77 results about "Web modeling" patented technology

Disclosed in the present invention is a four-network integration architecture for an unmanned swarm system. The four-network integration architecture has the capabilities of heterogeneous platform resource pooling, intelligent dynamic computing power allocation, and timely decision planning, so as to maximize the overall benefit. The present invention focuses on abstracting and integrating independent submodules to form a mesh topology of a swarm. The present invention designs a four-network integration architecture for an unmanned swarm system, which comprises a computing power network, a perception network, a decision network and a communication network as core modules. The structure aims to achieve efficient cooperation of all parts in the swarm, thereby improving the overall performance and adaptability of the system. The system integrates environmental perception, a swarm network modeling component, a knowledge base and a resource pool, providing an intelligent environmental perception strategy and a network modeling strategy for the interior of the swarm. Therefore, the perception of environments, tasks and networks by nodes can be facilitated, thereby completing establishment of intelligent networks, so as to ensure the characteristics of the stability and flexibility of networks.

The invention relates to the technical field of metallurgical raw material batching control, and discloses a bottom blowing smelting furnace automatic batching intelligent control method and system based on deep learning, and the method comprises the steps: achieving intelligent batching through multi-source data fusion, physical constraint modeling and dynamic optimization control; edge calculation is adopted to realize data space-time alignment and purification, and physical and economic mixed features are constructed; modeling a reaction path based on a graph neural network, and embedding conservation law constraint to synchronously predict key process parameters; and finally, in combination with gradient sensitivity analysis and reinforcement learning, constructing a differentiable optimization framework to realize multi-target dynamic ratio decision and real-time compensation control, and forming a perception-decision-execution closed loop. The system comprises a global sensing and data purification module, an intelligent decision-making and optimization batching module and a high-precision execution and closed-loop control module. According to the invention, the batching strategy is adaptively adjusted, and optimal resource allocation and maximum economic benefit are realized.

The invention discloses an equipment health state monitoring method and system based on the Internet of Things. According to the method, data in a multi-source sensor is obtained and preprocessed, a real-time operation data set of equipment is obtained, and a health quantification deviation value and a health state trend are determined through the real-time operation data set. And when the health state is abnormal, recording a starting point coordinate of an abnormal event, marking an abnormal event triggering timestamp, and analyzing the dependency relationship between the equipment by utilizing graph network modeling. Furthermore, the potential risk probability is analyzed through the long-short-term memory network, real-time monitoring, fault propagation prediction and risk assessment of the equipment health state are achieved, and the operation reliability and the maintenance efficiency of the industrial equipment are improved.

The invention discloses an AI-based digital project performance evaluation data processing method and system, and relates to the technical field of artificial intelligence and digital project management, and the method comprises a multi-source data collection module which is used for obtaining data; the entity alignment module is used for establishing an association relationship among cross-system data entities; the dynamic index generation module is used for dynamically adjusting the evaluation index weight by utilizing a reinforcement learning framework; the efficiency prediction module is used for modeling a task dependency relationship according to the time sequence diagram convolutional network and outputting a delay risk probability; and the visual interface is used for displaying the data. According to the method, cross-system entity alignment, dynamic index weight adjustment, task delay risk prediction and root cause analysis are realized through AI technologies such as BERT semantic matching and a graph neural network, visual data display and interaction functions are provided by virtue of a visual interface, the problems of data dispersion, index static state, risk lag and the like in traditional project evaluation are solved, and the project evaluation efficiency is improved. And the decision support capability of project management is improved.

The invention discloses an industrial control safety target range construction method and system based on a digital twinborn technology. The method comprises the following steps: constructing a digital twinborn simulation environment for an actual industrial scene, wherein the digital twinborn simulation environment comprises physical system modeling, communication network modeling and co-simulation modeling; defining an interactive interface between the simulation environment and an external tool, including selecting a communication mode according to a data exchange demand, defining a logic communication channel based on the selected communication mode and standardizing a data exchange format; converting an attack instruction initiated by an external tool into a simulation operation instruction in the simulation environment through the interactive interface, and feeding back a physical state change caused by the simulation operation instruction to the external tool; and an automatic target range resetting mechanism is established, and the state of the simulation environment is backed up and recovered through a snapshot mechanism. According to the invention, the management efficiency of the industrial control safety target range is improved, and the test reliability is improved.

The invention discloses a cross-domain AI knowledge aggregation method based on collaborative filtering. The method comprises the steps that S1, multi-source heterogeneous AI knowledge data and user behavior data are collected and preprocessed; s2, constructing a double-tower cross-domain embedded network, and outputting a cross-domain semantic fusion sequence; s3, modeling and analyzing user preferences through the improved Bi-GRU network, and generating user behavior preference vectors; s4, performing semantic diffusion and neighborhood reasoning on the cold start user, and complementing interest features; s5, adopting a double-tower recall structure and an XGBoost model to sort and generate a cross-domain recommendation list; s6, constructing a context rule base to execute context adaptability judgment, and generating a matched knowledge aggregation recommendation list; and S7, performing incremental learning according to user feedback information, and dynamically updating the double-tower cross-domain embedded network and the improved Bi-GRU network. According to the method, the knowledge matching precision, the cold start adaptability and the scene adaptation capability of cross-domain recommendation content are improved.

The invention discloses a lightweight cross-domain recommendation method and system based on user alignment Agent driving. The method comprises the following steps: firstly, acquiring historical behavior data of a user in multiple fields, fusing multi-modal contents such as texts and images, generating a fine-grained interest prototype through a cross-domain semantic encoder, and constructing a personalized Agent to simulate the intention of the user; then, in a multi-field collaborative environment, an Agent behavior strategy is optimized by utilizing reinforcement learning and a mixed reward mechanism, general preference and field specific preference are modeled through a hierarchical strategy network, and knowledge fusion is realized through a gating mechanism; and then, in combination with a preference distillation technology, extracting transferable characterization from Agent behaviors, and constructing a lightweight cross-domain knowledge graph. Finally, behavior track compression and cross-domain preference mapping are adopted, and efficient and low-consumption personalized recommendation is achieved. According to the method, the problems of cross-domain data sparsity and model complexity are effectively relieved, recommendation accuracy and system response efficiency are improved, and the method is suitable for real-time recommendation service of multiple scenes.

A method for training a machine learning model using enhanced multilayer direct disassembly networks (MDDNs). The method includes converting MDDNs into knowledge graphs and using them to train a generative and a discriminator component. The generative component produces synthetic disassembly structures, which are evaluated by the discriminator. Feedback from the discriminator is used to refine the generative model, improving the accuracy of disassembly network modeling.

The invention relates to the technical field of mine safety production, in particular to a mine ventilation parameter dynamic optimization method based on industrial internet of things, which comprises the following steps: S1, data fusion perception and edge calculation; s2, modeling a digital twin ventilation network; s3, performing multi-objective optimization solution; s4, instruction issuing and feedback; and S5, model self-correction and knowledge base updating. According to the scheme, the multi-dimensional sensor cluster and the dual-mode communication network are constructed through the industrial Internet of Things, real-time sensing and differential transmission of mine ventilation parameters are achieved, the data cleaning and feature extraction technology of edge calculation is combined, the authenticity and effectiveness of input data are ensured, and true and effective data are provided for dynamic optimization of the ventilation parameters.

The invention relates to an enterprise risk early warning method and device based on big data, equipment and a medium. The method comprises the following steps: respectively extracting time dynamic characteristics and space topological characteristics of nodes through a dynamic heterogeneous risk map, and generating unified space-time embedding representation through cross-modal fusion; an event logic enhanced nonlinear propagation effect of a dynamic heterogeneous graph attention network modeling risk event is adopted, and enhanced node representation and global representation are output; calculating the deviation between the node and the community mean value through topological residual detection to obtain a topological residual score, and calculating the minimum similarity between the node and the known risk prototype through similarity analysis to obtain a representation similarity score; the two scores are linearly combined and normalized to generate an unknown risk score, and a known risk probability is calculated based on a pre-training model; the comprehensive risk index is obtained through maximum value operation, comprehensive monitoring and early warning of known and unknown risks of an enterprise are achieved, and the coverage range, accuracy and timeliness of risk identification are remarkably improved.

The invention discloses a cross-version vulnerability recognition system based on a depth map neural network, and the system comprises a multi-version code metadata collection module which is used for collecting software source codes of multiple versions and analyzing the software source codes to generate a multi-version code metadata set; the multi-version code graph generation module is used for constructing a multi-version code graph set; the feature representation matrix generation module is used for generating a feature representation matrix of the multi-version code graph; the depth map neural network modeling module is used for obtaining an embedded matrix of each node by adopting a GLEM model; the potential vulnerability node identification module is used for identifying a cross-version potential vulnerability node set; and the result output module is used for outputting a cross-version vulnerability identification result and generating a detection report. According to the method, the depth map neural network and the cross-version modeling method are adopted, multi-version vulnerability automatic recognition is achieved, and the method has the advantages of being high in intelligence, high in adaptability and accurate in detection.

The invention relates to the technical field of thermal energy engineering and industrial artificial intelligence crossing, in particular to a boiler combustion multi-target cooperative control method based on PINN and reinforcement learning, which comprises the following steps: collecting boiler combustion related data through a multi-source sensor network, and fusing edge data; based on physical information neural network modeling training, predicting a key physical field in the boiler; constructing and training a reinforcement learning model based on PINN state embedding; and a multi-target cost function is constructed based on the reinforcement learning model, the multi-target cost function is continuously evaluated through the reinforcement learning model, the weight is automatically adjusted, and multi-target cooperative control over boiler combustion is achieved. The method is obviously superior to a traditional scheme in the aspects of physical consistency, adaptive capacity, real-time performance and multi-target cooperation, and a reproducible theory-engineering integrated new normal form is provided for efficient, clean and flexible operation of a coal-fired power plant boiler combustion system under the double-carbon background.

In order to solve the key problems of difficulty in multi-source information fusion, insufficient spatial topology modeling, attenuation of long-term prediction precision and the like in an existing aircraft ground taxiing trajectory prediction method, a parallel processing architecture of a historical trajectory encoder and a pavement path encoder is designed, and trajectory time sequence features are extracted by using a long-short-term memory network; modeling a spatial topological relation of a control path by adopting a graph attention network, and realizing effective integration of heterogeneous information through a feature fusion layer; a multi-component loss function fusing the position, the speed and the acceleration is provided, and the continuity and the smoothness of a prediction track are constrained; a lightweight data enhancement strategy and an adaptive residual connection mechanism are designed, the model generalization ability is improved, and long-term prediction error accumulation is relieved. According to the method, multi-source trajectory information can be effectively fused, and the accuracy and stability of aircraft ground taxiing trajectory prediction are remarkably improved.

The application provides a bearing health state online evaluation method and system based on morphological profile analysis and federal evolution hypergraph, aiming at solving the problems of poor model self-adaptability, difficult cross-device knowledge migration and easy to be submerged early weak fault characteristics of the prior art under dynamic working conditions. The method captures the geometric profile evolution of bearing micro-damage by constructing a morphological multi-scale profile feature extraction engine, topologically preserving morphological decomposition of the vibration signal; adopts a Bayesian Poisson online learning algorithm to realize dynamic threshold adaptive updating and early warning of the health index; introduces an evolutionary hypergraph neural network to model the high-order multi-element fault propagation relationship between the bearing and the adjacent components; finally, through a federal edge collaborative framework, the incremental aggregation and knowledge migration of the cross-device model are realized under the premise of protecting data privacy. The application significantly improves the robustness of bearing fault diagnosis under variable working conditions and the sensitivity of early warning, and provides a lightweight and evolving solution for intelligent operation and maintenance in distributed industrial scenarios.

The application discloses a kind of based on the transport hub maintenance scheduling efficiency optimization method of deconstruction of variable topology super network, comprising the following steps: step one, the maintenance element involved in the current real-time maintenance scheduling scene of transport hub and its interrelated relationship mapping modeling is modeled into supergraph network model;Step two, the basic uncertainty characteristics of supergraph network CNN topological structure are characterized;Step three, according to the supergraph network model, the evaluation value of measurement index is calculated;Step four, the three indexes calculated are compared with the reasonable value interval range of historical scheduling result, and the rationality of current scheduling state is judged;Step five, based on the measurement index, the maintenance scheduling efficiency of transport hub is improved.The beneficial effects of the present application are that the execution efficiency, resource collaboration degree and dynamic robustness of maintenance scheduling can be effectively improved through super network modeling and quantitative analysis, and bottleneck positioning and optimization are realized.

The application discloses a kind of cross-version vulnerability identification systems based on depth map neural network, comprising: multi-version code metadata acquisition module, for collecting multiple versions of software source code and parsing generation multi-version code metadata set;Multi-version code graph generation module, for constructing multi-version code graph set;Characteristic representation matrix generation module, for generating the characteristic representation matrix of multi-version code graph;Depth map neural network modeling module, for using GLEM model, obtains the embedding matrix of each node;Potential vulnerability node identification module, for identifying the potential vulnerability node set across version;Result output module, for outputting cross-version vulnerability identification result and generating detection report.The application adopts depth map neural network and cross-version modeling method, realizes multi-version vulnerability automatic identification, with the advantages of strong intelligence, high adaptability, detection precision.

The invention discloses a risk identification method based on an environment open source component, and relates to the technical field of network security, and the method comprises the steps: collecting heterogeneous data of the open source component in a development, deployment and operation environment, carrying out the semantic alignment of the heterogeneous data, and generating a semantic vector set with a unified dimension; inputting the semantic vector set into a multi-environment risk identification model, focusing on risk semantics of development, deployment and operation environment generality, and generating a risk tag; based on a cross-modal self-attention mechanism, performing dynamic weighting on the development, deployment and operation environment features to generate context feature vectors; constructing an environment component association graph based on the context feature vector, and modeling a risk propagation path by using a graph attention network; and based on the risk propagation weight matrix, the risk tag and the environment component association graph, calculating a dynamic risk score, and generating a multi-environment fusion risk score list. According to the method, through a multi-environment risk identification model, codes and dependency features are extracted, general risk semantics are focused, and accurate risk tags are generated.

This invention provides a modeling method and simulation system for complex mobile communication networks, belonging to the field of communication network simulation technology. The system includes a communication network model initialization module, a scenario data parsing module, a communication network state update module, a communication network model data management module, a communication link connectivity calculation module, a communication link seed edge selection module, a classical graph theory algorithm module, and an information communication performance calculation module. This invention first uses a seed edge algorithm to select the primary link, reducing redundant communication links to backup links. Then, it uses a classical graph theory algorithm to abstract the communication network in operation into a weighted undirected graph, thus simplifying, abstracting, and modeling complex mobile communication networks, reducing the complexity of topology analysis, and simplifying communication network solution. This invention supports communication performance calculation during the simulation process, effectively meeting the needs of communication network models in the simulation field.

The application discloses a defense strategy self-generating method and system for a smart device cluster, and belongs to the technical field of network security of a smart device cluster. The system comprises a network modeling module, an attack strategy integration module and a defense strategy dynamic generation module. The method comprises the following steps: a network attack and defense game model is constructed to represent a network attack and defense scene of the current smart device cluster; the network topology of the smart device cluster is acquired, and an attack strategy and an action path are selected according to the network topology; attack behaviors are performed on the smart device cluster according to the attack strategy; the defense strategy dynamic generation module acquires a defense matrix of the current smart device cluster and the detection capability of each network node cluster, and a deep reinforcement learning model is used to generate a defense strategy. The application solves the problems of poor adaptability, insufficient cooperation and unreasonable resource scheduling in the traditional defense technology of the smart device cluster network, and greatly improves the overall efficiency of network security defense.

The invention discloses a multi-dimensional data fusion asset assessment accurate modeling method and system. The method comprises the following steps: S1, carrying out multi-modal asset feature deconstruction and dynamic weight pre-configuration; s2, carrying out intelligent alignment and conflict resolution across spatio-temporal data fields; s3, value factor dynamic topology network modeling is carried out; s4, generating value confidence driven by a game fusion engine; and S5, dynamic self-optimization of the elastic evaluation architecture. Through an industry weight template of the feature deconstruction unit, dual-channel arbitration of the space-time anchoring unit and a cooperation-confrontation mechanism of the game fusion unit, penetrating coupling of a physical state-market fluctuation-law event is realized; the distributed arbitration unit and the man-machine cooperation unit form a legal and physical dual disaster recovery system, Bayesian evolution and gene recombination drive cognitive continuous optimization, and three industry bottlenecks of multi-dimensional splitting, risk lag and legal blind areas are systematically solved.

The invention provides a visual TMS logistics transportation management system and method, and relates to the technical field of logistics management. The visual TMS logistics transportation management system comprises a data fusing module which is used for accessing and fusing heterogeneous logistics data from an order management system, a warehouse management system, a vehicle positioning device and a carrier interface, and endowing each piece of data with a credibility weight and an influence potential energy label. According to the technology, heterogeneous data can be integrated, abnormal propagation risks can be quantified and a global optimal intervention scheme can be generated through dynamic influence network modeling and a collaborative entropy reduction mechanism, a network state and a decision process can be visually presented through a visual interaction module, fundamental transformation from isolated management to dynamic association and from local optimization to global entropy reduction is realized, and the network state and decision process can be visually presented. And the insight, the cooperation efficiency and the decision-making quality of logistics management can be greatly improved.

This invention discloses an efficient learning method for 6G edge networks based on graph-aware distributed neural networks, belonging to the field of communication technology. The method involves: modeling the edge network as a graph structure based on 6G edge network environment information; constructing a hybrid discrete-continuous non-convex joint optimization problem to find the node cluster partitioning method and distributed encoding / decoding parameters; dividing the hybrid discrete-continuous non-convex joint optimization problem into two sub-problems and solving them separately; iteratively optimizing the joint optimization problem until a preset maximum number of iterations is reached, outputting the final node cluster partitioning strategy. This invention unifies the graph partitioning decision of the network layer and the distributed feature learning of the model layer into an end-to-end joint optimization framework, achieving globally optimal inference accuracy and communication efficiency; simultaneously, by employing an alternating optimization framework and an energy-aware stochastic topology evolution algorithm, the complexity of solving the hybrid integer non-convex problem is effectively reduced.

The invention belongs to the field of information dissemination management, and discloses a node importance evaluation system based on information dissemination dynamics, which comprises a network modeling module for abstracting a social network into a directed graph, configuring a foundation for edges and enhancing two types of influence probabilities, and the latter is larger; a reverse reachable set generation module randomly selects a root node, generates a corresponding set through edge activation probability sampling reverse wide search, and records potential enhancement candidate nodes into a set at the same time; a propagation gain estimation module calculates and quantifies propagation expectation gain after node set enhancement according to multiple groups of independent set statistics; the enhanced node screening module iteratively selects k nodes with the maximum gain by a greedy algorithm according to a preset k value; and the result output module outputs a final enhanced node set. The method does not depend on seed nodes, adapts to a scene without a preset propagation starting point, and is wide in application range; the propagation gain of the screened enhanced node is prominent; based on the Chernov bound and the greedy strategy, estimation is reliable, the effect is stable, and the problem of performance fluctuation of a traditional heuristic algorithm does not exist.

In order to solve the key problems of the existing aircraft ground sliding trajectory prediction method, such as multi-source information fusion difficulty, space topology modeling deficiency and long-term prediction accuracy decay, the history trajectory encoder and the pavement path encoder parallel processing architecture are designed, the long short-term memory network is used to extract the trajectory time sequence characteristics, the graph attention network is used to model the space topology relationship of the control path, and the heterogeneous information is effectively integrated through the feature fusion layer; The multi-component loss function is proposed to fuse position, speed and acceleration, and the continuity and smoothness of the predicted trajectory are constrained; The lightweight data enhancement strategy and the adaptive residual connection mechanism are designed to improve the model generalization ability and relieve the long-term prediction error accumulation. The method can effectively fuse multi-source trajectory information, and significantly improve the accuracy and stability of the aircraft ground sliding trajectory prediction.

The application relates to a regional logistics demand prediction and transport capacity pre-scheduling method, which comprises the following steps: obtaining multi-source data of order quantity, weather, holidays, traffic state and economic activities, performing high-precision space-time alignment and normalization preprocessing, modeling space-time dependence relationship by using a deep Q network embedded with a multi-head self-attention mechanism, combining weighted back propagation to calculate characteristic attribution scores, generating a trend and cause-related semantic explanation based on business rules, driving an intelligent scheduling strategy of multi-objective optimization, and adaptively correcting prediction model parameters through a closed-loop feedback. The scheme realizes high precision, strong interpretability and intelligent collaborative scheduling response of logistics demand prediction, and effectively improves transport capacity allocation efficiency and system stability.

The application belongs to the technical field of cross-language entity linking, and discloses a cross-language entity linking method, system, medium, equipment and terminal.The cross-language entity linking method comprises the following steps: searching for candidate entities and constructing an entity name index; constructing an entity linking model based on different clues; performing vector embedding and fusion based on attribute-based entity linking; and performing multi-clue entity linking based on co-occurring entities.The application uses three kinds of clues to realize entity linking, the clues are entity attributes, co-occurring entities and context descriptions, and neural network structures such as LSTM, CNN and GCN are combined to perform word embedding representation and network modeling.Compared with traditional entity linking technology, the application has stronger semantic representation ability and the linking method has cross-language capability.Through analyzing different forms of information, the application realizes a multi-clue cross-language entity linking algorithm with three different structures, and completes accurate linking of news text cross-language entities such as persons and institutions and a given knowledge base.

The invention provides a large model load GPU (Graphics Processing Unit) modeling method based on a neural network, which comprises the following steps of: fusing GPU physical constraint and neural network adaptive learning, and classifying a load into five types of operators such as GEMM (Good Empirical Mode Model) according to a kernel name keyword through an operator splitter; the execution predictor calculates calculation time consumption Ccomp and memory access time consumption Cmem based on a formula, and key parameters alpha comp, a and b are learned by a neural network; a correction factor is generated through a multi-layer neural network modeling function r (x), and hardware deviation is dynamically corrected through Latency = (Ccomp + Cmem) * (1 + r (x)); and the period combiner accumulates various operator delays to obtain the total execution time.

The invention relates to the technical field of computer networks and communication, and discloses a network calculation-based key task system network end-to-end delay analysis method, which comprises the following steps of: modeling network equipment by adopting a pNC-QQU method; abstracting a processing unit jointly formed by at least one data queue and at least one quality of service (QoS) strategy associated with the data queue into an independent QQU (Quality of Service) unit in the network equipment; using a pNC-QoSTree method to construct a QoS strategy tree for describing the work logic of the complex QoS strategy in each QQU unit for each QQU unit; traversing each QQU (Quality Quality Unit) passed by the to-be-analyzed service flow f on the path R; for each QQU unit, respectively calculating an arrival curve and a service curve of the service flow f in the QQU unit based on the QoS strategy tree constructed by the QQU unit, and calculating according to the arrival curve and the service curve to obtain a local delay upper bound Dq of the service flow f passing through the QQU unit; and for each network device passing through the service flow f, superposing the local total delays Dd of all network devices on the path R to obtain an end-to-end delay upper bound of the service flow f. According to the invention, the network modeling difficulty of a user can be reduced.

The invention discloses a cloud edge collaborative computing power resource allocation method based on game analysis, and belongs to the field of information management and information systems. The invention provides a computing power network modeling and solving method aiming at the problem of how to select a proper computing power center to reduce the cost and improve the effectiveness under the data calculation requirement of an enterprise. According to the method, enterprise nodes, edge computing nodes and center computing nodes are constructed into a computing power service network, a cost function model covering data transmission, processing and computing links is established, and the network structure is simplified by mapping node cost into edge weights. On the basis, a variational inequality is introduced to describe an equilibrium condition in a calculation power center selection process, so that reasonable distribution of data traffic and minimization of total cost of an enterprise under different path selections are ensured. And a successive average method is further adopted to carry out iterative solution on the flow distribution, so that the equilibrium solution is converged and the path cost consistency is met. Scientific decision support can be provided for data task allocation of an enterprise in a multi-computing power center environment, the data computing cost is effectively reduced, the computing power utilization efficiency is improved, and technical support is provided for computing power resource allocation.

Methods and systems for utilizing transfer learning to deal with sparse datasets in wireless networks, particularly those associated with system-level network modeling. The methods and systems are designed to overcome the limitations posed by the sparsity and dynamicity of real network data, which is often difficult to collect due to the substantial costs and potential performance degradation associated with conducting system-level experiments on large numbers of base stations.