42 results about "Dependence network" patented technology

The invention relates to the technical field of intelligent substations, in particular to a rapid detection system, method and equipment for a process level switch of an intelligent substation, and aims to solve the problem that existing detection only depends on network transmission performance indexes and cannot comprehensively evaluate the business function correctness and operation reliability of the switch. The system, the method and the equipment comprise a power system fault scene modeling and message generation unit, a high-precision synchronous message injection and capture unit, a multi-dimensional state association and comprehensive judgment unit and a diagnosis report generation unit. According to the scheme, comprehensive detection, high-fidelity scene and deep diagnosis are realized, and service reliability determinacy evaluation is provided.

The invention relates to the technical field of data processing, and discloses a patient rehabilitation trajectory prediction and personalized nursing plan generation method and system. The method comprises the steps of collecting multi-dimensional function evaluation data to calculate a rehabilitation rate index sequence, recognizing a rehabilitation stage and configuring a feature weight vector, predicting a multi-dimensional rehabilitation trajectory through a rehabilitation stage gating sequential network, constructing a nursing task dependency network based on prediction data, and generating a personalized nursing plan through multi-objective optimization. According to the method, the accuracy of rehabilitation track prediction and the individuation degree of the nursing plan are improved, and comprehensive optimization of rehabilitation benefits, safety risks and resource consumption is realized.

The invention belongs to the technical field of tunnel matching construction early warning, and particularly relates to a process collaborative operation early warning system based on tunnel large-machine matching construction, which performs multi-source cross validation on process events uploaded in tunnel large-machine matching construction by fusing space positioning information and real-time operation state signals of corresponding responsible equipment of the process events. According to the method, credible judgment on the authenticity of an original process event is realized, meanwhile, the process event passing cross verification is taken as a node, a process dependency network is constructed based on the sequence dependency and parallel cooperation relationship between the processes, and the time sequence coupling logic of multiple processes is explicitly described; and after the collaborative deviation is identified, recursively deducing a conduction path of the delay under the constraint of a sequence chain and parallel by taking a deviation node as a starting point. Therefore, early warning is upgraded from single-point overtime judgment to global evaluation based on full-link influence, a prospective basis is provided for active scheduling, and meanwhile, early warning triggering depends on actual influence diffusion, so that early warning timeliness and accuracy are improved.

The invention relates to an earth pressure balance shield construction safety toughness dynamic evaluation system and method based on an extended cloud model and a network analysis method, and the system comprises an index system construction module which builds a multi-level toughness evaluation system based on a literature measurement and factor analysis method; an entropy weight TOPSIS weight calculation module objectively calculates the initial weight of the index through an information entropy theory; the ANP network weight optimization module corrects and optimizes the global weights of the indexes by constructing an inter-index nonlinear dependency network and a feedback mechanism; and the extended cloud toughness evaluation module quantifies qualitative indexes into membership degrees for different toughness levels based on expectation, entropy, hyper-entropy and other digital characteristics, comprehensively integrates weights and the membership degrees, and finally outputs soil pressure balance shield construction safety toughness levels and targeted optimization strategies. According to the method, multi-dimensional toughness quantitative evaluation and dynamic prevention and control of the construction safety risk of the earth pressure balance shield (EPB) can be realized under the complex stratum condition.

The invention discloses a multi-granularity data cascade updating method and device based on a dynamic topological graph and a medium, and relates to the technical field of data processing. The method maintains a multi-level data dependency model (DAG) in memory. The system recursively updates a target node by monitoring the change of a source node and utilizing an incremental propagation algorithm. For a high-concurrency scene needing isolation verification, the method introduces an overlay shadow topology mechanism: on the premise of not destroying a main graph structure and not copying a total graph, only instantiating a shadow copy of an affected node based on a Context ID, and constructing an overlay Mapping pointing to a father node; when the propagation path is calculated, the numerical value of the shadow node is preferentially read based on the context, and the propagation path of the original node is logically blocked in the current context. The invention further provides a path convergence mechanism and rendering frame synchronization technology based on topology in-degree, and the problems of resource competition, data consistency and front-end rendering flicker under the complex dependence network are effectively solved.

The invention discloses a dynamic and static track irregularity mapping method, which comprises the following steps: collecting dynamic and static data, and preprocessing the dynamic and static data; inputting the preprocessed data as a dynamic and static data mapping model, sorting dynamic data according to two dimensions of mileage and time, and extracting local features from the mileage by adopting a CNN network; meanwhile, the method depends on a transform network to capture a time sequence relationship, so that higher prediction accuracy is realized; according to the method, the existing detection technology can be optimized, the detection frequency is improved, the detection efficiency can be further improved, and the operation safety of the high-speed railway is guaranteed.

The invention provides a streaming processing method for real-time cleaning of agricultural product marketing data, and belongs to the technical field of big data processing.The method includes the steps that a field dependency directed graph is constructed, dependency intensity is quantified through mutual information, a dependency network is decomposed into weak coupling sub-graph sets through a minimum cut algorithm, and the weak coupling sub-graph sets are inserted into intermediate materialization nodes to achieve decoupling; performing Hamiltonian path inspired operator fusion optimization on each sub-graph to generate a super operator, converting the super operator into a machine code through a just-in-time compiler, establishing a multi-stage cache affinity scheduling mechanism and a dynamic watermark mechanism to process out-of-order data, and adopting a two-stage dynamic repartitioning strategy and a work stealing queue to cope with data skew; a hierarchical storage architecture is constructed, and capacity distribution is adjusted through a storage redundancy optimization function, so that the technical problem that the parallel processing capacity is limited due to a strong dependency relationship between fields in the agricultural product marketing data real-time cleaning process is solved.

The invention discloses a virtual machine monitoring method and device, equipment, a storage medium and a product, and relates to the technical field of cloud computing, and the method comprises the steps that a preset security agent in a virtual machine is started, the virtual machine is created by calling a virtualization management controller in a target cluster according to a preset virtual machine mirror image, and the preset virtual machine mirror image is integrated with the preset security agent; activating a data transmission channel between the virtual machine and the virtualization management controller through a preset security agent; and obtaining virtual machine operation information of the virtual machine based on the data transmission channel. The data transmission channel created by the virtualization management controller is activated through the pre-integrated preset security agent, and the bottom layer implementation of the data transmission channel does not depend on a TCP / IP network stack, so that the virtual machine operation information of the virtual machine can still be acquired even if a network is not set for the virtual machine, and the monitoring of the virtual machine is realized.

The application provides a virtual large language model LLM device, and belongs to the technical field of virtualization, and comprises an LLM app, a virtio_llm device, a virtio_llm driver, physical hardware Host, a virtual machine monitor Hypervisor, a virtual machine Guest and a large language model inference device LLM. The application does not rely on network transmission data, and does not need to passthrough computing resources to the virtual machine.

The invention relates to the technical field of computers, and discloses a data task processing method and device, equipment and a medium, and the method comprises the steps: obtaining a query statement in a data task, and analyzing the query statement to obtain grammar element information corresponding to the query statement; wherein the grammar element information comprises a table name, a field name, an operation mode and a connection condition; performing semantic pairing on the query statement based on the language element information to obtain a transfer mapping pair set between an input field and an output field in the query statement; constructing a transmission path between the input field and the output field according to the transfer mapping pair set to generate a directed dependent network structure; and mapping the directed dependent network structure between task nodes of the data task to obtain a blood relationship graph corresponding to the data task. The method can be applied to business management program systems such as financial science and technology, medical treatment, health and pension and the like, and can adaptively identify structural dependence of data tasks and quickly realize data consanguinity extraction.

The application belongs to the field of structure fracture damage simulation, and provides a structure fracture damage simulation method and system based on deep learning to solve the problem that the prediction accuracy of the whole process of complex fracture behavior depends on network architecture design and optimization algorithm, and the universality is limited. The structure fracture damage simulation method based on deep learning comprises the following steps: constructing a neural network model and setting initial parameters; optimizing the initial parameters of the neural network model based on a sample set and a loss function to obtain a trained neural network model; and predicting the initiation, expansion and bifurcation and merging of cracks in the structure domain based on the current structure geometric parameters, material parameters, boundary conditions and initial crack information by using the trained neural network model to obtain displacement field, stress field and damage evolution results. The method can enhance the adaptability of the model to different working conditions by introducing multi-scale and multi-physical field information.

The invention discloses a colorectal cancer individualized risk artificial intelligence assessment method and system, and the method comprises the following steps: carrying out the adaptive cross-domain multi-dimensional feature extraction of individual multi-source heterogeneous data, completing the missing value completion through semantic coding and channel attention, unifying the feature dimensions, and obtaining an alignment feature tensor; extracting a causal dependency relationship through a cross-domain multi-dimensional causal inference network to form a causal dependency network; extracting local-global fusion features by using a concentric circle Mangbar network; executing multi-head attention aggregation by taking the causal dependency network as a query vector and the fusion feature as a key value to obtain a causal-oriented fusion feature; and finally, outputting a colorectal cancer individualized risk probability value through a full-connection network. According to the method, the complex nonlinear relation can be effectively modeled, and the dependence on artificial feature selection is reduced.

The invention provides a road property state monitoring data analysis method and system applying deep learning, and the method comprises the steps: obtaining a road property data set, carrying out the cross-modal feature correlation mapping of the road property data set, generating a cross-modal correlation feature matrix, constructing a feature evolution dependence network based on the cross-modal correlation feature matrix, and carrying out the analysis of road property state monitoring data. And outputting a road property state reasoning result through the characteristic evolution dependence network. According to the invention, the comprehensiveness and accuracy of road property state monitoring can be effectively improved.

This invention relates to the field of computer software maintenance, and more particularly to a method, apparatus, and storage medium for software product information management based on big data. By constructing a knowledge graph of information items for the target software product, it associates event data and information items in multi-source event streams and calculates the comprehensive coupling degree. Information items exceeding a preset activation threshold are marked as vulnerability propagation sources. The cumulative vulnerability heat value of each information item is calculated, as well as the estimated remaining time for each information item to degrade to a preset risk state. Finally, a maintenance work order sequence is generated based on the estimated remaining time and the cumulative vulnerability heat value. This achieves an upgrade in proactive risk propagation quantitative assessment, enabling the identification of vulnerability propagation paths and affected scopes within the software's internal dependency network. Furthermore, it prioritizes maintenance based on degradation trend predictions, ensuring that maintenance resources are focused on critical information items with imminent risks and broad impact, thereby improving the scientific nature of maintenance decisions and the efficiency of operational resource utilization.

The invention relates to a collaborative scheduling method, device and system for a building energy system. The method comprises the steps that under the condition that the building energy system is in an operation stage, the building energy system selects an optimal strategy according to module information and preset information, outputs first information, adjusts operation data and outputs second information; and the coordination control module of the building energy system periodically calculates the Pareto boundary of the whole system and outputs coordination scheduling indication information. The problems that in an existing building energy system distributed photovoltaic, energy storage and heat pump system, equipment integration is complex, a control framework is concentrated, operation depends on a network, and self-adaptive optimization capacity is lacked can be solved.

The invention relates to a technology in the field of neural network application, in particular to a demand prediction method based on multimode power adaptive space-time dependent learning, which comprises the following steps of: constructing a power decision network in an offline stage and randomly initializing a modeled demand prediction model; the demand prediction model is composed of a feature representation module, a time network module, a space network module, a space-time dependence network module, a space-time fusion network module, a demand prediction network module and a loss calculation module. Performing feature representation on the multimode power flow data by using a feature representation module, forming a core of a demand model through a time network module, a space network module, a space-time dependence network module and a space-time fusion module, and forming a demand prediction model; and in the online use stage, through a demand prediction model obtained through training, power flow demand prediction is realized based on the input multimode power state data.

The invention provides a call link tracking method, system and device based on static analysis and a medium, belongs to the technical field of software engineering, and aims to solve the problem that a complete call link in a front-end application adopting a state management mechanism cannot be accurately tracked in the prior art. The method comprises the following steps: statically analyzing a front-end source code to extract interface definition information; constructing a front-end dependency relationship network comprising a module import relationship and a calling relationship, wherein the calling relationship covers a direct calling mode and an indirect calling mode via a state management mechanism; and based on the network, upward traversal is carried out from a calling point of a target back-end interface, and a complete calling link from the front-end user interface element to the interface is generated. According to the method and the device, a breakpoint-free dependency network can be constructed, complete and accurate link tracking is realized, the universality is high, and the research and development efficiency is improved.

This invention provides a method and system for analyzing road asset condition monitoring data using deep learning. By acquiring a road asset data set, cross-modal feature association mapping is performed on the data set to generate a cross-modal association feature matrix. A feature evolution dependency network is constructed based on the cross-modal association feature matrix, and the road asset condition inference result is output through the feature evolution dependency network. This invention can effectively improve the comprehensiveness and accuracy of road asset condition monitoring.

This invention discloses a context-based adaptive entropy coding method for audio, comprising the following steps: First, the quantized codebook index sequence Q and quantization side information S of the audio signal are sequentially fed into a causal convolutional context model, a cross-band attention network, and a cascaded codebook dependency network for processing, to obtain temporal context feature vectors, frequency context feature vectors, and quantization-level context feature vectors for the audio signal in the time domain, respectively. Second, the temporal context feature vectors, frequency context feature vectors, and quantization-level feature vectors are input into a neural probability prediction model for processing, estimating the conditional probability distribution of each codebook index. Third, the predicted conditional probability distributions are input into an adaptive arithmetic encoder to generate the encoded bitstream, and necessary header information is added to the encoded bitstream to form a complete bitstream. This invention can further improve the compression efficiency of audio coding.

The application belongs to the field of electric power automation, and discloses a power internet of things security risk assessment method, system, device and storage medium, which comprises the following steps: based on the dependence network theory, an information network diagram, a power network diagram and an inter-network dependence relationship of the power internet of things are established; an attacked node in the information network diagram is obtained and a state value thereof is assigned as a preset attack success probability, then the state values of all nodes in the information network diagram are updated according to an information side diagram calculation method; the attack success probability of each node in the power network diagram is determined through the inter-network dependence relationship according to the state values of all nodes in the updated information network diagram; the state values of all nodes in the power network diagram are determined through a physical side diagram calculation method according to the attack success probability of each node in the power network diagram; and a security risk assessment value of the power internet of things is obtained according to the state values of all nodes in the power network diagram. The application has great improvement in the accuracy and comprehensiveness of security assessment, and provides guidance for designing or improving the topology of the power internet of things.

The invention discloses a system performance analysis and evaluation method based on a fuzzy function dependence network, and the method comprises the following steps: constructing a performance evaluation model based on the fuzzy function dependence network through the analysis of an equipment system, and the fuzzy dependency strength and the fuzzy dependency criticality of the dependency relationship of each pair of nodes and the autonomous efficiency value of each node are specified, and the output efficiency value of the nodes is calculated through hierarchical step-by-step recursion, so that the comprehensive evaluation of the behavior efficiency of the system is realized. According to the method, attribute parameters are processed through the triangular fuzzy number and the fuzzy logic, and the defect that values are too subjective and absolute when two parameters of dependency intensity and dependency criticality of the node dependency relationship are determined during system performance analysis and evaluation based on a traditional function dependency network is overcome. The fuzziness of manually set parameters is effectively reflected, the scientific reliability and accuracy of analysis are improved, and the system efficiency analysis and evaluation work can be scientifically, reasonably and comprehensively implemented.

The invention relates to the technical field of computers, and discloses a user identity authentication method and device based on a dynamic verification code, and the method comprises the steps: generating an encryption object in response to an identity authentication request, the encryption object comprising a random number and a random key; performing asymmetric encryption on the encrypted object based on a pre-stored user public key to generate an encrypted ciphertext; converting the encrypted ciphertext into a two-dimensional code, sending the two-dimensional code to a client for display, and activating a dynamic verification code input interface; receiving a dynamic verification code input by a user through a client, and performing encryption operation on the random number through the random key to generate a corresponding reference code; and comparing the dynamic verification code with the reference code to generate a user identity authentication result. According to the invention, safe and networking-free user identity authentication can be realized, the security and convenience of user identity authentication are improved, and the problems of insufficient security, dependence on a network or time synchronization and the like in a traditional mode are solved.

The invention belongs to the field of structural fracture damage simulation, and provides a structural fracture damage simulation method and system based on deep learning in order to solve the problem that the whole-process prediction precision of complex fracture behaviors depends on network architecture design and an optimization algorithm, and the universality is limited. The structure fracture damage simulation method based on deep learning comprises the following steps: constructing a neural network model and setting initial parameters; optimizing initial parameters of the neural network model based on a sample set and a loss function to obtain a trained neural network model; and on the basis of current structure geometric parameters, material parameters, boundary conditions and initial crack information, crack initiation, expansion and bifurcation combination are predicted in a structural domain by using the trained neural network model, and a displacement field, a stress field and a damage evolution result are obtained. The adaptability of the model to different working conditions can be enhanced by introducing multi-scale and multi-physical field information.

The invention provides virtual large language model (LLM) equipment, which belongs to the technical field of virtualization, and comprises an LLM app, a virtiollm device, a virtiollm driver, physical hardware Host, a virtual machine monitor Hypervisor, a virtual machine Guest and large language model reasoning equipment LLM, and the LLM app, the virtiollm device, the virtiollm driver, the physical hardware Host, the virtual machine monitor Hypervisor, the virtual machine Guest and the large language model reasoning equipment LLM. The method does not depend on a network to transmit data, and the computing resource does not need to be passcored to the virtual machine.

The invention provides a deep learning-based building carbon emission data processing method and a computer system, and the method comprises the steps: obtaining a building carbon emission data set, executing feature correlation mining, generating a correlation intensity tensor, enabling a first dimension of the correlation intensity tensor to correspond to a function unit, enabling a second dimension of the correlation intensity tensor to correspond to a time sequence position, and obtaining a building carbon emission data set; the third dimension corresponds to a feature association type, and the tensor element represents the association degree between the carbon emission feature of the corresponding functional unit at the corresponding time sequence position and the carbon emission features of other functional units; constructing a time-varying dependency network based on the association strength tensor to obtain time-varying dependency network parameters; tracking a multi-scale evolution trajectory through time-varying dependent network parameters to generate a multi-scale evolution trajectory set; and executing attention mechanism key path identification through the multi-scale evolution trajectory set, and outputting a carbon emission data processing result. According to the invention, the effectiveness and reliability of the overall carbon emission data processing result can be improved.

The application discloses a kind of communication link perception multi-agent dynamic event trigger consistency methods, comprising S1: the communication network topology graph of information exchange between multi-agent is established;S2: determine the dynamics equation of multi-agent system;S3: introduce adaptive control, design complete distributed consistency control input strategy based on event trigger;S4: the solution of algebraic riccati equation obtains feedback gain matrix;S5: based on Lyapunov theorem and channel capacity parameter, design communication link perception dynamic event trigger control strategy;S6: reach multi-agent consistency by complete distributed control input strategy and communication link perception dynamic event trigger control strategy;The communication link perception dynamic event trigger mechanism of the application is completely distributed, through adaptive coupling gain, control input adaptive adjustment, without relying on global information related to network topology, realize and apply in large networked multi-agent system.

This disclosure provides methods, components, devices and systems for coordinated scheduling-based protocols for multi-hop traffic flows in a controller-based network. Some aspects more specifically relate to a coordinated scheduling of time slots for multi-hop traffic flows across both fronthaul and backhaul links in controller-based networks. In some examples, an access point (AP) within the network may provide information indicative of one or more established schedules of time slots to a network controller and the network controller may use such information for conflict resolution between different schedules of time slots. In accordance with the propagation of schedules of time slots to the network controller, the network controller may obtain a complete view of traffic patterns within the network and leverage such a complete view for the conflict resolution. Additionally, or alternatively, an AP may rely on a configuration by the network controller of one or more coordinated schedules of time slots.

The invention relates to the technical field of psychological assessment, and discloses a 2.4 G communication networking technology-based group psychological assessment system, which comprises a 2.4 G networking module, an assessment task distribution module, a data acquisition and return module, an analysis processing module and an output module, a wireless communication network of the master and slave devices is established through the 2.4 G networking module, dependence on a traditional network is eliminated, stable data transmission in an environment without an external network is realized, evaluation tasks are synchronously issued in cooperation with the evaluation task distribution module, the group evaluation efficiency is greatly improved, and the problem that single-person evaluation consumes long time is solved; answer data are collected and stored in real time through the data collection and return module, psychological laws are mined in combination with JAVA data cleaning and a machine learning algorithm of the analysis processing module, and then the visual display and report export functions of the output module are utilized, so that the hardware and labor cost is reduced, the data processing efficiency and the analysis application value are improved, and the application prospect is wide. A complete evaluation technical scheme is constructed, and the problem that multi-person evaluation depends on a network or multi-device linkage is effectively solved.

The application discloses a kind of weak supervision based on self-evolution's video anomaly detection method, comprising: 1) the original feature is extracted by video feature extractor to data set;2) the feature rich in time information is extracted by time dependence network to original feature;3) the feature rich in time information is formed confidence matrix and similarity label matrix by self-evolution method;4) according to similarity label matrix, form fragment level pseudo label;5) the pseudo label forms cross entropy loss to guide network optimization to complete anomaly detection task.The application makes deep network only focus on the sample that feature is more obvious in the early stage of training by self-evolution method, gradually learns more difficult to classify sample with the update iteration of network, uses fragment level pseudo label to replace video level weak label for training, so that all fragments participate in training, so as to more effectively utilize data set, and can improve the precision of video anomaly detection.

This invention provides a multimodal spatiotemporal fusion method and system for 3D target detection, relating to the field of target detection technology. In the multimodal and spatiotemporal fusion architecture, it innovatively introduces decoding branches for each modality and a current quality signal autonomously constructed from each initial detection result. This current quality signal serves as a unified adjustment mechanism, dynamically allocating feature fusion weights in both the temporal fusion and cross-modal fusion stages, breaking the limitation of traditional fusion methods that overly rely on implicit learning. Even when facing common challenges in real-world autonomous driving scenarios such as sparse LiDAR point clouds, camera obstruction by rain and fog, and glare from strong light, it can accurately detect targets. Without adding any external supervision labels, it can significantly improve the overall robustness, stability, and reliable localization capability of 3D target detection in complex scenarios.