31 results about "Structure learning" patented technology

The application relates to an excitation system fault recording and event record analysis and diagnosis method and system, belonging to the field of excitation systems. The method comprises collecting recording files and event sequence records of the excitation system; performing multi-domain feature extraction based on the recording files to obtain a multi-domain feature tensor; performing space-time causal structure learning based on the multi-domain feature tensor and the event sequence records to output a causal adjacency matrix, a causal diagram and a time delay matrix; performing double-channel interpretable fault classification based on the multi-domain feature tensor, the causal diagram and an event time tag list E in the event sequence records to output a fault type label M and an attention space-time heat map; performing counterfactual causal tracing to obtain a root cause variable set and a causal propagation path, and outputting a diagnosis report. The application realizes intelligent diagnosis of the excitation system with signal analysis capability, causal reasoning capability and diagnosis interpretability.

The application discloses a source-load joint probability prediction method and system of a physically constrained graph attention network. The method collects multi-dimensional feature data of source-load nodes in a prediction area to construct an initial node feature matrix. A similarity matrix is generated through differentiable graph structure learning. A dynamic adjacency matrix is generated through normalization and introduction of a sparse mask. Spatial feature aggregation is performed through a multi-head graph attention network to obtain node spatial encoding. The node spatial-temporal hidden state is output through an encoder. The node spatial-temporal hidden state is input into a probability prediction head to output Gaussian distribution parameters of the source-load node power. A joint loss function is constructed. The joint loss function is used for soft constraint training to output a probability prediction result. Posterior projection hard constraint correction is performed in an inference stage to obtain a corrected prediction result. The application solves the problems of lack of physical consistency and inability to quantify uncertainty in the prior art.

The application discloses an aero-engine RUL prediction method based on MDGODE, which comprises the following steps: (1) collecting aero-engine performance degradation data and constructing RUL labels; (2) establishing an MDGODE model, including a multi-scale autocorrelation decomposition module, a multi-scale dynamic graph structure learning module, a space-time coupling graph neural ordinary differential equation module and a multi-scale attention fusion module; (3) training the MDGODE model by using the collected aero-engine performance degradation data; (4) extracting the comprehensive space-time representation of the aero-engine performance degradation data to be predicted by using the trained MDGODE, and inputting the comprehensive space-time representation into a regressor to predict the final remaining useful life value of the aero-engine. The MDGODE has strong space-time representation capability and robustness for the complex non-stationary performance degradation process of the aero-engine, and can significantly improve the RUL prediction accuracy of the aero-engine.

A storm surge intelligent forecasting method integrating causal structure and graph neural network, relating to the fields of artificial intelligence and marine storm surge forecasting technology, includes: Step 1. Multi-source data fusion; Step 2. Data preprocessing and standardization; Step 3. Using UNet-LSTM to achieve fusion modeling of storm surge meteorological spatiotemporal characteristics and water level temporal characteristics; Step 4. Causal structure inference and graph structure learning; Step 5. Using graph convolutional neural network to obtain a UNet-LSTM-DGCN combined model for final storm surge water level forecasting; Step 6. Combined model training and validation; Step 7. Visualization of prediction results; Step 8. Interpretability analysis. This invention constructs an intelligent method combining dynamic graph neural networks for storm surge forecasting, effectively capturing the dynamic causal relationships in water level changes at multiple stations under the influence of typhoons. While improving forecast accuracy, it reveals the decision-making basis within the model through visualized graph structures.

The application discloses a kind of track vehicle running gear anomaly detection methods based on symbolic regression and generative adversarial network, comprising the following steps: collecting the sensing data of track vehicle running gear, and save in database;Establish structural learning model and generative adversarial network model, the result of structural learning model and generative adversarial network model is superimposed as health baseline data;Obtain the sensing data of track vehicle running gear in target period as real-time monitoring data, calculate real-time deviation according to real-time monitoring data and corresponding health baseline data;Calculate the average error of all real-time deviations in target period, if average error is greater than preset alarm threshold, it is judged to appear abnormality.The application can accurately analyze the mechanism of the structural characteristics of track vehicle running gear, realize the dynamic real-time tracking of anomaly detection.

The application relates to the technical field of dialogue sentiment computing and natural language processing, and particularly discloses a dialogue sentiment-reason pair extraction method based on a double-channel graph encoder network. The method comprises the following steps: preprocessing an input dialogue to construct a dialogue directed graph; inputting dialogue features and an adjacency matrix into a global structure learning channel and a local feature extraction channel in parallel, and modeling long-term dependence and local implicit clues respectively; integrating double-channel features through an adaptive gating fusion module; calculating matching scores of all candidate sentiment-reason dialogue pairs based on the integrated features, and outputting final sentiment-reason pairs.

The application discloses a water channel pipe network robot control method based on reinforcement learning, relates to the technical field of robot intelligent control, and comprises the following steps: collecting multi-source data of a water network pipeline and preprocessing the multi-source data to generate basic observation data; based on robot motion state data and driving feedback data, a fluid disturbance prediction model is constructed to inversely obtain fluid characteristic disturbance information by fluid action, and the stability state of the robot is constructed in combination with robot posture data; the basic observation data, the fluid disturbance characteristic information and the stability state are adaptively fused to construct an enhanced environment state, the pipe structure features are identified based on the enhanced environment state, the pipe network topology memory structure is updated in a graph structure learning mode, and path guide information is generated. Through the multi-target collaborative analysis of the reinforcement learning control driven based on the enhanced environment state, the control stability and adaptability of the robot in the complex pipe network environment are effectively improved.

The application belongs to the field of fault prediction, and discloses a server fault prediction system based on deep learning, which comprises a multi-scale time sequence feature extraction module, receives multi-dimensional monitoring index time sequence data of each node of a server cluster, extracts short, medium and long time scale time sequence features through a plurality of parallel causal convolution layers with different expansion rates, a dynamic graph structure learning module, learns an implicit dependency relationship of a node from monitoring data based on an attention mechanism, and generates a synchronously evolving dynamic adjacency matrix, a cross-scale attention fusion module, aligns and fuses different scale time sequence features and dynamic graph structure features by using a cross-attention mechanism, and generates a fusion feature representation, and a fault prediction module, which generates a fault prediction result of each node based on the fusion feature. Compared with the prior art, the system solves the defect that multi-scale time sequence and dynamic graph structure deep coupling cannot be captured simultaneously through multi-module cooperation, and improves the prediction accuracy.

This invention provides a reinforcement learning-based intelligent adversarial system for game NPCs, belonging to the field of game adversarial systems. The system includes a data acquisition module that collects all raw interaction data from the target game engine and generates standardized trajectory data packets; a feature transformation module that uses a preset computational model for feature extraction and semantic encoding, and uses an adaptive segmentation algorithm to segment tactical fragments and mark target decision points; a causal path acquisition module that constructs a temporal causal discovery dataset, executes a preset temporal causal structure learning algorithm, generates a domain knowledge-enhanced causal graph, and acquires the target causal path; and a counterfactual inference module that proposes counterfactual intervention hypotheses, performs structured counterfactual inference using a preset inference method, and generates causal comparison training samples. This invention improves the information density of learning samples and the generalization ability of strategies, enriches game content, extends the game lifecycle, and directionally improves the intelligence level of NPCs without human intervention.

The application discloses a kind of underground cavern group complex construction ventilation environment space-time prediction method based on deep learning, including data acquisition, data cleaning, data noise reduction and normalization processing: construct and include sequence perception global Token generation module, graph structure learning module, graph aggregation module, time coding module and trend perception attention module's space-time prediction model: normalized data is divided into training set, verification set and test set, constructs joint loss function, utilizes optimization algorithm to train the space-time prediction model, and adopts swarm intelligence optimization algorithm to automatically optimize hyperparameter;Real-time monitoring data is input after pre-processing into trained space-time prediction model, and the wind speed, dust concentration prediction value of future time step is output, and the operating frequency of fan of ventilation system is dynamically adjusted according to prediction value.

The present application relates to a kind of gas stove operation chain risk prediction method and system based on causal inference, belong to gas risk prediction field.Therein, the method includes collecting original data and forming multivariate time series;Time series causal structure learning is carried out based on multivariate time series, and time-delayed time series causal directed acyclic graph is output;Causal effect quantification is carried out based on multivariate time series and time series causal directed acyclic graph, and the causal effect function of each edge is obtained, and structural causal model is output;Counterfactual path deduction is carried out based on multivariate time series, time series causal directed acyclic graph and structural causal model, and a group of possible future causal paths are obtained, and each path corresponds a probability value;Chain risk prediction is carried out based on future causal path and its probability value, and early warning information is obtained.The present application provides a kind of gas stove safety warning method capable of understanding risk transmission mechanism, identifying key causal path, providing interpretable early warning.

The application relates to the technical field of computer application, and provides a Chinese character virtual learning method and device combining VR and mortise and tenon. The method obtains user input information through a VR device to determine a target Chinese character, then carries out skeleton extraction and topological analysis on the target Chinese character to obtain a structure unit set of the target Chinese character, matches a target mortise and tenon template based on connection type information in the structure unit set, determines mortise and tenon geometric parameters based on stroke information in the structure unit set, and finally obtains user splicing information in real time through the VR device to determine the user's operation until the user completes the splicing of the target Chinese character. The method realizes automatic mapping of Chinese character structure and three-dimensional mortise and tenon components, improves the learning experience of Chinese character form in a three-dimensional space, improves the interaction efficiency and recognition accuracy in the Chinese character structure learning process through multi-modal information cooperation.

This invention discloses a method for constructing a scientific graph based on multi-level knowledge structure recognition, comprising: collecting multi-source heterogeneous data from preset aspects; preprocessing the multi-source heterogeneous data; dividing the multi-source heterogeneous data into multi-level knowledge structures to obtain hierarchical knowledge data; constructing a hierarchical weight allocation mechanism based on knowledge transfer entropy and hierarchical contribution using the hierarchical knowledge data; obtaining a cross-level association mechanism based on the hierarchical knowledge data; constructing a multi-level knowledge structure recognition scientific graph model based on a multi-level graph neural network based on the hierarchical knowledge data, the hierarchical weight allocation mechanism, and the cross-level association mechanism; optimizing the multi-level knowledge structure recognition scientific graph model using dynamic topology learning; inputting the data to be constructed into the multi-level knowledge structure recognition scientific graph model; and outputting the target scientific graph.

The application discloses a kind of multi-view clustering methods based on global and local anchor point learning, belong to data processing technical field, including the following steps: S1, generate target function for all views;S2, based on target function, anchor graph structure learning and global consistency structure learning are carried out, and double-layer anchor graph optimization target function is constructed;S3, based on double-layer anchor graph optimization target function, global and local anchor point learning multi-view clustering model is constructed using mapping matrix;S4, global and local anchor point learning multi-view clustering model is solved, and classification result is obtained.The application learns the local anchor point of specific view under the guidance of cross-view global anchor point selection, so as to effectively capture the local information of specific view and the cross-view global consistency of multi-view data.

The present application relates to a kind of power transmission network impedance envelope fast prediction method based on graph structure learning and its electronic equipment, for the fast performance evaluation under complex environment, method includes: first, according to standard unit cell theory, PDN is discretized into Unit cell, and training and test data set containing different environment and process parameters are constructed by EDA tool simulation.First, pre-processing is carried out to data, relevant parameters are extracted from PDN netlist file, and they are abstracted as graph structure: node represents Unit cell, includes L, C, position and five attributes of distance from port distance;Edge indicates connection relationship and contains R attribute.Subsequently, graph data is normalized and converted into node feature vector, and graph neural network model is constructed, with the second upper envelope line of impedance-frequency curve as label, graph level regression training is carried out using training set.Finally, the model is verified using test set, to realize the fast, high-precision prediction of PDN impedance under unknown parameters.The method has good generalization and universality.

This invention proposes an interpretable respiratory event detection method based on prior-guided graph structure learning. The method includes: preprocessing and segmenting multimodal physiological signals to obtain time-segment sequences; performing segment-level graph learning and feature extraction on the time-segment sequences to obtain segment-level global representations and segment adjacency matrices; using the segment-level global representations for global-level learning and long-range dependency extraction to obtain global-level global representations and global adjacency matrices; constructing an objective function using the segment adjacency matrices and global adjacency matrices; updating the detection model using the objective function to obtain an updated detection model; and obtaining prediction results using the updated detection model. This invention effectively filters random noise and spurious associations in the data by introducing a graph structure learning mechanism guided by clinical prior knowledge, utilizing the continuity of physiological signals or specific coupling relationships between modalities as constraints.

The application discloses a kind of based on CIM standard's power grid multidimensional net loss fine management method, system, equipment and medium, belong to distribution network net loss management field, the method of the present application includes: generating power grid digital twin topological structure;Quantitative measurement data are carried out space-time alignment preprocessing, and dataset is constructed;Theoretical line loss estimation value and equipment loss contribution degree sorting of each branch are generated by using graph structure learning network to carry out deep feature cross fusion;Line loss estimation value is input into multidimensional attribution analysis engine, and equipment health characteristics, load composition characteristics, topological structure characteristics are fused, main cause category is located and loss reduction strategy is generated;It is transformed into control instruction and is executed, and after the execution, the measured data are collected to calculate loss reduction effect, and the result is fed back to the feature extraction link, and the feature weight distribution is corrected.The management accuracy and adaptability of the system are improved persistently when power grid structure and load characteristics change, and the evolution of distribution network net loss management to intelligent, precision and self-direction is promoted.

The application discloses a kind of aviation maintenance risk assessment and intelligent diagnosis method based on multidimensional bayesian network, belong to aviation maintenance safety and data analysis technical field.The existing risk analysis method is aimed at solving the problem that there is dependence on expert subjective experience, it is difficult to represent the nonlinear coupling relationship of multidimensional factor, cannot effectively extract unstructured text risk information, lack of dynamic early warning and accident cause accurate tracing ability.Proceeding with the pretreatment and inspection of maintenance text;Cascade network node of four dimensions of man, machine, environment and management is constructed;Risk Boolean matrix is generated using keyword matching;Under the constraint of forbidden edge, structure learning is carried out, and parameter estimation is executed;After using AUC to evaluate model, forward risk prediction and cause reverse diagnosis are executed.It can realize the mapping from unstructured text to quantitative causal model, effectively represent complex nonlinear coupling relationship, realize dynamic early warning and accurate tracing, improve the objectivity and reliability of aviation maintenance safety management.

The embodiment of the application provides a method for determining a causal relationship diagram, a method for determining a fault solution, equipment and a medium, the method comprises: taking production parameter information and fault state information of a product sample in a production process as learning data, performing structural learning on production data of the product sample by using a directed acyclic graph under the constraint condition of the directed acyclic graph, obtaining a graph structure learning result of the directed acyclic graph, if a user evaluation result of the graph structure learning result is learning end, optimizing an edge weight of the graph structure learning result, obtaining a causal relationship diagram, and then using the causal relationship diagram to analyze a solution of a target fault. The technical solution can improve fault analysis accuracy and analysis efficiency of a product to be tested such as a circuit board, and improve effectiveness of a fault solution.

This invention implements a unified multimodal information extraction method for social media platforms. Inputting text and image information from the social media platform, the method performs entity and relation extraction for both text and images through a multimodal feature fusion module and a multimodal information intelligent extraction module. The unified information extraction module uses intelligent decision-making to identify and classify entities in the data, employs reinforcement learning-based intelligent decision-making to determine the execution order of extraction tasks, and uses Q-learning to perform multimodal information extraction and intelligent task selection to find the optimal task execution order. This invention fully utilizes the diversity of information modalities on social media, proposing to construct a heterogeneous graph from information of different modalities. Graph structure learning and causal intervention are used to optimize the graph structure and graph neural network, respectively, thereby obtaining text and image representations with richer semantics.

The application discloses a kind of graph neural network and malicious traffic monitoring method and system of adaptive concept drift, it is related to internet, network security technology, including: the network data packet of original network traffic data collected is aggregated as session flow, and network traffic features are extracted;Any session is regarded as a node of network traffic graph, and network traffic graph is constructed;Using graph convolution operator, for each session node, the features of its associated node are aggregated to perform graph structure learning;Using independently initialized learnable weight matrix, the result of graph structure learning is enhanced;Based on the result of graph structure learning and enhanced features, train adversarial boundary generator, evidence classifier;For any input traffic sample, use the trained evidence classifier to monitor malicious traffic.The application constructs the traffic monitoring method capable of responding to encryption, collaboration, dynamic network threats.

The expressway OD and path flow prediction method, electronic equipment and storage medium belong to the technical field of expressway intelligent prediction management. In order to solve the problem of real-time prediction of expressway OD flow and path flow, the present application constructs a space-time convolution model, including a graph structure learning unit, a time convolution unit, a space convolution unit and an output unit, which is used for predicting the time series of expressway entrance flow and outputting the predicted entrance flow; a hierarchical flow model is constructed; the predicted entrance flow is mapped to the OD matrix and the path matrix through two fully connected matrices in turn, and the reverse propagation mechanism of the neural network is used for predicting the expressway OD flow and the path flow; then, model training and model prediction are carried out. The present application realizes an integrated model for predicting entrance flow, OD flow and path flow, and the hierarchical flow model solves the problem of OD and path flow prediction under the condition of being unable to obtain real-time OD and path.

The application provides a cross-level time series prediction method and computer device based on multi-source digital footprint, which is used for modeling and inferring the data asset value trend and risk interval for future period, and forming forward-looking regulation under budget constraint. It includes multi-source footprint collection and structure learning; hierarchical state modeling and parameter self-adaptation; integrated time series prediction and uncertainty quantification; counterfactual correction and online re-calibration; rolling window resource and strategy linkage optimization. The application is suitable for data asset predictive management, low-carbon energy efficiency regulation and automated decision-making scenarios in complex computing environments.

The application discloses a scene graph robust representation learning method based on uncertainty guidance causal intervention. For a target image sequence, a trained scene graph robust representation model is used to obtain a risk prediction result, wherein the scene graph robust representation model comprises: a target feature construction module for converting the target image sequence into a structured entity feature representation; a variational scene graph generation module for performing entity layer variational modeling and relationship layer variational modeling on the structured entity feature representation to obtain a probabilistic scene graph; an uncertainty perception causal intervention module for performing soft intervention correction on an environment-related bias in a representation space to obtain a corrected relationship representation; a differentiable structure learning module for constructing a sparse scene graph through a sparsification strategy; and a time sequence aggregation and prediction module for obtaining a risk prediction result through graph-level feature aggregation and cyclic time sequence modeling based on the sparse scene graph. The application can improve the reliability of an automatic driving scene understanding system in a complex environment.

This invention relates to an underwater path planning method based on data fusion, belonging to the field of intelligent inspection technology for special equipment. The method uses a multi-source heterogeneous sensor array to collect real-time hull status data and construct a comprehensive environmental dataset. The dataset is then differentiated, divided into subsets based on material, deposit density, and water flow disturbance characteristics, and compensation factors are generated. These subsets and compensation factors are input into a graph structure learning network path planning model, outputting cleaning paths and equipment action commands. A multi-dimensional attribution analysis engine is used to locate the main causes of cleaning efficiency deviations and generate compensation strategies. After the strategies are converted into control commands and executed, the collected effect data is fed back to the differentiation processing stage to correct the subset division and compensation factors, forming a continuous evolutionary closed loop. This application improves the targeting and operational efficiency of path planning, reduces energy consumption and repetitive operation rates, promotes the transformation of underwater inspection towards intelligence and adaptability, and thus improves the accuracy of underwater inspection of special equipment.

The application discloses a dynamic graph anomaly detection method and system fusing space-time feature enhancement and potential structure mining, and the method comprises the following steps: acquiring a dynamic graph data stream to be detected, and dividing the dynamic graph data stream into a plurality of discrete time snapshots based on timestamps; learning potential correlations between nodes based on node features of a current time snapshot, and generating a sparse potential graph structure; fusing an original graph structure and the potential graph structure to generate an enhanced graph structure; constructing a vertex-edge transition matrix, and iteratively performing bidirectional feature propagation and updating of vertices to edges and edges to vertices on the enhanced graph structure by using the vertex-edge transition matrix to obtain updated node features; and inputting the updated node features into an anomaly detection model, calculating anomaly probabilities of each node, and outputting a detection result. The application mines potential correlations by learning dynamic graph structures, reduces computational complexity by using a vertex-edge interaction mechanism, and improves the accuracy and robustness of anomaly detection.

This invention provides a graph data processing method based on graph neural networks, aiming to solve the problems of excessive smoothness, difficulty in capturing long-range dependencies, and sensitivity to noise in the original graph structure of existing graph neural networks. Its core includes: multimodal feature extraction and fusion of the original graph data to form enhanced feature representations of nodes; constructing an innovative graph neural network model, which sequentially includes a bidirectional information propagation layer, a hierarchical attention fusion layer, and an adaptive graph structure learning layer; using this model for information processing, where the bidirectional propagation layer simultaneously performs message passing from node to neighbor and from neighbor to node, the attention fusion layer dynamically integrates features at different depths of the network, and the structure learning layer adaptively optimizes the graph topology; finally, the model is trained through a multi-task joint learning framework to complete downstream tasks. This invention significantly improves the accuracy and robustness of graph data processing tasks.

The present invention relates to an artificial intelligence-based graph structure learning method and system, and a graph-based data processing method and system using same, and provides an artificial intelligence-based graph structure learning method and system, and a graph-based data processing method and system using same, which are capable of more effectively analyzing and learning graph data.

The application relates to a method for determining an anomaly based on a graph attention mechanism, comprising: acquiring multivariate time series data, wherein the time series data is data generated by a plurality of sensors of a device in a semiconductor manufacturing process; acquiring final embedding vectors corresponding to the plurality of sensors according to a graph structure learning model, wherein the final embedding vectors are used to represent the association relationship between the plurality of sensors; inputting a target subsequence in the time series data in a time sequence into a time series feature extraction model to obtain a time series feature vector; splicing the time series feature vector and the final embedding vector and inputting the spliced result into a prediction model to obtain predicted data corresponding to the plurality of sensors; and determining whether an anomaly occurs at a target time point corresponding to measured data according to the predicted data and the measured data following the target subsequence in the time series data.

The application discloses a kind of knowledge graph guided diffusion model anomaly detection method and system, it is related to data anomaly detection technical field, comprising: obtaining multi-source heterogeneous time series data is constructed as multi-modal time series graph representation, and based on regional knowledge graph constructs knowledge graph embedding vector embedded field knowledge;Condition diffusion model is constructed, and multi-modal time series graph data under normal state is used for training, with knowledge graph embedding vector as conditional input, guide reverse denoising process recovers original graph structure from noise graph, learns the distribution of normal data under knowledge guidance;The data reconstruction of the multi-modal time series graph to be detected is carried out, the difference between original graph and reconstructed graph is calculated, and multi-scale anomaly scoring mechanism is used to generate anomaly score, and the area exceeding threshold is output as anomaly detection result.The application can effectively fuse multi-modal information and field knowledge, realize accurate positioning and explainability detection of anomaly.