147 results about "Relationship mining" patented technology

The invention relates to the technical field of multi-modal data processing, and discloses a multi-modal threat perception method and system based on a space-time diagram neural network, and the method comprises the steps: obtaining a multi-modal original data set in a vehicle insurance claim settlement link, and carrying out the business relation mining and space-time dynamic analysis, and obtaining an entity space-time relation diagram; inputting the entity space-time relation graph into a space-time graph neural network for space-time fusion to obtain a node threat embedding vector; performing graph contrast learning and cross-modal feature discrimination on the node threat embedding vector to obtain a vehicle insurance threat feature vector; and carrying out fraud space-time propagation modeling based on the vehicle insurance threat feature vector, and generating a vehicle insurance threat blocking strategy, the method can accurately predict a propagation path and an influence boundary of gang fraud in a vehicle insurance ecological network, and identifies potential threats and starts prevention measures before a fraud behavior is completely displayed.

The invention relates to the technical field of document intelligent analysis, and discloses a document content self-adaptive analysis method and system based on a large model. The method comprises the following steps: acquiring original data flow of a to-be-analyzed document, wherein the original data flow comprises a text coding sequence, a layout structure mark and a multimedia embedding feature; the data stream is input into a pre-trained multi-modal large model, and a document semantic graph structure, a concept-containing node set, a relation edge weight matrix and a cross-modal alignment index are generated through context sensing analysis; performing dynamic hierarchical clustering on the semantic graph structure to obtain a hierarchical topic tree containing core topic branches, secondary topic branches and leaf node association strength; extracting a document logic framework containing chapter division suggestions, key information positioning coordinates and a cross reference mapping table according to the topic tree; a result is generated based on an adaptive analysis strategy optimization framework, and the strategy adjusts clustering granularity and relation mining depth according to document type features.

The invention discloses a model intelligent verification and parameter correction method, which comprises the following steps of 1, building an unmanned ship task high-value verification point set based on expert experience and a large model collaboratively, screening boundary points and extreme points, and performing simulation-actual measurement data consistency verification; if the normalization error of the simulation data and the actual measurement data exceeds a threshold value, triggering correction; step 2, constructing a sensitive factor set according to global sensitivity analysis and a parameter association relationship mining result, screening high-sensitivity parameters as a priority correction target, and avoiding redundancy optimization; and step 3, optimizing the high-sensitivity parameters, and forming a verification-correction-update closed loop by verifying and iteratively updating the priority of the factors to realize the intelligent correction of the parameters of the unmanned ship model.

The invention discloses a multi-modal data association relationship mining method based on potential feature perception, which belongs to the field of multi-modal data analysis and feature association modeling in artificial intelligence and data mining technologies, and comprises the steps of multi-modal data acquisition and preprocessing, multi-modal feature mining based on potential semantic alignment, multi-modal data analysis and feature association modeling. Performing multi-stage feature fusion and time sequence association representation learning, and constructing a cross-modal semantic association graph. According to the method, under the conditions of noise interference, unbalanced sample distribution and weak semantic association of the multi-modal data, robust fusion and semantic consistency expression of the multi-modal features in a potential space can be realized through adaptive anomaly correction and a multi-level feature alignment mechanism, mismatching caused by noise pollution and shallow association is avoided, and the robustness of the multi-modal features is improved. And accurate mining of the high-order potential semantic relationship is realized. Meanwhile, the semantic edge and the time sequence edge can be subjected to separation modeling according to the internal structure of the multi-modal data under the conditions of modal isomerism and time sequence overlapping, and meanwhile, a unified cross-modal association graph is constructed. Furthermore, in order to improve the accuracy of time sequence relation modeling, time sequence comparative learning and dynamic consistency constraint are utilized, effective distinguishing between real time sequence dependence and multi-mode repeated representation is achieved, and the precision and robustness of multi-mode correlation analysis are remarkably improved.

The invention discloses a science and technology big data service system and method for promoting transformation of science and technology achievements, and relates to the technical field of science and technology services and big data, and the system comprises a data collection and integration module, a data analysis and mining module, an intelligent matching and recommendation module, and a transformation service support module. The data acquisition and integration module is used for acquiring data from multiple data sources and cleaning, preprocessing, integrating and storing the data; and the data analysis and mining module is used for carrying out feature extraction, market demand analysis and association relationship mining on the data. According to the science and technology big data service system and method for promoting the transformation of the science and technology achievements, the science and technology achievements and related information can be comprehensively and accurately collected through the multi-source data collection and integration module, the problem that the information is dispersed and incomplete is solved, and a solid data basis is provided for subsequent analysis and service. And the data cleaning and preprocessing unit ensures the data quality and improves the accuracy and reliability of data analysis.

The invention discloses a course recommendation method and system based on learner multi-behavior relation mining. The method comprises the following steps: firstly, constructing a learner multi-behavior heterogeneous graph; extracting a plurality of single-behavior sub-graphs from the multi-behavior heterogeneous graph of the learner; based on each single behavior sub-graph, fusing the embedded representation of the behavior into message transmission of GCN, and learning the embedded representation of the learner node and the embedded representation of the course node to obtain the embedded representation of the learner node and the embedded representation of the course node under each behavior; then, performing multi-behavior generality fusion on the embedded representations of the learner node and the course node based on the meta-path, and further performing multi-behavior generality enhancement on the embedded representations of the learner node and the course node to obtain the enhanced embedded representations of the learner node and the course node; and finally, calculating a correlation score of the learner-course pair, and recommending a course to the learner according to the correlation score. The behaviors are integrated into the embedded representation learning of the learner and the course nodes, so that the recommendation accuracy and the user satisfaction are remarkably improved.

The invention provides an intelligent Internet of Things safety monitoring method and system, and the method comprises the steps: obtaining an equipment operation record and a behavior operation record, carrying out the event serialization coding of the equipment operation record, obtaining a sensor event sequence vector, carrying out the behavior mode vectorization processing of the behavior operation record, obtaining a behavior event sequence vector, and carrying out the monitoring of the behavior event sequence vector. Inputting the sensor event sequence vector and the behavior event sequence vector into a pre-trained association analysis model, carrying out cross-sequence causal relationship mining processing, generating an abnormal event association graph, and based on event node attributes and association edge weight values in the abnormal event association graph, executing risk conduction path analysis to obtain a risk conduction path; and obtaining a key conduction path set and a risk accumulation intensity value of the house safety risk, and generating a safety monitoring result containing the risk level identifier and the risk source positioning information according to the key conduction path set and the risk accumulation intensity value. According to the invention, the accuracy and comprehensiveness of intelligent internet-of-things safety monitoring are effectively improved.

The invention provides a project contract construction period management method and system based on causal relationship modeling, and the method comprises the steps: firstly obtaining a construction period element set which comprises a contract agreed process execution sequence, a resource configuration scheme and historical similar project construction period influence records; the method comprises the steps of obtaining a causal dependency relationship chain among process nodes, generating a dynamic construction period intervention scheme based on the causal dependency relationship chain and process execution state data collected in real time, updating a reference construction period plan of a project contract according to the dynamic construction period intervention scheme, and obtaining an adjusted construction period plan. And comparing and verifying the adjusted construction period plan with the contract construction period constraint condition, and generating a construction period management report, thereby improving scientificity, accuracy and flexibility of project contract construction period management, and reducing construction period delay risk.

The invention discloses a smart farm knowledge warehouse establishment method and system. The smart farm knowledge warehouse establishment method comprises the steps of obtaining farm basic data and an agricultural field ontology library; preprocessing the farm basic data to obtain preprocessed data; acquiring data with traceability information according to the preprocessed data; generating a semantic enhanced association rule base; and generating a multi-modal agricultural knowledge representation library, wherein the multi-modal agricultural knowledge representation library comprises a feature layer, a semantic layer and a rule layer. According to the semantic enhanced association rule base constructed by the method, an agricultural field ontology base and a machine learning algorithm are fused, and rule dynamic updating and implicit relation mining are realized. The mechanism can automatically adapt to climatic change, variety improvement and other scenes, for example, irrigation rules are adjusted based on real-time environment data, and decision accuracy is improved.

The invention belongs to the technical field of artificial intelligence, and particularly relates to an AI-based file data content accurate and deep analysis and interpretation method, which comprises the following steps: acquiring multi-format file data and file meta-information, constructing an AI analysis network, extracting text semantic vectors, image visual features, table structure information and document layout features, and constructing a multi-dimensional semantic map. According to a user query intention, semantic extension is performed in combination with a domain knowledge base, an enhanced semantic description vector is generated through a graph attention mechanism, semantic reasoning and relation mining are performed by adopting an improved knowledge distillation Transform model, and a deep analysis conclusion is generated through a multi-hop reasoning model in combination with file data complexity, information density and user requirements. And generating a personalized interpretation report in combination with a user role and a task scene, and outputting an analysis result through a visual interface. Therefore, the problems of poor understanding ability, poor file adaptability and the like in the prior art are solved.

The new energy ship fault causal relationship construction method provided by the invention comprises the following steps: when a fault occurs, fusing multi-modal heterogeneous data and screening to obtain a core feature set; obtaining an environment invariant feature matrix through invariant risk minimization learning, evaluating causal edge strength among variable features based on mutual information to obtain a causal edge strength matrix, and mining a core fault causal skeleton in combination with a causal edge strength threshold and a condition independence test; based on the skeleton and the hierarchical node system, constructing an initial hierarchical fault causal graph, fusing a causal edge strength matrix to determine an initial causal edge weight, and dynamically updating by using a meta-learning model to obtain a target hierarchical fault causal graph; and after the target hierarchical fault causal graph is corrected through an anti-fact sample, a reasoning path is optimized, and then the fault causal relationship of the new energy ship is obtained through structured reasoning and combined with a large language model to mine an implicit causal relationship. Therefore, stable and self-adaptive fault causal relationship mining under a complex dynamic working condition is realized.

The invention provides an enterprise cooperation relationship mining model training method and a friend recommendation method and system, and the method comprises the steps: firstly integrating multi-source data, constructing a structured enterprise label through optical character recognition and a natural language processing technology, and constructing an enterprise relationship graph with an enterprise as a node and a plurality of business relationships as edges on the basis of the structured enterprise label; then, node pairs with business exchange are selected from the atlas as positive samples, node pairs which do not exchange but meet specific conditions are selected as negative samples, and a training set is formed; a model based on an inductive graph neural network encoder, a multi-relation graph attention layer, a cross-relation fusion layer and a link prediction layer is adopted for training, and parameters are optimized through a marginal contrast loss function. During application, a trained model is utilized to calculate the cooperation probability between a target enterprise and an unknown enterprise, and potential cooperation partners are recommended to a user through instant messaging according to the cooperation probability. According to the invention, the efficiency and the intelligent level of business expansion of enterprises are obviously improved.

The invention discloses a relay protection defect diagnosis method and device based on a dynamic knowledge graph, and belongs to the field of relay protection defect diagnos.The method comprises the steps that a first data set in an electric power system is collected in real time, standardization processing is conducted on the first data set, and a second data set is generated; the second data set is input into a MacBERT-BiLSTM-CRF model for entity extraction, and a plurality of entities are obtained; updating graph structure data according to the second data set and each entity, and inputting the graph structure data into a GNN model for relationship mining to obtain a plurality of entity relationships; updating a knowledge graph according to each entity and each entity relationship, performing rule reasoning based on the knowledge graph and real-time data of the target relay protection device, and performing probabilistic reasoning through a Bayesian network after a first diagnosis result is obtained to obtain a second diagnosis result; based on the second diagnosis result and the real-time data, grey correlation degree analysis is conducted, a consistency evaluation result is obtained, and when the consistency evaluation result is larger than a first threshold value, a diagnosis report is generated and output.

The invention belongs to the technical field of wind power plant fire assessment, and discloses a multi-modal fusion wind power plant fire hazard multi-source data space-time synchronous assessment method and system, and the method is characterized in that a reference calibration module builds a digital twin simulation, double closed-loop dynamic calibration and credibility quantification three-in-one mechanism; the time synchronization adopts a triple strategy of GPS time service, local clock compensation and transmission delay prediction, and the transmission delay is compensated in advance in combination with an LSTM network; the space calibration depends on a three-dimensional digital twin model of the fan, and mounting deviation and vibration drift are corrected through visual identification and coordinate matching; the quality grading module is used for constructing a three-dimensional quality model, distributing weights according to data quality grading, and reducing evaluation deviation caused by data heterogeneity; the feature fusion module adopts a spatial-temporal feature, modal feature and quality weight cross fusion mechanism; and capturing data time sequence association through an overlapped time window, and mining spatial association in combination with a digital twin spatial topological relation.

The present application relates to the technical field of complex equipment component repair, in particular to a device component performance and unit body maintenance level relationship mining method capable of effectively improving the use efficiency of an aero-engine, which first carries out expansion processing of repair samples, and then selects a support vector machine regression method which is better in the condition of small sample problems to solve the mapping relationship between the component performance before repair, unit body maintenance level and component performance after repair. Since a component is generally composed of multiple unit bodies, each component has multiple maintenance levels, and the mapping relationship between the component performance before repair, unit body maintenance level and component performance after repair is a many-to-one mapping relationship. In order to improve the accuracy of support vector machine regression, a hybrid kernel function method is used to optimize it, and a particle swarm algorithm is used to optimize the related parameters.

The invention relates to an alarm information processing method and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: receiving a plurality of original alarm events of a target system; aggregating the plurality of original alarm events based on the alarm time to obtain an aggregated alarm event; performing alarm text semantic analysis on each aggregated alarm event to obtain an alarm object related to the aggregated alarm event and at least one alarm index item corresponding to the alarm object; obtaining historical value data of each alarm index item, and performing association relationship mining on a change condition of the historical value data to obtain an index value association relationship between the alarm index items; determining an alarm causal relationship between the aggregated alarm events according to the index value association relationship; and merging the aggregated alarm events according to the alarm causal relationship to obtain a target alarm event corresponding to the target system. By adopting the method, the alarm analysis efficiency and accuracy can be improved.

The application relates to the field of knowledge graphs, and discloses a method and system for generating a resource map of a person's network based on a relationship graph. Multi-source interaction behavior data is collected to construct a heterogeneous graph structure, and a graph neural network is used to jointly encode the interaction frequency, time decay factor and spatial trajectory co-occurrence degree to calculate the weight of a strong implicit relationship. A relationship graph is generated based on the weight of the strong implicit relationship, and a space-time dimension index is embedded in the database structure. The space-time reachable probability of a resource node is calculated based on the combination of the geographical topological constraint, the weight of the strong implicit relationship and the space-time dimension index. The resource nodes are visualized and rendered based on the space-time reachable probability to generate a resource map of a person's network. The application overcomes the defect that a static graph cannot represent dynamic relationships, improves the accuracy of strong implicit relationship mining, realizes quantitative calculation of resource space-time reachability, and reduces the redundancy of graph search traversal.

The application discloses a data blood relationship mining method and device of an advanced enterprise application programming language, and the data blood relationship mining method comprises the following steps: performing word segmentation processing on the code of the advanced enterprise application programming language to obtain a mark list; wherein the word segmentation processing is performed in the minimum unit of words and symbols; the mark list is parsed according to the grammar rules of the Backus Normal Form to obtain a semantic expression list, and the semantic expression list comprises grammar keywords and grammar contents; the semantic expression list is formatted according to the grammar rules of the Backus Normal Form to obtain a semantic expression dictionary list; the semantic expression dictionary list is parsed, nodes are created, the blood relationship between different nodes is found and saved. The ABAP program code is subjected to fine-grained word segmentation processing, and semantic analysis is performed on the basis of the fine-grained word segmentation processing, so that the association relationship between data is finally obtained, and the blood relationship between nodes is determined.

The invention discloses a mail fusion governance device with an AI identification technology, and particularly relates to the technical field of mail governance, and the mail fusion governance device comprises a certification data receiving module, a multi-mode intelligent analysis engine, a mail account relation modeling module, a customs data collision engine and an Excel result output module. According to the method, intelligent structured conversion of multi-format attachments is realized by constructing a mail governance architecture fusing AI multi-modal identification and relation mining, and unstructured service data scattered in the attachments are automatically extracted and mapped into standard database fields by utilizing cooperative work of a large language model and a multi-modal model, so that the mail governance efficiency is improved. Establishing a dynamic mail entity relation graph, and generating an association network based on mail contact characteristics to strengthen the business traceability; a cross-source data intelligent collision engine is designed, structured mail data is compared with third-party systems such as customs declaration in real time, risk clues or business evidence closed loops are automatically output, and the time cost of manual checking is greatly reduced.

The disclosure provides a generalized small sample segmentation method and system for class relationship mining, relating to the technical field of image segmentation, which performs generalized small sample segmentation on remote sensing images through a two-stage trained segmentation model; in the two-stage training, class relationship mining is performed using background consistency modeling technology and inter-class relationship mining technology; in the first stage of base class training based on a large number of labeled samples, the background consistency modeling technology is used to optimize the feature distribution between the background and the new class, and between the new classes; in the second stage of new class training based on a small number of labeled new class samples, the inter-class relationship mining technology is used to enhance the new class prototype using the base class prototype, and based on the base class prototype and the new class prototype, base class query features, new class query features and background features are generated for segmenting objects of the corresponding classes; the present application quickly adapts to new classes using a small number of samples, and efficiently completes the semantic segmentation task of remote sensing images by dealing with the problems of new classes and new scenes that may occur in remote sensing scenes.

This invention relates to the field of bioinformatics processing and discloses a method and system for mining the structure-activity relationship (SMR) of porcine neurotrophic peptides based on sequence features. The method includes constructing an original sample index table and fusing multi-source production data, extracting peptide sequence features to generate a sequence feature matrix, constructing a sequence-process joint graph containing peptide nodes and process state nodes, training a structure-activity relationship graph neural network to mine SMR relationships, and deriving a process control decision table based on a process response sample set generated by the network, thereby achieving online optimization of the porcine neurotrophic peptide preparation process. This invention solves the problem of SMR mining caused by the separation of process parameters, sequence information, and activity data, achieving accurate characterization of the synergistic effect of sequence and process and reverse optimization of process parameters, thus improving the targeted enrichment efficiency and bioactivity retention level of target neurotrophic peptides.

The invention discloses a multi-modal recommendation method for cross-modal semantic alignment and graph relation enhancement. The multi-modal recommendation method comprises the following steps: S1, initializing a model and setting parameters; s2, constructing an article heterogeneous relation graph fusing semantic information and behavior information; s3, executing cross-modal local-global multi-granularity semantic alignment; and S4, performing model optimization and preference prediction based on the aligned multi-modal representation. Through the integrated design of heterogeneous relation fusion and cross-level alignment, the problem that in an existing method, article relation mining is one-sided and cross-modal feature alignment is insufficient is effectively solved, and recommendation accuracy and robustness are remarkably improved.

The application discloses a human-object interaction detection method based on AutoHOINet, comprising the following steps: obtaining the visual modal vector of the to-be-detected image from the to-be-detected image; through a candidate image construction module, extracting the human object and the object object in the to-be-detected image, and reconstructing the same to generate a candidate image; using a human latent relationship mining module, performing human correlation analysis and latent relationship mining to find various latent interaction relationships between the human object and the object object; with the help of a human relationship reasoning module, screening out target relationships from the latent relationships to generate pseudo labels to guide the learning process of the HOI model; finally, using the generated pseudo labels to guide the learning of the HOI model, so as to realize the detection of human interaction. The application can be better than the current weakly supervised HOI model using image-level interaction labels on two benchmark datasets without using artificial labeled HOI label data and weak supervision.

The invention provides a chip heat dissipation layout optimization method based on an improved Transform model, and the method employs the improved Transform model to carry out the preprocessing of training data, strengthens the analysis capability of a model for the temperature intensity under a spatial position relation for the structured spatial data of a chip position coordinate, can effectively capture an interconnection relation during the processing of the structured sequence data, and achieves the optimization of the heat dissipation layout of a chip. The complex relationship between different data sequences can be accurately represented and predicted; and meanwhile, a nonlinear feature enhancement layer is added into the network model, so that the defect of mining a nonlinear feature relationship by a Transform model feed-forward network is made up, and the ability of the network model to learn the relationship between the layout of the chip position and the heat influence temperature is enhanced. According to the method, the limitation of a traditional chip based on human experience is broken through, the design process can be accelerated, the effective heat dissipation control of the optimal layout of the circuit board chip position can be improved through accurate layout optimization, the service life of electronic equipment is finally prolonged, and the maintenance cost of an electronic equipment system is reduced.

The application provides an event causal relationship mining method and system based on atlas constraint and time sequence optimal transmission, which comprises the following steps: cutting a burst event description text into text blocks; performing atomic event extraction on each text block, and formalizing each atomic event into a binary feature group; mapping any two atomic events and into event nodes and in a preset knowledge atlas, calculating the shortest path topological distance of the event nodes and in the atlas, and constructing a teacher model probability distribution according to the shortest path topological distance; calculating the cosine distance between the vectors and as a basic semantic transmission cost, and performing time sequence constraint on the basic semantic transmission cost to construct a time sequence transmission cost matrix, searching for an optimal transmission matrix that minimizes the total transmission cost; and obtaining an event causal relationship that meets the time sequence optimal transmission under the constraint of a field knowledge atlas by minimizing a joint loss function of the teacher probability distribution and the optimal transmission matrix.

The invention discloses a digital asset traceability method based on a hybrid model. The traceability method comprises the following steps: S1, preprocessing digital asset data; s2, carrying out initialization processing on the model; s3, training the model, and carrying out iterative optimization; and S4, performing traceability reasoning on the digital assets, and performing association relationship mining. Through the RNN and attention mixed model, the explicit and implicit relationships among massive digital assets are solved, and an efficient and credible solution is provided for digital asset traceability and security in the financial industry.

The invention provides a bid inviting and purchasing penetration type supervision AI large model construction method and system fused with multi-modal data. The method belongs to the technical field of artificial intelligence and intelligent auditing. The method comprises the steps of performing multi-source data fusion processing on internal system data and external data related to bid invitation purchasing, generating a cross-domain associated data set, performing implicit associated feature mining, obtaining implicit associated data among bidders, deploying a graph neural network, and constructing an association relationship mining network. According to the industrial design review method based on artificial intelligence, through multi-source data fusion and implicit association mining, design details and potential association can be accurately captured, and the comprehensiveness and accuracy of design review are improved.

The invention discloses a financial data index dynamic mining method and system based on machine learning, belongs to the technical field of financial science and technology and data processing, and solves the problems that dynamic change and real-time fluctuation of a financial market are difficult to cope with when an existing method performs big data mining through index classification results and subject entity relationship mining, and the mining efficiency is low. The mining result lags behind the market development, and the timely response to the index evolution process is lacked. The method comprises the following steps: pre-constructing a dynamic mining model, acquiring real-time associated data associated with a target financial index, pre-processing the real-time associated data, performing associated identification on a pre-processed data set by the dynamic mining model, and outputting a dynamic mining result comprising a candidate index set. Through trend analysis, anomaly detection and multi-dimensional feature extraction, a financial index system with real-time performance can be automatically generated based on market changes, and therefore the response capability and accuracy are improved in the aspects of financial risk monitoring, investment decision making, macroeconomic prediction and the like.