature points by using Bayesian probability propagation, and removing the dynamic feature points; obtaining the mask, the boundary box and the sparse 3D point cloud of the extracted multiple static objects, and obtaining the initial pose of the multiple static objects by using PCA; after multi-frame observation and sufficient data, the appearance implicit code and the optimized pose of the multiple static objects are obtained by using a three-dimensional reconstruction network, and three-dimensional reconstruction is performed on the multiple static objects. The application can reduce the absolute trajectory error of a traditional SLAM method in a dynamic environment, improve the system pose estimation accuracy and robustness, and improve the semantic degree of a map.
Owner:TAIYUAN UNIVERSITY OF TECHNOLOGY

Complex equipment response-oriented hybrid uncertainty propagation method, system, equipment and medium

The invention discloses a complex equipment response-oriented hybrid uncertainty propagation method, system and device and a medium, and the method comprises the steps: finding a sample which determines a CDF boundary in an iteration mode through a sequence simulation strategy, so as to gradually approach the upper and lower boundaries of a structural response P-box; in the sequence solving process, in each step of CDF solving, a high-dimensional probability propagation problem is converted into a moment analysis problem of a univariate subsystem through a univariate dimension reduction integral method. And after the interval of the system response statistical moment is obtained, fitting the response statistical moment in the interval range through Johnson fitting distribution to obtain the CDF. Compared with a direct simulation method, the sequence simulation strategy disclosed by the invention has the advantages that the response probability box analysis of the complex equipment is more effectively carried out, and under the condition of ensuring the same calculation precision, the relatively accurate system response is obtained by only using less sample point information. According to the method, the sampling efficiency problem in a high-dimensional problem is solved by combining the advantages based on interval sequence simulation and statistical moment analysis.
Owner:NANJING INST OF TECH

Robust denoising processing method and system enhanced by acoustic signal self-supervised learning

The application provides a robust denoising processing method and system enhanced by acoustic wave signal self-supervised learning, relates to the technical field of signal processing, and comprises the following steps: performing feature enhancement on an initial time-frequency representation through a dynamic self-adaptive mask strategy; constructing a self-supervised reconstruction task based on the mask time-frequency representation; separating noise and signal subspaces in a semantic manifold space; combining time-domain continuity features to establish a probability propagation network to model local dependent relationships; and finally generating denoising weights and realizing semantic fidelity optimization. The application can effectively improve the denoising effect and semantic integrity of acoustic wave signals in a noise complex environment.
Owner:BEIJING GUANYU INFORMATION TECHNOLOGY CO LTD

A method for selecting virtual scene object groups based on probability propagation and position rearrangement

The embodiments of the present disclosure disclose a method for selecting a group of virtual scene objects based on probability propagation and position rearrangement. A specific implementation of the method includes: constructing a spatial grid to accelerate the probability propagation process, querying the minimum side length of the collision body among all objects, and setting it as the side length of the smallest unit of the spatial grid; presetting the grouping center position to ensure that the position between each other is the maximum value; recording a grouping probability vector for each object; detecting the user's selection operation for ungrouped objects; based on the probability propagation algorithm of two strategies, updating the grouping probability vector and determining the grouping priority of the remaining objects to be selected; based on the object position rearrangement algorithm of the grouping probability vector, relocating the position of the target object. This implementation can increase the speed at which users classify randomly arranged multiple objects in a virtual reality scene, reduce the number of operations and difficulty, and improve the user experience.
Owner:BEIHANG UNIV

A digital marketing probability trigger-based split marketing method and system

The application discloses a kind of digital marketing probability trigger type split marketing method and system, it is related to probability marketing propagation method field.The application constructs dynamic social fission graph structure structure, extracts user behavior state vector and state transition tensor, and introduces nonlinear fission function and response function and other model mechanisms, realizes the accurate determination of user fission potential and the high-dimensional modeling of transition path, so as to be able to identify seed node and key intersection node with high fission potential, construct reinforced propagation path matrix, realize marketing resource in high-value node Concentration and preferential allocation.Improves the propagation efficiency, conversion accuracy and the intelligent level of marketing strategy as a whole, significantly makes up the defects of existing probability propagation model in dynamic and sensitivity.
Owner:CHENGDU NABAO TECH CO LTD

Space-time joint planning and decision-making method and system based on uncertainty perception

ActiveCN122126313BQuantitative uncertaintyAvoid passing blindlyProbability propagationHeat map
The application provides a spatio-temporal joint planning and decision-making method and system based on uncertainty perception, which obtains a detection result of detecting a moving target in a preset range of an autonomous vehicle, and static map data in the preset range of the autonomous vehicle, generates a spatio-temporal probability occupation heat map of the moving target, fuses the static map data and the spatio-temporal probability occupation heat map, updates an occupation probability distribution at a current time based on a probability propagation mechanism, constructs a three-dimensional probability occupation grid, maps a value in the three-dimensional probability occupation grid and a collision time to a risk field, performs opportunity-risk analysis on a preset candidate strategy, determines an output decision, determines a planning space, and determines an optimal perception trajectory in the planning space; the application can avoid blind passing in a high-uncertainty condition, effectively transfer perception uncertainty to a planning layer, and significantly improve the robustness and safety of an autonomous driving system in a long-tail scenario.
Owner:HONEYCOMB (WUHAN) MICROSYSTEM TECH CO LTD

OODA heterogeneous network link prediction method, device, equipment and medium

The invention discloses an OODA heterogeneous network link prediction method, device and equipment and a medium, and relates to the technical field of OODA heterogeneous network link pre-processing, and the method comprises the steps: employing a CROTS algorithm based on an OODA heterogeneous network, and constructing a teacher-student model; and mapping nodes in the OODA heterogeneous network to a low-dimensional vector space, predicting the OODA heterogeneous network through a teacher model by adopting a probability propagation algorithm to obtain a prediction result, and training a student model based on the prediction result to perform link prediction. According to the method, the learning thought of teachers and students is introduced into OODA heterogeneous network link prediction, so that the link prediction speed is greatly improved. The teacher model effectively avoids the problem that the prediction time of a high-accuracy link prediction algorithm is long, and meanwhile, the high accuracy is kept; after the student model completes the teacher-student learning process, OODA heterogeneous network link prediction is carried out according to the topological characteristics, so that high accuracy can be kept while the prediction speed is high.
Owner:NAT UNIV OF DEFENSE TECH

Quantitative evaluation method for coal mine rock burst risk

The invention discloses a coal mine rock burst risk quantitative evaluation method, and the method achieves the unified organization and expression of monitoring data, geological conditions and mining parameters through constructing a three-dimensional grid data model combining space coordinates and time information, provides a structured data basis for risk evaluation, and improves the risk evaluation efficiency. According to the method, multi-dimensional feature vector extraction and causal association graph modeling are combined, causal paths and strength among monitoring parameters, geological factors and mining conditions are described, it is ensured that an evaluation result has a traceable logic basis, historical time sequence data are utilized to train a dynamic causal evaluation model, conditional probability propagation is executed in a real-time monitoring stage, and a dynamic causal evaluation result is obtained. The danger probability value of each grid unit is dynamically output, the space migration and time evolution rule of danger along with mining disturbance can be reflected, and the defects that the result is static and updating is lagged in a traditional method are overcome.
Owner:CHINA UNIV OF MINING & TECH

Gas meter digital measurement uncertainty evaluation method based on deep learning

PendingCN121959250AAccurately characterize nonlinear characteristicsbreak through dependenceEnsemble learningBiological modelsData setProbability propagation
The invention discloses a gas meter digital measurement uncertainty evaluation method based on deep learning, and relates to the field of metering, and the method comprises the steps: collecting verification data of a gas meter, carrying out the preprocessing of the verification data to obtain a known data set, and dividing the known data set into a training set and a verification set; training the deep learning model by using the training set and verifying by using the verification set; calculating the mean value and the standard uncertainty of the input quantity based on the known data set; in combination with the Monte Carlo method and the prediction uncertainty probability distribution of the model instance, calculating the probability distribution of the measurement through random sampling and probability propagation; based on the measured probability distribution, the measured mean value and the standard uncertainty are calculated, and evaluation of the digital measurement uncertainty of the gas meter is completed. The evaluation process is full-automatic, and subjective errors caused by manual introduction are avoided; and the requirements of measurement development planning are met, digital measurement construction is promoted, and verification credibility is improved.
Owner:GUANGZHOU INST OF MEASURING & TESTING TECH

A multi-modal knowledge graph sparse information enhancement system

The application relates to a multimodal knowledge graph sparse information enhancement system and belongs to the technical field of data processing.The system comprises a rule mining module, which mines a rule set from a data-scarce knowledge graph; a feature extraction module, which synchronously extracts entity structure, visual and text features on the basis of a learnable structure vector of the graph, and outputs three-modal embedding of a unified dimension after dimension reduction and normalization; a scoring module, which calculates scores of each mode by using a scoring function corresponding to the mode, and obtains a probability of a relationship triple by fusing the scores through learnable weights; a rule set is a skeleton, and the probability is a priori; a posterior distribution of unknown facts is output by global probability propagation in combination of known facts and learnable rule weights; and a variational inference module is used for iteratively optimizing the posterior distribution through a variational EM algorithm, outputting high-confidence unknown facts, and completing knowledge graph completion.The application realizes the generalization ability and robustness of multimodal knowledge graph reasoning under a low-resource setting.
Owner:JIAXING UNIV

Virtual scene object group selection method based on probability propagation and position rearrangement

The embodiment of the invention discloses a virtual scene object group selection method based on probability propagation and position rearrangement. A specific embodiment of the method comprises the following steps: constructing a space grid to accelerate a probability propagation process, querying a minimum side length value of a collision body in all objects, and setting the minimum side length value as the side length of a minimum unit of the space grid; presetting a grouping center position so as to ensure that the positions of the grouping center positions are maximum values; recording a grouping probability vector for each object; detecting a selection operation of a user for ungrouped objects; based on a probability propagation algorithm of two strategies, a grouping probability vector is updated, and grouping priorities of the remaining to-be-selected objects are determined; and relocating the position of the target object based on an object position rearrangement algorithm of a grouping probability vector. According to the embodiment, the speed of classifying the randomly arranged multiple objects in the virtual reality scene by the user can be increased, the operation frequency and difficulty are reduced, and the user experience is improved.
Owner:BEIHANG UNIV

Defect hotspot detection method based on two-dimensional contour extraction

The invention provides a defect hot spot detection method based on two-dimensional contour extraction, which relates to the technical field of semiconductor processing, and combines a dynamic gradient tracking method (DGT) and a probability propagation network (PPN). The limitation that traditional static edge detection is insufficient in adaptability to complex patterns is overcome; the local contour features and the global layout information are fused through the adjacency graph propagation probability, the limitation of a traditional isolated analysis method is broken through, the fusion can not only predict potential defect hotspots in the design stage, but also dynamically adapt to different process nodes and layout densities, the detection precision and robustness are remarkably improved, and the method is suitable for popularization and application. And the unprecedented early optimization capability is provided for photoetching research and development, so that the defect occurrence rate in the manufacturing process is reduced, and the product marketing time is shortened.
Owner:上海芯无双仿真科技有限公司