22 results about "Laplacian matrix" patented technology

The present application relates to the technical field of agricultural robot control, and particularly provides an anti-interference convergence method for an agricultural robot cluster based on interaction link dynamics. The method comprises obtaining state information and speed information of each interaction link, determining an interaction topology structure and a connection weight matrix between the interaction links; constructing an edge Laplacian matrix according to the connection weight matrix of the interaction links, and determining a control gain based on all non-zero eigenvalues of the edge Laplacian matrix; determining a control amount of each interaction link according to the state information and speed information of each interaction link, the state information and speed information of a neighbor interaction link, the control gain, and a random interference term; and controlling the state and speed of the corresponding interaction link according to the control amount of each interaction link. The method effectively suppresses the influence of external interference on the link state, guarantees the convergence of the link state, and improves the robustness and cooperative control ability in a complex airspace environment.

This invention discloses a robust fusion method for large models based on semantically aligned fuzzy clustering. The method includes: first, obtaining a sequence of probability distribution vectors from the outputs of multiple heterogeneous large-scale pre-trained models for the samples to be processed; then, introducing non-negative reliability weights to weight this sequence to obtain an aggregated probability matrix, and constructing a graph Laplacian matrix accordingly; second, approximating the aggregated probability matrix into a fuzzy membership matrix and a semantic prototype matrix, constructing a cost objective function by combining the graph Laplacian matrix, and iteratively updating each matrix and weight until convergence; finally, solving for the maximum value of the converged fuzzy membership matrix to obtain the sample prediction category result. This invention can effectively suppress the influence of inferior models in unsupervised environments, significantly improving the accuracy and robustness of fusion prediction.

The present application relates to the technical field of data processing, in particular to a kind of spectrum radius feature extraction and Hamiltonity determination method for mixed graph structure.It includes the following steps: obtaining mixed graph data and constructing adjacency matrix, Laplacian matrix and unsigned Laplacian matrix;Three spectrum radii are extracted using power iteration acceleration algorithm, and a spectrum radius pyramid is constructed by multi-layer coarsening, forming a multi-scale spectrum feature vector;The feature vector is input into the classification model to output the Hamiltonity probability value;The confidence score is calculated by multiple random dropout, and it is dynamically decided whether to activate the accurate determination engine;Finally, the determination result is output.The present application combines multi-scale spectrum features and statistical features, uses graph neural networks to achieve high-precision probability prediction, and improves the reliability of boundary samples through confidence evaluation and cascade determination architecture, suitable for graph data analysis in the fields of communication networks and bioinformatics.

The application belongs to the technical field of electric digital data processing, and relates to a road facility semantic topology lightweight graph neural network modeling method. Road facility data is parsed into a topology coordinate sequence and a semantic feature sequence; a topology and semantic double-flow branch is constructed, a graph Laplacian matrix is constructed in the topology branch according to the coordinate sequence, a first-order Chebyshev polynomial without a nonlinear activation function is used for graph filtering to extract a topology representation vector; in the semantic branch, the coordinate information is stripped, one-dimensional convolution is performed on the semantic feature sequence to extract a semantic representation vector; the two are spliced to generate a fusion representation vector; a pre-trained large graph neural network with an isomorphic topology is used as a teacher model, a mean square error of a teacher vector and the fusion representation vector is calculated to construct a distillation loss function, and network weights are updated through back propagation to obtain a lightweight model. The application reduces the amount of nonlinear operation and floating point operation, blocks the back propagation of topology noise, and shortens the inference response time.

This invention discloses a multi-task partial differential equation solution method based on spectral-physical manifold learning and graph diffusion operators. It acquires tasks with different geometric topologies (computational domain, boundary conditions, low-resolution sparse observation data); extracts spatial geometry, multi-scale dynamics, and energy mode features, fuses them into a hybrid spectral-physical embedding, constructs a task graph, and derives the graph Laplacian matrix; selects a central prototype task based on composite centrality; constructs a parameter-decoupled physical information neural network, dividing the parameters into an encoding module sharing an invariant differential operator and an adaptive module adapting to the geometric boundaries; uses the graph diffusion operator along the task graph topology to transfer the encoded parameters of the prototype task to the target task to initialize the network; after fine-tuning with data fitting and physical residual joint loss, it outputs a high-resolution continuous physical field prediction; this invention is not dependent on specific equation forms and is applicable to systems such as fluid mechanics, heat and mass transfer, and diffusion reactions.

This application provides a method, electronic device, and program product for estimating soil salinity in vegetated areas. By introducing two synergistic constraint mechanisms, it effectively overcomes the problem of unstable unmixing in traditional NMF (Natural Motion Modeling) in high vegetation cover (FVC>53%) scenarios. **Unique Spectral Constraint on Vegetation Endmembers:** Based on a pre-constructed vegetation calibration spectral library, similarity constraints are applied only to the spectra of vegetation endmembers, ensuring stable extraction of vegetation spectral features while cleverly avoiding excessive restriction on the decomposition space of soil endmembers, thus fully preserving key spectral information of soil salinity. **Smoothing Spatial Constraint on Abundance Matrix:** Utilizing the spatial correlation of hyperspectral imagery, a smoothing constraint is applied to the abundance of adjacent pixels through a Laplacian matrix regularization term, significantly reducing error accumulation and enhancing the spatial consistency and overall stability of the unmixing results.

The application belongs to the technical field of graph signal processing, and particularly relates to a space-time graph diffusion generation method based on spectral truncated latent space, which comprises the following steps: S1, acquiring graph topology information and node signal data; S2, constructing an adjacency matrix and a normalized Laplacian matrix; S3, extracting a low-frequency spectral truncated basis; projecting the node signal to the spectral truncated latent space to obtain latent variables; S4, performing diffusion processing in the spectral truncated latent space; S5, reconstructing the current latent variables to the node domain to obtain the current node domain state; estimating the noise or correction corresponding to the current step by using a node domain denoising module; S6, updating the current latent variables in the spectral truncated latent space according to the noise estimation result, and judging whether the preset termination step is reached, and if not, returning to S4; and S7, outputting the generated graph signal. The application can solve the problems of high state dimension, large reverse denoising calculation amount, and difficulty in considering local structure, global correlation and time information of the existing graph generation method on the edge device.

This disclosure presents a multimodal remote sensing image segmentation method, system, and electronic device. The method involves acquiring multimodal remote sensing images, preprocessing these images (including optical images and elevation images), extracting optical features from the optical images, and extracting elevation features from the elevation images. A graph Laplacian matrix is ​​constructed based on the optical and elevation features, and feature fusion is performed using this matrix to obtain target fusion features. A segmentation model is then invoked to predict the category probability based on the target fusion features, resulting in a predicted segmentation result for the multimodal remote sensing image. The target loss is determined based on the predicted segmentation result and the label segmentation result, and the segmentation model is trained based on this target loss. Finally, the trained segmentation model is invoked to segment the remote sensing image to be processed, resulting in a segmentation result for the remote sensing image to be processed. This method improves the segmentation accuracy and efficiency of multimodal remote sensing images.

The application discloses a carbon emission intensity evaluation method and device based on dynamic area division, a terminal equipment and a storage medium, relates to the technical field of power grid analysis, and comprises the following steps: acquiring real-time operation data of a power grid and a winning amount of a new subject, wherein the real-time operation data comprises a node impedance matrix and a time sequence of net power injection of each node; calculating an electrical distance according to the node impedance matrix; calculating a power fluctuation similarity according to the time sequence of net power injection; constructing a comprehensive similarity matrix according to the electrical distance and the power fluctuation similarity, then constructing a Laplacian matrix and obtaining an eigenvector; dividing a sub-region through clustering; for each sub-region, determining a regional carbon emission factor according to the real-time operation data and the winning amount, and evaluating the carbon emission intensity. Through implementation of the application, the problem that the existing carbon emission factor calculation is limited by administrative regions and ignores the physical correlation of the electrical distance and power flow of the power grid can be solved, and the accuracy of carbon accounting results is improved.

The application discloses a hyperspectral anomaly detection method fusing graph attention and beta wavelet graph network, which regards the pixel of a hyperspectral image as a graph node, regards spectral features as node attributes, establishes edge connection through a K nearest neighbor algorithm, and constructs a graph data structure containing spatial and spectral information. Firstly, the graph attention network is used to dynamically aggregate neighborhood information and enhance spatial context features. Secondly, a beta wavelet filter bank is constructed based on a graph Laplacian matrix, multi-scale frequency domain filtering is performed on node features, and the right shift phenomenon of spectral energy caused by an abnormal target is effectively captured. Finally, multi-scale filtering features are fused and mapped into an abnormal probability, and pixel-level anomaly detection is realized. Through the joint mining of nonlinear spatial correlation and unique spectral anomaly features of hyperspectral data by the graph neural network, the accuracy and robustness of anomaly detection in a complex scene are significantly improved.

This invention discloses a method, apparatus, and medium for multi-scale wavelet analysis of brain networks based on manifold learning. Using T1-weighted MRI and DW-MRI images, combined with Desctrieux mapping and probabilistic fiber tractography based on surface seeds, an initial adjacency matrix is ​​obtained. The average adjacency matrix is ​​calculated. Based on the node degree, betweenness, PageRank, and assignment coefficient of the average adjacency matrix, several nodes are selected from the brain network. Masks at different scales are calculated on the nodes. Multi-scale wavelets are initialized. The eigenvectors of the Laplacian matrix of the average adjacency matrix are solved using a power iteration method to obtain the optimal multi-scale wavelet. Protein signals in the brain network are projected onto this wavelet to obtain new biomarker signals from the brain network. Therefore, this embodiment of the invention uses manifold learning to calculate the mean of a brain network group, which better maintains the geometric topology of the network and, considering the hierarchical modularity and centrality of network nodes, better uncovers some potential physiological and pathological mechanisms in brain diseases.

The application relates to the technical field of computers and discloses a knowledge graph relationship reasoning optimization method, which comprises the following steps: acquiring, by a processor, graph structure data corresponding to a knowledge graph, and constructing a Laplacian matrix based on the graph structure data; performing, by the processor, feature decomposition operation on the Laplacian matrix to obtain a corresponding feature value set and a feature vector set; in a low-resolution spectral layer, performing aggregation processing on a plurality of entity nodes based on the similarity relationship between the feature vectors to generate a virtual super node; in a high-resolution spectral layer, preferentially reasoning the short path relationship between the entity nodes based on local spectral smoothing operation; and when the knowledge graph is updated, performing incremental update operation only on the spectral components affected by the update. By converting the adjacency matrix of the knowledge graph into spectral domain representation, using the Laplacian matrix to generate feature spectrum and dividing the feature spectrum into a low-resolution layer and a high-resolution layer, the application effectively solves the path explosion problem in traditional path searching.

This invention discloses a graph regularized slow feature analysis method for infrared nondestructive testing data, comprising: acquiring an infrared image sequence and preprocessing the infrared image sequence; subsequently, constructing a pixel-level graph Laplacian matrix using a block-parallel strategy, calculating pixel adjacency relationships within local image blocks based on bilateral weights of spatial distance and temperature similarity, and merging them to generate a global sparse adjacency matrix through a maximum value fusion strategy in overlapping regions; constructing a joint objective function that fuses temporal slowness constraints and spatial smoothness constraints, and optimizing the solution using an adaptive dual whitening and matrix trace ratio dynamic adjustment strategy; finally, solving the objective function to extract slow feature components and reconstructing the infrared image. This invention overcomes the computational bottleneck of high-dimensional thermal imaging data through an efficient computational strategy, and utilizes a physically-aware graph regularized slow feature analysis algorithm to achieve feature decoupling of multiple types of defects under spatiotemporal constraints, significantly enhancing the thermal feature edges and signal-to-noise ratio of defects.

The application provides a non-stationary industrial process monitoring method and system, and the method comprises an offline training stage: a time Laplacian matrix and a space Laplacian matrix are calculated based on a historical data matrix; a target function of a stationary subspace analysis method is constructed, and a time constraint term of the time Laplacian matrix and a space constraint term of the space Laplacian matrix are added to the target function; the target function is solved to obtain a stationary projection matrix; stationary components and monitoring indexes of each sample in the data matrix X are calculated in turn based on the data matrix X; a control limit is determined by using a kernel density estimation method; and an online monitoring stage: based on real-time running data x, stationary components of the real-time running data x and corresponding real-time monitoring indexes are calculated according to the stationary projection matrix, and if the real-time monitoring indexes are greater than the control limit, it is judged that a non-stationary process operation has a fault; and the application can improve monitoring accuracy.

The application discloses a non-stationary industrial process fault detection method based on graph representation learning and belongs to the technical field of industrial process monitoring. The method comprises the following steps: acquiring operation data, dividing the operation data into multiple continuous time segments; analyzing statistical differences of operation data of each time segment, and constructing a statistical difference objective function; constructing a similarity matrix of sample points across time segments, constructing a transition probability matrix and an adjacency matrix; generating a graph Laplacian matrix, and constructing a graph representation learning objective function; introducing a weight coefficient to jointly optimize the statistical difference objective function and the graph representation learning objective function, and obtaining a stationary projection matrix; determining a control limit by using the stationary projection matrix; and collecting real-time operation data and performing fault detection on the real-time operation data. Through fusion of stationary subspace analysis and graph representation learning, the data stability is realized, the internal characteristics of the data are revealed, and the accuracy of fault detection is improved.

This application discloses a method and system for predicting EEG motor imagery based on a multi-residual graph convolutional network. The method includes: first, using Chebyshev polynomials, converting the EEG electrode position relationships corresponding to the acquired EEG motor imagery signals of the subject into a Chebyshev bias matrix; then, employing a multi-residual graph convolutional network, sequentially performing the following operations: extracting layer-by-layer coarsening maps from the EEG motor imagery signals to obtain multi-level coarsening maps; processing the multi-level coarsening maps to obtain the Laplacian matrix of each level of coarsening map; fusing the Chebyshev bias matrix and the Laplacian matrices of each level of coarsening map based on a multiple residual computation mechanism to obtain a multi-dimensional graph representation vector; and performing feature projection on the graph representation vector to predict four types of motor imagery tasks for the subject. This application can improve the accuracy of predicting four types of motor imagery tasks for the subject.

The application provides a knee osteoarthritis typing model training method and device and a typing method, and the method comprises the following steps: constructing a data set by taking information of a patient in a phenotype characteristic as a data sample; taking two-dimensional coordinates obtained by dimension reduction as initial clustering centers, and constructing a minimum spanning tree according to connection lines between the clustering centers; generating an adjacency matrix and a Laplacian matrix according to the connection relationship; calculating the probability of the data sample being distributed to the clustering center according to the distance between the two-dimensional coordinates and the clustering center, and generating a soft distribution probability matrix and a diagonal matrix; solving a pre-set target function based on the above matrices to obtain an updated matrix set; solving the target function based on the updated matrix set, and when the target function value converges, taking the minimum spanning tree topology as a knee osteoarthritis typing model. The application can make the knee osteoarthritis typing result meet the requirements of clinical interpretability, continuous phenotype characteristics and stable and reliable results at the same time.

The present invention, with respect to a method of processing video data, provides a method of processing video data, provides a method characterized by comprising the steps of: performing a clustering for the video data; generating at least one data cluster as a result of the clustering; generating at least one Graph laplacian matrix corresponding to the at least one data cluster; performing conversion optimization on the basis of multiple graph-based models, wherein the multiple graph-based models respectively include at least one graph laplacian matrix; and generating an optimized conversion matrix according to the results of performing the conversion optimization.

The application discloses an intelligent analysis system and method for network operation state based on artificial intelligence, and relates to the technical field of network operation and maintenance.The method comprises the following steps: a network graph comprising a network node set, a link set and a topology adjacency matrix is constructed; operation indexes of each network node are collected to form an original observation data set; time sequence regulation processing is performed on the original observation data set to obtain aligned index data under a unified time axis; a historical baseline is constructed based on historical data of the network nodes, the aligned index data is compared with the historical baseline and is standardized to obtain a node comprehensive abnormal intensity and a network node abnormal intensity vector; a graph Laplacian matrix is calculated based on the topology adjacency matrix, a graph heat diffusion model is constructed to simulate an abnormal conduction process, and core propagation features are calculated; after normalization, root cause scores are obtained by weighting the core propagation features, root cause nodes are screened according to a root cause score threshold, and an abnormal propagation chain is combed and output.

This invention provides a public opinion analysis method based on LNMF, comprising: Step S1, collecting comment texts and associated metadata of public opinion events from target social media platforms to form multiple samples, and constructing a public opinion data matrix through feature extraction and normalization; Step S2, calculating the similarity between samples based on the public opinion data matrix to construct a Laplacian matrix reflecting the local geometric structure of the data; Step S3, constructing a manifold regularization term using the Laplacian matrix, decomposing the public opinion data matrix V into the product of a coefficient matrix and a basis matrix, and jointly optimizing the coefficient matrix and the basis matrix by minimizing the objective function that incorporates the manifold regularization term; Step S4, performing cluster analysis on the dimensionality-reduced feature representation based on the optimized coefficient matrix, and calculating the comprehensive influence index of the public opinion event based on the clustering results, comment texts, and associated metadata. The beneficial effect is that this invention enables more accurate, stable, and rapid analysis and insight into public opinion.