44 results about "Label propagation" patented technology

The invention relates to the technical field of data processing, and discloses a fault analysis method and system based on a label propagation algorithm. The method comprises the following steps: acquiring operation state data through a monitoring equipment network, extracting the data to obtain a system topology data set, performing label propagation fault analysis on the system topology data set to obtain node fault information, and performing fault influence weight analysis on each node based on the node fault information to obtain a maintenance priority, and inputting the influence weight data into the graph neural network for coping strategy analysis to obtain a target processing strategy. According to the invention, the efficiency of fault propagation analysis and the accuracy of decision making are improved.

The invention provides a large-model intelligent label synthesis and data automatic labeling integration method and system, and belongs to the technical field of label synthesis and data labeling, and the method comprises the steps: carrying out the semantic embedding and robust clustering of text data, and obtaining a stable cluster set; when new data is introduced, semantic consistency alignment of cross-round clustering results is realized through a confusion matrix matching strategy, and label drift is inhibited; maintaining an editable hierarchical label directed acyclic graph to support label system evolution; driving a large language model to generate a high-quality and interpretable cluster-level semantic tag based on the representative sample; carrying out automatic annotation and confidence evaluation by using large model context learning for clustering non-attribution or low-confidence samples; and propagating the cluster-level labels to the instances, and combining the cluster-level labels with an automatic labeling result to construct a full-process traceable label management mechanism. According to the method, the efficiency, quality and consistency of text labeling are improved, and powerful support is provided for large model training and intelligent data management.

The invention discloses a knowledge graph fusion method, device and equipment based on label propagation. The method comprises the following steps: acquiring a label propagation command, wherein the command comprises a knowledge graph list and a result output path; respectively executing a label propagation algorithm on each knowledge graph in the knowledge graph list, and determining initial label value distribution of each knowledge graph; determining an equivalent entity set among different knowledge maps; based on an initial label value and a local topological structure of each equivalent entity in the equivalent entity set in the knowledge graph, fusing the initial label value and the topological structure to obtain a fused entity, a fused label value and a fused topological structure; and according to the fusion entity, the fusion label value and the fusion topological structure, fusing the at least two knowledge maps in the knowledge map list to obtain a fusion result, and storing the fusion result in a result output path. According to the embodiment of the invention, cross-graph complex semantic calculation and conjoint analysis are supported.

The application provides a local consistency guided sparse label enhancement method, which is suitable for road drivable area detection and belongs to the technical field of images. The method aims to solve the problem that the existing deep learning model depends on a large amount of pixel-level label data. Firstly, the input image is sparsely labeled, and context enhancement features are constructed according to local and global image representations to establish the similarity relationship between superpixel nodes. Then, a label propagation model is constructed based on a graph convolution network to propagate sparse labels to unlabeled areas to generate pseudo labels. In the application, a local consistency guided weak supervision training strategy is adopted, and a joint loss function is designed to cooperatively supervise the labeled areas and the unlabeled areas, thereby improving the reliability of the pseudo labels and the overall segmentation accuracy. Experimental results show that the application can be applied to various road drivable area detection tasks, and the obtained high-quality pixel-level pseudo labels can be used for subsequent full-supervised model training.

The invention relates to the technical field of power system stability control, and provides a new energy black-start partition optimization method considering unconventional risks, which comprises the following steps: simulating unconventional risks possibly encountered by a power system by adopting Monte Carlo sampling; extracting the output characteristics of wind power under the unconventional risk by adopting variational mode decomposition, and performing output prediction in combination with a long-short-term memory neural network; a power system recovery partition is divided based on an LPA algorithm, and the LPA algorithm is improved by considering a partition size balance principle, so that the partition area is more balanced, and the tag oscillation effect is reduced. According to the method, the black-start partition optimization model fusing new energy output uncertainty modeling and unconventional risk scene generation is designed, and the partition scheme is cooperatively solved by adopting the improved label propagation algorithm, so that the robustness and the recovery efficiency of the black-start partition are effectively improved; and rapid and stable power supply of a power system under complex disturbance can be realized.

The application provides a large model intelligent label synthesis and data automatic labeling integrated method and system, and belongs to the technical field of label synthesis and data labeling. The method comprises the following steps: performing semantic embedding and robust clustering on text data to obtain a stable cluster set; when new data is introduced, the semantic consistency alignment of cross-round clustering results is realized through a confusion matrix matching strategy to suppress label drift; an editable hierarchical label directed acyclic graph is maintained to support label system evolution; a large language model is driven based on representative samples to generate high-quality and interpretable cluster-level semantic labels; for samples that are not attributed or have low confidence in clustering, automatic labeling and confidence evaluation are performed through context learning on the large model; the cluster-level labels are propagated to instances and combined with the automatic labeling results to construct a full-process traceable label management mechanism. The application improves the efficiency, quality and consistency of text labeling, and provides strong support for large model training and intelligent data management.

The invention discloses a data processing method, electronic equipment, a storage medium and a computer program product, and relates to the technical field of large model technology and data query. The method comprises the following steps: acquiring seed data; pre-labeling the seed data in a label propagation mode to obtain labeled data; performing data quality inspection on the labeled data to obtain training data; and training the initial intention classification model by adopting the training data to obtain a target intention classification model. The technical problems that the training cost of a model with a tool selection function is high, the efficiency is low, and the tool selection accuracy and the query accuracy of the model obtained through training are poor in the prior art are solved.

The invention discloses a power grid partitioning method based on partial differential morphology and label propagation, and the method comprises the steps: abstracting a power network into a weighted graph, and based on a selection strategy of electrical centrality, namely selecting k nodes with the maximum electrical centrality as seed nodes, obtaining a seed node set, and distributing a unique label for each seed; morphological diffusion and label propagation are combined in a power grid partition, and outward diffusion propagation is carried out from a seed node set; label propagation and morphological diffusion run in parallel to dynamically adjust label attribution of nodes, and each node is endowed with a final label after multi-round iteration when the cost of all the nodes does not change any more, so that division of the power grid is completed. According to the method, efficient partition control and management of the power grid can be effectively achieved, accurate identification and reasonable division of the topological structure of the power grid are achieved, and scientific support is provided for planning and management of a power system.

The invention discloses a label propagation text classification method and device generated by fusing weak supervision information, and relates to the technical field of text classification. The method comprises the steps of selecting an initial category word set, pre-training an initial multi-label text classification model, inputting an original text into an encoder layer to obtain deep potential features, and inputting the deep potential features into a prediction layer to obtain an initial classification prediction result; determining a pseudo label set; gradually updating the pseudo label set, and further training the pre-trained multi-label text classification model; an integrated pseudo label set is obtained; determining an adjacent matrix of the k-neighbor graph; determining a label correlation matrix; performing noise correction on the integrated pseudo label set, and performing final training on the multi-label text classification model; and according to the trained multi-label text classification model, obtaining a label corresponding to the to-be-classified text. According to the invention, the noise supervision information is corrected by using the text neighbor relation and the label correlation, so that the classification accuracy is improved.

The application discloses a hardware design security vulnerability qualitative analysis method and system, according to a mixed attribute label propagation logic element library and a mixed attribute model construction method based on discrete mapping, so that a corresponding mixed attribute model can be constructed for any HDL design in linear time; the label designed by the mixed attribute model integrates two types of attributes, so that the safety behavior related to the safety attribute and the clock attribute can be modeled simultaneously; the method provided by the application can realize effective discrimination of hardware Trojan horses and hardware time measurement channels by verifying the clock attribute and the safety attribute; the application is deployed in the design and verification stage of the EDA process, so that the HDL design security vulnerability can be detected early, and the design basis is provided for high-reliability hardware design.

The invention relates to a multi-level community identification method with a priority core node and a similar structure in a population flow network, and belongs to the technical field of community detection and complex network processing. The method comprises the following steps: quantizing the significance of a flow edge weight through an exceeding probability, and repeatedly extracting core node pairs to combine and form a microscopic community; identifying a core node by adopting a C-index centrality measurement index, and performing label alignment on the microscopic community and the core node; based on the node destination travel distribution sequence, calculating structural similarity by using a dynamic programming algorithm; and carrying out label propagation by combining with structure similar neighbors to generate a final macroscopic community. According to the method, the diversity of flow directions, the relative importance of flow edges and the distribution characteristics of travel structures in a flow network are comprehensively considered. Compared with a traditional algorithm, the method can more accurately describe the central distribution and the multi-layer structure of the population flow network, and can provide a scientific basis for urban agglomeration evolution analysis, traffic planning and sustainable city development.

The invention relates to the technical field of voice keyword detection, in particular to an anti-noise detection method and device for voice keywords, equipment and a storage medium. Comprising the steps of performing acoustic feature extraction on an original audio signal to obtain an acoustic feature sequence; calculating an attention weight vector in combination with the target phoneme sequence and the acoustic feature sequence; performing weighted fusion on the acoustic feature sequence based on the attention weight vector to obtain a fused feature sequence; performing feature coding on the fused feature sequence to obtain a task feature sequence, and obtaining posterior probability distribution; updating a preset decoding graph according to the attention weight vector to obtain a dynamic decoding graph, and performing identification transmission decoding based on the dynamic decoding graph and the posterior probability distribution to generate at least one candidate path; and determining confidence scores of all candidate paths, and determining a keyword detection result corresponding to the original audio according to the confidence scores. According to the application, the reliability of the keyword wake-up technology in the application can be improved.

The invention relates to a hyperspectral image dimension reduction method, which comprises the following steps of: firstly, constructing a similarity matrix of space-spectrum joint features of marked and unmarked samples based on a direct push learning strategy, and iteratively generating reliable pseudo labels of the unmarked samples through a label propagation algorithm; secondly, taking the training sample as a center, constructing three-dimensional tensor blocks by using a space window with a fixed size, and forming a training set; thirdly, taking a tensor block as a node, measuring similarity through a covariance descriptor and an earth movement distance (EMD), constructing a similar graph based on a locality preserving projection (LPP) rule, and embedding the constructed graph into a dimension reduction model; then, low-rank decomposition is introduced on the basis of the graph embedding model, and noise reduction and dimension reduction are synchronously achieved; the category information of the marked sample is used as a constraint, and the low-dimensional feature discrimination is improved; and finally, solving the integrated model by using an alternating direction multiplier method (ADMM) to obtain an optimal projection tensor, acting the optimal projection tensor on the hyperspectral image to obtain low-dimensional features, and inputting the low-dimensional features into an SVM classifier to complete ground feature classification.

The invention provides an overlapping community detection method based on attribute perception and dynamic similarity fusion, and the method comprises the steps: introducing a graph convolutional network (GCN) model to achieve the collaborative modeling of node structure proximity and attribute proximity, obtaining a low-dimensional representation vector of the network, carrying out the dot product of the low-dimensional representation vector, and obtaining a low-dimensional representation vector of the network; the global similarity of fusion structures and attribute information among the nodes is obtained, and then the problem that the structures and the attribute information of the nodes cannot be captured at the same time through a label propagation method is solved; and dynamic community correlation among the nodes is introduced, and semantic consistency in the community is captured, so that in the label propagation process, not only is the local neighborhood relationship of the nodes considered, but also the similarity of the nodes in the community semantic level can be reflected, and the semantic coherence of community division is improved. The node labels are updated according to a fixed sequence, the importance of all nodes in the network is calculated, the nodes are sorted from large to small, and the node labels are updated according to the sequence. When the node labels are updated, the label sets of the nodes to be updated are updated through the main labels of the neighbor nodes, and finally iterative training is carried out till the method converges.

The application discloses a noise label robust training method and device for an image recognition model, and belongs to the technical field of image recognition model training, and comprises the following steps: inputting image training samples into a deep neural network for feature extraction; constructing a multi-granularity granular ball structure in a feature space based on a feature vector; performing hierarchical correction on the labels of the image training samples; performing label propagation in the granular ball, and calculating the propagation confidence distribution and consistency score of each image sample; screening a clean sample subset according to the consistency score, and iteratively training the deep neural network to update network parameters. Through the adaptive multi-granularity granular ball division mechanism, the application can fully depict the local structure characteristics of data in the feature space, avoid the problems of excessive fragmentation or insufficient purity caused by traditional fixed-granularity clustering in a high-noise environment, and provide a stable and reliable structure prior for subsequent label correction and information propagation.

The application provides a kind of telecommunication anomaly detection method and detection device.The method comprises: obtaining the social relationship information and portrait feature information of each user in personnel relationship heterogeneous graph, obtaining the fusion feature information and relationship feature information of each user according to label information, portrait feature information and social relationship information;According to fusion feature information and relationship feature information, adopt neighborhood aggregation algorithm, obtain the node feature information of label user;Based on semi-supervised learning mechanism, according to social relationship information and node feature information, through improved attention mechanism model, obtain the anomaly probability of each user, to realize through label propagation algorithm, the label information of label user is propagated to unlabelled user, and anomaly detection is completed.The method of the application utilizes the context information of multiple relationships in heterogeneous graph and the characteristics of semi-supervised learning, effectively improves the detection accuracy and adaptability of the model in the label scarce scene.

The invention relates to a software aging prediction method and device based on feature reconstruction and tag guidance. The method comprises the following steps: acquiring a source code of target software and an aging tag sample corresponding to the source code; extracting a metric feature and a graph structure feature from the source code, and reconstructing the metric feature and the graph structure feature into a feature image; based on the feature image, obtaining aging feature representation of the source code through self-training of contrast learning optimization; generating a pseudo tag for an unlabeled code unit in the source code by using tag propagation assisted by a buffer area and the aged tag sample; constructing a data set based on the source code, the aged tag sample and the pseudo tag; and based on the aging feature representation and the data set, training a prediction model and performing aging prediction on the target software. Through feature reconstruction and buffer area assisted label propagation, the stability and reliability of the pseudo label are improved, and the problem that the quality of the pseudo label is poor in the initial training stage of a traditional self-training method is solved.

The invention belongs to the technical field of machine learning, and relates to a heterogeneous federal map learning method and system. The method comprises the following steps: in each client, constructing a globally shared symbiotic space, and generating a unified target semantic prototype through a label propagation mechanism; generating a prototype distribution matrix, and uploading the prototype distribution matrix to a server side; generating a global prototype by aggregating the prototype distribution matrix of each client, and distributing the global prototype to each client; embedding and mapping the nodes into a hash bucket by using a hash function, generating aligned local hash codes and local anchor point embedding, and uploading the local hash codes and the local anchor point embedding to a server side; aggregating the local hash codes and the local anchor point embedding to obtain global hash codes and global anchor point embedding; global hash coding and global anchor point embedding are optimized through a graph auto-encoder and consistency constraint, and alignment of the concentrated graph is optimized. According to the invention, the flexibility and efficiency of federal map learning are improved, the data privacy and security are ensured, and the communication overhead is reduced.

The application discloses a data processing method and related device, obtains a training sample matrix composed of labeled image samples of a source domain and unlabeled image samples of a target domain, a sample label of the labeled image samples is used for identifying category information of the labeled image samples, an initial field alignment matrix, an initial global similarity matrix and an initial predicted label matrix are updated according to the training sample matrix, a target alignment matrix, a target global similarity matrix and a target predicted label matrix are obtained, a check parameter is constructed according to the training sample matrix, the target alignment matrix, the target global similarity matrix, the target predicted label matrix and a source domain sample label matrix, if the check parameter does not satisfy a first convergence condition, iterative updating is performed until the first convergence condition is satisfied, and it is considered that training of an image recognition model of the target domain is completed. Label propagation can make the labeled image samples of the source domain be used in image recognition of the target domain, and improve training efficiency of the image recognition model of the target domain.

The invention discloses a noise label robust training method and device for an image recognition model, and belongs to the technical field of image recognition model training, and the method comprises the steps: inputting an image training sample into a deep neural network for feature extraction; constructing a multi-granularity particle-ball structure in the feature space based on the feature vector; performing hierarchical correction on the labels of the image training samples; performing label propagation in the globules, and calculating propagation confidence distribution and consistency score of each image sample; and screening out a clean sample subset according to the consistency score, and carrying out iterative training on the deep neural network to update network parameters. According to the method, local structure features of data can be fully described in a feature space through a self-adaptive multi-granularity particle and ball division mechanism, the problem of excessive fragmentation or insufficient purity generated by traditional fixed-granularity clustering in a high-noise environment is avoided, and stable and reliable structure priori is provided for subsequent label correction and information propagation.

This invention discloses a multi-temporal information joint hyperspectral change detection method based on graph transformer guidance. By constructing a topological graph, the network can propagate labels to unlabeled samples in a semi-supervised learning mode, reducing the dependence on a large number of training samples during training. A graph transformer module with significant relationship enhancement is proposed. This module applies the transformer to a graph structure with superpixels as nodes, enabling the acquisition of affinity relationships between any two sub-regions under low computational cost, breaking through the traditional definition of adjacency relationships and capturing change information in the entire image. A gated change information fusion unit is proposed, which aims to effectively inject the change features guided by the GTrans module into the original dual-temporal stitching features, realizing the organic fusion of multi-temporal information.

The invention discloses a large-scale data-oriented feature self-weighted anchor graph learning and structured clustering method, which comprises the following steps of: performing information acquisition and feature extraction on a large-scale data object, and performing feature normalization preprocessing to obtain a data feature matrix; establishing a structured anchor graph learning model based on feature self-weighting; optimizing constraint conditions of the structured anchor graph learning model through an alternating direction multiplier method to obtain an anchor graph and a class label thereof; and calculating a clustering result of the data feature matrix through a class label propagation method. According to the clustering method, efficient graph learning and clustering can be carried out on large-scale data influenced by noise and redundant information in a complex scene, and the robustness and accuracy of a traditional subspace clustering method are improved.

The application relates to the technical field of intelligent education management, in particular to an information-based teaching test system with adaptive learning progress, which comprises the following contents: a knowledge node construction module, a learning state recognition module, a test task generation module, a path scheduling adjustment module and a feedback data backtracking module.In the application, the knowledge point fine numbering management is realized through the construction of a directed edge numbering set, the dependence direction is set in combination with a label propagation algorithm, the semantic expression capability of the node relationship is enhanced, the state weight sequence is formed based on the matching of the answer behavior matching weight and the stability coefficient, the stability of the classification recognition is improved, the cognitive ability coding and the time consumption median value linkage state weight are introduced, the questions are filtered and a priority queue is constructed, the precision of the task pushing is improved, the path chain is constructed by fusing the numbering and the sorting weight, the path branch is expanded by calculating the dependence strength, and the individual adaptability and the task rhythm coordination capability of the path scheduling are enhanced.

A computer-implemented method for determining a set of target items to be annotated to train a machine learning application. The method includes providing a training dataset having a set of data samples and an autoencoder having a classifier. The autoencoder includes an embedding model that maps the set of data samples to a set of compressed feature vectors. The set of compressed feature vectors defines a compressed feature matrix. The method further includes: defining a graph associated with the compressed feature matrix; applying a clustering algorithm to identify node clusters in the graph and applying a centrality algorithm to identify the central nodes of the node clusters; retrieving node labels from annotator nodes for the central nodes; propagating the annotated node labels to other nodes in the graph; and performing training of the embedding model and the classifier using the annotated node labels and the propagated node labels.

The application discloses a kind of based on the incremental semi-supervised image clustering method and system of double-layer label propagation, mainly used to solve the problem of low efficiency caused by repeated calculation when static semi-supervised clustering method faces incremental image data and incremental pairwise constraint.This application uses double-layer label propagation to process the clustering problem of increasing image data and constraint condition.In the first layer label propagation, propagate and diffuse pairwise constraint information in image data samples, and combine the membership matrix of component of image data samples at the last time, incrementally calculate the membership matrix of component of image samples at the current time.In the second layer label propagation, use the clustering result at the last time to mark cluster label information in component, and let known cluster label information propagate in component structure, then gradually expand cluster label information to the whole image dataset through the membership relationship of image sample pair component, so as to realize effective semi-supervised clustering of incremental image data.

This disclosure relates to a method and system for multi-sensor fusion in the presence of missing and noisy labels. Prior methods for multi-sensor fusion do not estimate and correct labels for learning effective models in semi-supervised learning methods. Embodiments of the present disclosure provides a method for learning robust sensor-specific autoencoder based fusion model by utilizing a graph structure to perform label propagation and correction. In the disclosed Graph regularized AutoFuse (GAF) method latent representation for each sensor is learnt using the sensor-specific autoencoders. Further these latent representations are combined and fed to a classifier for multi-class classification. The disclosure presents a joint optimization formulation for multi-sensor fusion where label propagation and correction, sensor-specific learning and classification are executed together.

The invention provides a local consistency guided sparse label enhancement method, which is suitable for detecting a drivable area on a road and belongs to the technical field of images. The method aims at solving the problem that an existing deep learning model depends on a large amount of pixel-level annotation data, firstly, sparse annotation is conducted on an input image, context enhancement features are constructed according to local and global image representation, and the similarity relation between super-pixel nodes is established; and then constructing a label propagation model based on the graph convolutional network, and propagating the sparse labels to the unlabeled areas to generate pseudo labels. According to the method, a weak supervision training strategy guided by local consistency is adopted, a joint loss function is designed, and collaborative supervision is carried out on a labeled region and an unlabeled region, so that the reliability of a pseudo label and the overall segmentation precision are improved. Experimental results show that the method can be suitable for various road drivable area detection tasks, and the obtained high-quality pixel-level pseudo label can be used for subsequent fully supervised model training.

The present application relates to the technical field of three-dimensional point cloud data processing, and particularly relates to a point cloud data fast labeling method and system. The method comprises the following steps: obtaining point cloud data and initial point labeling data; constructing a superpoint structure according to the point cloud data to obtain point cloud superpoint structure data; mapping the initial point labeling data to the point cloud superpoint structure data to obtain label mapping data; performing local topological fidelity label propagation on the label mapping data to obtain first label propagation data; performing projection-guided aggregation on the label mapping data to obtain second label propagation data; obtaining label propagation feature data according to the first label propagation data and the second label propagation data; fitting the confidence of the first label propagation data and the second label propagation data according to the label propagation feature data to obtain label propagation data; and performing label reflection on the label propagation data to obtain point cloud labeling data. The present application realizes efficient and accurate expansion of point cloud label information.

The invention provides an abnormal intelligence message identification method and device, equipment, a storage medium and a program product, which can be applied to the technical fields of big data, information security and financial science and technology. The method comprises the steps that message nodes of a pre-constructed knowledge graph are finely adjusted, an embedded vector is obtained, the message nodes of the knowledge graph are constructed according to intelligence information, and user nodes of the knowledge graph are constructed according to user information; according to the edges of the knowledge graph, positive sample pairs and negative sample pairs of message nodes in the knowledge graph are determined, the positive sample pairs are two nodes having first-order edges with the same user node, and the negative sample pairs are two nodes having no first-order edges with the same user node; optimizing an embedding vector of the message node according to the positive sample pair and the negative sample pair to obtain an optimized embedding vector; according to the optimized embedded vector and the message nodes marked with the labels, the labels are propagated to the message nodes not marked with the labels; and determining an abnormal message node according to the label.

The invention provides an adaptive source positioning method and device based on a conditional diffusion model, and a storage medium, relates to the technical field of network propagation source positioning, and aims to solve the problems of low precision and robustness of network propagation source positioning. The method comprises the following steps: acquiring network topology information and an observation infection state vector of a target network, wherein the observation infection state vector is used for representing an infection state of a node in the network; based on the network topology information and the observation infection state vector, a label propagation algorithm is adopted to calculate a source estimation vector, and the source estimation vector is used for representing the initial confidence of each node as a propagation source; the network topology information, the observation infection state vector and the source estimation vector are input into a conditional diffusion model, a source position vector is generated by executing a reverse denoising process, and the conditional diffusion model is a generation model for guiding denoising through conditional information; and determining the position of the network propagation source node according to the source position vector.