132 results about "Incidence matrix" patented technology

The invention discloses a large model prompt project optimization system and method fusing a domain knowledge graph. The system comprises an analysis module, a template generation engine module, a large model interaction interface module, a feedback analysis module and an optimization strategy module. And the analysis module forms a constraint coding signal containing an entity attribute incidence matrix. The template generation engine module forms an enhanced prompt text stream with a reservoir physical property parameter slot; the large model interaction interface module receives the enhanced prompt text stream and generates a question and answer response data stream containing geological terminologies; the feedback analysis module forms a feedback signal containing semantic deviation measurement through a semantic error vector calculation algorithm; and the optimization strategy module forms a parameter optimization instruction signal and transmits the parameter optimization instruction signal to the analysis module to complete iterative updating of the constraint condition. According to the large model prompt project optimization system fusing the domain knowledge graph, the problem of low answer accuracy of a large model in the oil-gas exploration field due to lack of professional constraints can be solved.

The invention discloses a multi-point ditch flow monitoring method, and particularly relates to the field of branch ditch flow monitoring, which is used for solving the problem of complex branch ditch flow monitoring and optimization, and comprises the following steps of: constructing a branch ditch topological connection relation matrix, generating a flow characteristic curve in combination with flow and water level time sequence data, and marking interference points; generating an interference incidence matrix by utilizing two-channel interaction analysis, extracting upstream and downstream flow offset and disturbance propagation characteristics, judging an interference dominant mode according to a difference degree, and respectively calculating a flow offset compensation matrix and a disturbance propagation correction matrix; the recursive optimization model realizes segmented correction and reduces error residual errors, combines spatial-temporal feature clustering to identify abrupt change sections and generate a high-risk area key monitoring list, extracts refined data based on a high-frequency acquisition strategy to analyze abnormal causes, generates a dynamic correction result by using a prediction model, and updates a traffic balance data set. The multi-mode interference is accurately corrected, the flow distribution is optimized, and the monitoring precision and the system stability are improved.

The invention provides a virtual power plant platform source network load storage equipment real-time monitoring and optimization method and system, and relates to the technical field of virtual power plant intelligent optimization scheduling, and the method comprises the steps: collecting equipment operation data, carrying out the noise reduction and feature compression to obtain feature data, and recognizing the equipment fault type based on the feature data; carrying out decoupling reconstruction on the feature data in time and space dimensions to obtain an equipment incidence matrix, constructing an equipment topological graph based on the equipment incidence matrix and a credible fault type, carrying out bidirectional tracking on the equipment topological graph according to a fault diffusion probability, and generating a fault propagation path; grading and sorting the equipment on the fault propagation path, and outputting a fault influence prediction result and an equipment priority list; and carrying out cooperative scheduling optimization on the equipment in the virtual power plant by adopting a hierarchical reinforcement learning algorithm to realize cooperative scheduling optimization of the equipment in the virtual power plant.

The invention provides a self-adaptive scene analysis and target generation method and system based on deep learning, and relates to the technical field of computer vision, and the method comprises the steps: extracting the depth information of a scene image sequence, constructing three-dimensional point cloud data, calculating the distance and direction matrix between objects, and generating a scene space feature vector. Meanwhile, a semantic label graph is obtained through semantic segmentation, a semantic incidence matrix is constructed, and a scene semantic vector is calculated. Inputting the two into a variational auto-encoder to obtain a target feature vector, calculating an attention score based on the target feature vector, and generating a target description vector with a weight; and finally, target contour features are generated through a probability graph reasoning network, and a final target result is generated in combination with the regional attention graph. According to the method, the accuracy and efficiency of target generation are improved.

The invention relates to the technical field of data processing, and discloses a multi-granularity extraction and enhancement method, device and equipment for multi-source heterogeneous data. The method comprises the following steps: performing multi-level feature extraction on a multi-modal data set and original feature parameters to obtain a multi-modal feature set; performing correlation analysis on the multi-modal feature set to obtain a feature correlation matrix and optimized feature representation; performing cross-modal knowledge migration processing on the feature incidence matrix and the optimized feature representation to obtain migration feature parameters; performing dynamic fusion processing on the migration feature parameters to obtain target fusion features; and performing multi-task segmentation processing according to the target fusion features to obtain a plurality of sub-task sequences, performing time sequence enhancement and multi-task fusion analysis on each sub-task sequence to obtain target extraction data, and performing visual display on the target extraction data. According to the method and the device, the efficiency and the accuracy of multi-granularity extraction and enhancement for multi-source heterogeneous data are improved.

The invention discloses a carbon neutralization target-oriented electricity-carbon-green certificate multi-target scheduling optimization method and system, and relates to the technical field of carbon neutralization control, and the method comprises the following steps: obtaining first data of a target energy storage unit, constructing a constraint generation network, and dynamically updating an energy storage operation constraint domain based on a dynamic time warping algorithm; based on the updated operation constraint domain, establishing an electricity-carbon coupling model, extracting a nonlinear mapping relation between an electric energy output behavior and carbon emission intensity, and outputting a dynamic incidence matrix through a multi-dimensional time sequence predictor; constructing a carbon flow distribution topological graph according to the dynamic incidence matrix, and obtaining an optimal carbon flow path through a Dijkstra algorithm; based on the operation constraint domain and the optimal carbon flow path, establishing a multi-subject collaborative decision-making mechanism, solving a multi-target unit equilibrium solution by adopting a distributed alternating direction multiplier method, and generating a global scheduling instruction set; according to the method, a carbon flow topology and dynamic constraint domain cooperation mechanism is constructed, so that efficient solving of a multi-main-body carbon response optimization scheduling problem is realized.

The invention belongs to the field of multi-modal artificial intelligence risk identification. The core comprises a multi-source heterogeneous data parallel acquisition module; a modal exclusive feature extraction module; a graph attention driven dynamic fusion module; and a risk classification module with an attention mechanism. A real-time fusion weight is generated through a cross-modal incidence matrix, adaptive feature weighting and hidden risk association mining are realized, and the recognition accuracy and interpretability of a complex scene are remarkably improved. The method is applied to the fields of financial risk control and industrial monitoring.

The invention relates to the technical field of construction risk early warning, in particular to a BIM-based building construction risk early warning method and system, and the method comprises the following steps: constructing an incidence matrix between risk nodes based on time and space dimensions for the risk factor of each node in a construction process, calculating the risk propagation intensity value between the nodes, and calculating the risk propagation intensity value between the nodes; and judging the risk association relationship between the nodes, performing grading processing on the risk coupling degree according to the propagation rate, and outputting a risk coupling matrix value. According to the method, a high-risk area is extracted by combining spatial distribution data, dynamic partition management is achieved, dynamic influences of environmental variables such as temperature, humidity and vibration are quantified, the complex relation between environmental conditions and risk states is revealed, the accuracy and real-time performance of risk assessment are optimized, and the risk assessment accuracy is improved by analyzing the risk transfer trend and propagation path changes. And dynamically adjusting node priorities and resource allocation, updating the influence of newly added risk signals in real time, and realizing full-process dynamic optimization of construction risk early warning.

The invention discloses a conductor instance detection method and system based on key point matching, and the method comprises the steps: obtaining original aerial image data, generating a four-dimensional feature matrix through dynamic feature mapping and enhancement processing, obtaining three-dimensional projection feature data through projection transformation, and carrying out the feature enhancement to obtain enhanced feature data. Generating enhanced feature map data through multi-scale feature fusion; performing feature recombination on the enhanced feature map data to obtain feature matrix data, generating response matrix data through recursive decomposition and mapping, and calculating adaptive threshold data to screen key points; generating thermodynamic diagram data for the screened key point set through a multi-layer thermodynamic diagram, obtaining a spatial incidence matrix by combining spatial feature enhancement, and performing endpoint matching based on multi-dimensional constraints; and finally, through boundary feature analysis and cross-block instance matching, a global traverse map is constructed, and accurate detection of traverse instances is realized. According to the invention, the accuracy and robustness of wire detection can be effectively improved.

The invention discloses an intangible cultural heritage multi-modal knowledge graph construction method and system based on big data, and belongs to the field of knowledge graph construction, and the method comprises the steps: collecting intangible cultural heritage data of different modalities, carrying out the preliminary cleaning and labeling of the collected data, building the preliminary association between the data of different modalities, and carrying out the preliminary cleaning and labeling of the data of different modalities; obtaining a first edition multi-modal data set; processing the initial multi-modal data set through a multi-modal data cleaning and labeling model to obtain a final multi-modal data set, a labeling consistency matrix and a cross-modal incidence matrix; constructing a unified domain knowledge model, and performing structured modeling on intangible cultural heritage related concepts, entities and relationships thereof to form a semantic expression and knowledge organization form supporting multi-modal data; and constructing a structured multi-modal knowledge graph. According to the method, multi-source heterogeneous non-perpetual data can be effectively integrated, and the acquisition and utilization efficiency of non-perpetual knowledge is improved.

The invention discloses an intelligent terminal function test method and system, and belongs to the technical field of intelligent terminal function test. The method comprises the steps that a dynamic demand tracking mechanism is constructed, UI elements are analyzed, and a topological graph marking resource occupation types and hierarchical dependence is generated; establishing a demand-use case-defect incidence matrix to realize automatic update of the use case state when the demand is changed; executing a test based on the topological graph and the matrix, traversing a path according to a preset strategy, collecting data and comparing an expected result; and optimizing iteration according to a test result, adjusting a node weight and a matrix mapping relation, and generating a new test path and a use case suggestion. The system comprises a dynamic demand tracking module, an incidence matrix management module, a test execution module, an optimization iteration module and a data storage module. According to the method, the test range is accurately adapted to the demand change, the test coverage and the defect tracing efficiency are improved, a test strategy closed-loop optimization mechanism is formed, and the test requirements of a complex system and an agile development mode are met.

The invention discloses a hyperspectral and multispectral image fusion algorithm. The algorithm comprises the following steps: step 1, constructing a fusion model based on regular term constraint; step 2, constructing a multi-view clustering model of hypergraph manifold and low-rank tensor constraint; 3, performing spatial and spectral manifold learning based on a multi-view clustering model; 4, introducing a regular hyperspectral and multispectral image fusion model by utilizing the space and spectral manifold constraint obtained by learning; and 5, solving a fusion problem by adopting an ADMM algorithm. On the basis of inspiration of a low-dimensional manifold structure in a hyperspectral image, by considering space and spectrum height correlation and utilizing manifold constraints and high-dimensional characteristics to cluster an internal low-dimensional space of the hyperspectral image, a low-dimensional structure of the hyperspectral image in space and spectrum dimensions is described, and multi-view constraints are introduced to ensure the accuracy of an incidence matrix. A solution space of a to-be-restored hyperspectral image is constrained, self-expression constraints of a space and a spectral low-dimensional space are introduced, and a proposed algorithm is optimized by using an alternating direction multiplier method.

The invention relates to the technical field of precision machining, in particular to a cutting force signal synchronous calibration method based on multi-sensor data fusion, which comprises the following steps of: 1, dynamically preprocessing acquired signals; step 2, constructing a reference time axis based on the spindle current signal, and realizing time domain alignment of multichannel signals by adopting a dynamic time warping algorithm; 3, establishing a feature incidence matrix containing a force-vibration-temperature coupling relation, and generating a fusion cutting force feature vector through frequency domain energy weight distribution; 4, a dynamic transfer function model is constructed, model parameters are updated in real time, and the fusion feature vector is mapped into a calibration cutting force signal; and 5, carrying out dynamic compensation on the calibration signal by adopting a nonlinear inverse compensation and residual modal screening strategy. According to the method, the accuracy of cutting force signals can be remarkably improved, errors and uncertainty in the machining process are reduced, and the production quality and efficiency are optimized.

The invention discloses a management decision-making method and system based on a knowledge base construction technology. The method comprises the following steps: performing sequential relationship extraction on multi-source financial data to obtain a sequential relationship set related to query content; constructing an event-entity incidence matrix corresponding to the time sequence relation set; according to the time sequence relation set and the event-entity incidence matrix, constructing a dynamic knowledge graph; determining causal effect parameters in the causal graph structure by adopting a dual machine learning model; constructing a structural causal model according to the causal graph structure and the causal effect parameters; and generating an anti-fact prediction result by using the structural causal model, and generating a decision scheme corresponding to the query content based on the anti-fact prediction result. The technical problem that decision information including accurate causal basis and prospective simulation information cannot be generated due to the fact that the causal relationship between financial data is difficult to determine and the intervention effect cannot be dynamically deduced in a related management decision method is solved.

The invention discloses an information fraud identification method based on big data and artificial intelligence, and relates to the technical field of information security, and the method comprises the steps: S1, multi-source data collection and processing, S2, dynamic heterogeneous graph construction, S3, feature encoder design, S4, heterogeneous space-time attention mechanism, S5, multi-modal fusion, and S6, dynamic fraud detection model. According to the method, by introducing a heterogeneous space-time diagram attention mechanism, the fraud detection precision is improved, specifically, by designing a space-time incidence matrix and dynamically calculating the space-time attention weight between nodes, abnormal behaviors can be effectively recognized, and the fraud detection accuracy is improved. The space-time correlation matrix combines node features and time difference information, the dynamic evolution law of user behaviors can be captured, the complex fraud mode with space-time correlation can be more accurately recognized through the design, and the accuracy and reliability of fraud detection are improved.

The invention discloses a high-precision dynamic calibration system and calibration method for an ultrasonic partial discharge instrument, and relates to the technical field of dynamic calibration of the ultrasonic partial discharge instrument, and the system comprises a collection and enhancement module which is used for deploying an ultrasonic sensor in a target area to collect a partial discharge signal and carrying out the denoising and enhancement of the partial discharge signal; the feature extraction module is used for extracting features according to the denoised and enhanced signals; the matrix and calculation module is used for constructing a space-time incidence matrix based on the extracted features and dynamically calculating node weights; the error calibration and verification module is used for constructing an error model according to the node weight to carry out error calibration and verification; according to the method, the anti-noise capability is remarkably enhanced through a denoising method combining sparse decomposition and a sparse dictionary matrix, global and local features of the signals are extracted through Henkel matrix construction and singular value decomposition, and high-quality input is provided for the follow-up process through optimization reconstruction and standardized fusion of the signals; and high-precision calibration of the partial discharge signal in a complex environment is realized by combining the dynamically calculated node weight and the error calibration model.

The invention discloses a multi-mode federated learning method and system, computer equipment and a readable storage medium, and belongs to the technical field of federated learning. The multi-modal federated learning method comprises the following steps: on each client node, mapping local data of various modals into a plurality of vectors in a unified semantic space, determining an incidence matrix of the data of the various modals, and fusing the plurality of vectors according to the incidence matrix to obtain a local semantic vector; training a local model by using the local semantic vector to obtain local model parameters, and uploading the local model parameters to a server; on the server, identifying the difference degree between the data distribution condition of each client node and the global data distribution condition, and determining the node weight vector of each client node; and performing weighted aggregation on the corresponding local model parameters by using the node weight vector of each client node to generate global model parameters for next federated learning. Therefore, the performance of the training model can be improved.

The invention relates to the technical field of production process intelligent regulation and control, in particular to an intelligent regulation and control method and system for an instant gelatin production process, and the method specifically comprises the following steps: firstly, collecting process parameters such as reaction kettle temperature, acid-base concentration and the like in a production line sensor and control system to form production batch data; then, an optimization model and an intelligent regulation and control strategy are constructed, a dynamic incidence matrix is constructed through time-varying mutual information entropy, multiple types of features are fused to generate a process state descriptor, a constraint dominating relation and a population initialization strategy are optimized, and a self-adaptive genetic manipulation adjustment mechanism and a state-guided search and constraint processing method are designed; iteratively executing an improved NSGA-II algorithm, obtaining an optimized Pareto solution set, and evaluating diversity; and finally, selecting an optimal solution through a multi-criterion decision and a dynamic regulation and control strategy, and converting the optimal solution into an executable process parameter set value. According to the method, the accuracy and the high efficiency of the production process can be improved by realizing intelligent optimization regulation and control of the process parameters.

The invention discloses a multi-source heterogeneous data processing method and system based on a test chip, and the method comprises the steps: employing a distributed message queue to access a multi-protocol data stream in real time, collecting the multi-source heterogeneous data related to a chip test, and carrying out the normalization preprocessing; performing cross-domain spatio-temporal feature alignment and mapping on the preprocessed multi-source heterogeneous data, constructing a hybrid incidence matrix based on linear correlation degree and nonlinear correlation degree, extracting three-dimensional joint features, and generating a dynamic metadata tag; generating nodes and edges by using the incidence matrix and the extracted three-dimensional joint features, constructing a parameter association graph structure, performing graph embedding training and generating an index structure, positioning Top-K candidates based on the index structure by using query features and labels, and sorting and returning a result according to a comprehensive distance; the problem that cross-domain association analysis is difficult to carry out due to the fact that the format difference of multi-source heterogeneous data generated by existing chip testing is large is solved, and the precision and efficiency of multi-source heterogeneous data processing are improved.

The invention discloses a multi-modal data driven dynamic pressure test scene generation method and system, and the method comprises the steps: carrying out the event extraction and matching of a user operation log and a full-link API call sequence, and carrying out the feature enhancement of each alignment event pair, and obtaining a local feature vector; statistical features are extracted from each user session, a graph attention network is constructed, and a group feature matrix is obtained; obtaining a global incidence matrix according to the mapping between the service load and the resource consumption; and splicing the user session feature matrix and the group feature matrix to generate group enhancement features, performing matrix multiplication on the group enhancement features and the global incidence matrix to obtain system-level risk features, splicing local feature vectors and the encoded user behavior logs, splicing the system-level risk features and the encoded performance indexes, and obtaining the user behavior log. And finally, carrying out multi-source fusion to generate a pressure measurement scene. The simulation precision, the dynamic adaptive capacity, the abnormal reproduction capacity, the resource utilization rate and the like are remarkably improved.

The invention provides an intelligent question and answer method and device based on domain knowledge lightweight fine tuning and a cross-domain dynamic knowledge base and a readable storage medium thereof.The method comprises the steps that the cross-domain dynamic knowledge base is constructed, and multi-domain heterogeneous data is processed through a grading and blocking strategy; dynamic expansion and incremental updating are realized through hierarchical vector indexing (bottom-layer knowledge block embedding + upper-layer field incidence matrix); a two-stage lightweight fine tuning strategy is designed, a high-rank adapter is used for roughly aligning a field semantic space in a pre-access stage, and a low-rank adapter is switched for optimizing a retrieval result in a post-access stage, so that the consumption of computing resources is remarkably reduced; and on the basis of a domain correlation reordering model, the retrieval result is optimized by fusing the original similarity, the term coverage and the historical matching degree, and the cross-domain knowledge correlation is improved. Through closed-loop cooperation of the knowledge base, the fine tuning model and reordering, the problems of poor migration ability, knowledge updating lagging, high customization cost and the like in the traditional system field are solved.

The invention discloses a hyperspectral remote sensing image red tide detection method and system, and the method comprises the steps: converting a hyperspectral remote sensing image into a hypergraph structure composed of vertexes and hyperedges, and representing the hypergraph structure through a hypergraph incidence matrix; inputting the vertex features and a hypergraph incidence matrix representing the relation between the vertexes and hyperedges into a hypergraph neural network, and training the hypergraph neural network by using a preset loss function so as to map the vertex features to a new feature space; and carrying out clustering on the vertex features in the new feature space, and identifying pixels or regions related to the red tide phenomenon in the hyperspectral remote sensing image according to a feature clustering result so as to realize detection of the red tide. According to the method, rich spectral information of a hyperspectral remote sensing image is utilized, and a label-free unsupervised red tide detection technology is constructed based on a hypergraph theory and a hypergraph neural network.

The invention relates to the technical field of multi-region data center management, and discloses a multi-region data center integrated management method and system, and the method comprises the steps: obtaining the physical position, business data and fault information of a data center, and constructing an adjacent matrix and an incidence matrix; obtaining a first resource group by adopting a clustering algorithm; constructing a resource availability model to obtain a resource demand, performing prediction by using a load prediction algorithm to obtain a resource prediction value, and if the resource prediction value exceeds a preset resource threshold value, triggering an allocation mechanism and obtaining an optimal scheme; calculating a fault isolation range according to the fault information, and if the isolation range contains an available data center, updating an association value to obtain an updated association matrix; optimizing by using a network topology optimization algorithm to obtain an optimization result; performing evaluation by using a preset delay evaluation model to obtain a delay influence degree and a second resource group; and generating a service deployment scheme according to the features. The method can improve the management efficiency of the multi-region data center.

The invention discloses a privacy protection effect evaluation method and system, and the method comprises the steps: carrying out the hierarchical analysis of a privacy protection effect evaluation index system, defining the definition and calculation boundaries of four first-level dimensions, i.e., data features, compliance, availability and safety, and subordinate second-level indexes, carrying out the adaptation of a calculation algorithm for different types of indexes, and carrying out the weighted integration of the second-level indexes through employing an analytic hierarchy process, thereby achieving the evaluation of the privacy protection effect. Multi-dimensional quantitative evaluation is realized; an index association rule base and a matrix are constructed through static analysis and dynamic mining, and linkage analysis is achieved; determining an index fluctuation reference according to scene characteristics and historical data, converting user demands into a variation amplitude of a fluctuation reference unit, and realizing index dynamic linkage updating based on an incidence matrix; and finally, the optimization rule is verified through simulation testing and historical data backtracking, and a dynamically adaptive and self-iterative optimization relation rule system is formed, so that the limitation of traditional evaluation is broken through, a scientific basis is provided for privacy protection strategy optimization, and the privacy governance efficiency is improved.

The invention provides a scheduling method, device and equipment of a wafer manufacturing system. The method comprises the following steps: constructing a time Petri net model by adopting a Petri net model according to a wafer processing route and constraint conditions; the constraint condition is the time delay added on the transition or library; an incidence matrix of the time Petri net model is calculated, and the incidence matrix is used for calculation to obtain the number of times of emission; the emission times are emission times corresponding to each transition from the initial identifier to the target identifier under the time Petri network model; operating the time Petri net model by using the launching times in combination with the shortest process time priority strategy, and processing the deadlock problem by adopting multi-step rollback in the operation process of the time Petri net model to obtain a scheduling result of the wafer manufacturing system; wherein the shortest process time priority strategy only reserves a preset number of enabling transitions under each identifier. And the generation efficiency of the scheduling result is improved on the basis of stable operation of the wafer manufacturing system.

The invention relates to the technical field of disease and pest detection, in particular to a disease and pest detection method and system based on a dynamic scene, and the method comprises the steps: obtaining a plurality of disease and pest initial images and disease and pest occlusion images in the dynamic scene, extracting an occlusion region for disease and pest recognition according to the disease and pest initial images and the disease and pest occlusion images, according to the feature information of the occlusion area, setting a form incidence matrix of the occlusion area; identifying feature interaction information of the pest initial image and the pest occlusion image according to the form incidence matrix of the occlusion area, and forming a self-adaptive interaction model based on the feature interaction information; determining a boundary distribution probability and a category probability of the initial image of the pest and disease damage by using a self-adaptive interaction model; according to the boundary distribution probability and category probability of the pest initial image, using an approximate matrix to carry out position adjustment, and delineating a pest prediction position; accuracy and precision of disease and pest detection in a dynamic scene are realized.

The invention discloses a real-time track fusion algorithm and system based on a radar and an AIS (Automatic Identification System) and application, and particularly relates to the technical field of shipping monitoring and management, which is used for carrying out batch elimination processing on targets subjected to batch elimination. And for the associated tracks, association judgment is performed on the associated radar tracks and AIS tracks in the latest time window in a manner of sliding the time window, and if the number of unassociated points exceeds a misassociation threshold, it is judged that the associated tracks are misassociated, and de-batching is performed on the associated tracks. The method comprises the following steps: calculating the correlation degree of a radar track to be correlated and an AI track to be correlated in pairs to obtain a correlation matrix, setting a coarse correlation time window, and performing correlation degree calculation on track points in the time window; and merging the regional associated tracks. And on the basis of an incidence matrix obtained by coarse correlation, through a KM algorithm, a radar and AI target combination with the highest correlation degree is screened out to serve as a correlation track.

The invention discloses a lightweight method, device and equipment for a large government affair model and a storage medium, and relates to the field of artificial intelligence model compression, and the method comprises the steps: determining metadata of the large government affair model, constructing a semantic incidence matrix based on a preset government affair field knowledge graph, and determining target information based on the metadata and the semantic incidence matrix; performing credibility evaluation on weight parameters of the government affair large model based on the target information, and determining a pruning strategy based on an obtained evaluation result; performing sensitivity analysis on the hierarchy of the government affair large model based on the target information and the weight parameter to obtain an analysis result, determining a quantization interval of the weight parameter, and determining a quantization strategy by using the analysis result and the quantization interval; and lightening the government affair large model by using the pruning strategy and the quantification strategy to obtain a corresponding target lightweight model. Therefore, the lightweight effect of the large government affair model can be improved.

The invention discloses an AI-based historical defective fresco digital restoration system, and relates to the technical field of AI restoration, a two-channel filter and a U-Net network are adopted to extract a fresco damage boundary, and edge detection is enhanced in combination with manual labeling; calculating a multi-scale Hurst index, and constructing a matrix quantization fractal feature change rule; dynamically generating a probabilistic grammar rule set based on the fractal incidence matrix, and defining a state transition probability function of texture growth; recursively executing texture generation driven by grammar rules from boundary points, and introducing random disturbance attenuating along with depth to simulate a stroke natural form; textures are generated through Laplacian pyramid layered correction, and the phase of a reference image is aligned in a Fourier domain to reserve structural consistency; and solving an affine transformation matrix with unchanged illumination in the transition zone, and globally optimizing color connection between the restoration area and the original wall painting. According to the method, the dynamic grammar rule base is established by analyzing the fractal features of the residual strokes, so that the generated texture follows creation logic, and the visual coherence of a recovery area is improved.

The invention provides a multi-domain power grid data collaborative modeling method and system based on a tensor game diagram, and the method comprises the following steps: firstly constructing a six-dimensional enhanced tensor model, constructing a six-order tensor based on the number of nodes, timestamps and other six dimensions, and obtaining a kernel tensor and a factor matrix through CPD-Tucker mixed decomposition; a joint optimization objective function containing reconstruction errors, game equilibrium and privacy risks is constructed, and an optimization kernel tensor and factor matrix is solved; and finally, inputting an optimization result into the MGGCN, processing double targets through a leader branch (a physical topology adjacency matrix guarantees reliability) and a follower branch (a market transaction incidence matrix optimizes economic cost), and generating a collaborative decision result through multi-head attention fusion. According to the method, the problems of low multi-source heterogeneous data fusion efficiency and difficulty in considering cross-domain collaborative privacy security and dynamic optimization are solved, the new energy output prediction error can be reduced, and the data utility loss caused by global encryption is reduced.