Network space geographic map monitoring system based on big data

Abstract

The invention relates to the technical field of intelligent monitoring, and particularly discloses a cyberspace geographic map monitoring system based on big data, which adopts an artificial intelligence technology based on deep learning to carry out semantic understanding on node description of a first node in a cyberspace geographic map, extracts semantic feature representation of the node description, and sends the semantic feature representation to the first node; meanwhile, time sequence analysis is conducted on the multi-source heterogeneous data of the first node, state information of the power grid nodes is mined from the multi-source heterogeneous data, and therefore node description is updated in combination with initial node description and the state information of the power grid nodes. Therefore, the real-time monitoring of the node state in the network space geographic map and the real-time updating of the map can be realized, the latest state of the node in the network space can be reflected in time, and the updating efficiency and accuracy of the network space geographic map are improved.

Description

Technical Field

[0001] The present application relates to the field of intelligent monitoring technology, and more specifically, to a cyberspace geographic map monitoring system based on big data. Background Art

[0002] Cyberspace, as a new spatial form of human activity and a common carrier of both people and information, scientifically characterizing cyberspace is a crucial foundation for analyzing cyber incidents, governing cyberspace, and ensuring network security. It also represents a new frontier for geographical research in the information age. Against the backdrop of fierce global competition in cyberspace, there is an urgent need to strengthen the cross-disciplinary integration of geography and cyberspace security, fostering theoretical and methodological innovation based on traditional geography to create a field of cyberspace geography.

[0003] Cyberspace geography is an emerging interdisciplinary field that intersects multiple disciplines, including geographic information science, computer science, and communications technology. It primarily studies geographic information, spatial relationships, and geographic phenomena in cyberspace. It explores geographic characteristics, spatial distribution patterns, and geographic location-related data in cyberspace, and constructs geographic information maps in cyberspace. Cyberspace geography maps represent various nodes (such as servers, routers, and devices) in cyberspace and the relationships between them as graphs, thereby presenting the geographic structure and spatial relationships in cyberspace.

[0004] As an important tool for describing the structure and properties of cyberspace, cyberspace geographic maps are crucial for network security management, configuration optimization, and troubleshooting. However, cyberspace geographic maps require real-time monitoring and updating to reflect dynamic changes in cyberspace. However, existing technologies lack effective methods for updating nodes in cyberspace geographic maps, making it difficult to timely reflect the latest status of nodes in the network. Therefore, an optimized cyberspace geographic map monitoring system based on big data is desired. Summary of the Invention

[0005] In order to solve the above technical problems, the present application is proposed. The embodiment of the present application provides a cyberspace geographic map monitoring system based on big data, which uses artificial intelligence technology based on deep learning to perform semantic understanding of the node description of the first node in the cyberspace geographic map, extracts the semantic feature representation of the node description, and simultaneously performs time series analysis on the multi-source heterogeneous data of the first node, thereby mining the status information of the power grid node, and then combining the initial node description and the status information of the power grid node to update the node description. In this way, real-time monitoring of the node status in the cyberspace geographic map and real-time updating of the map can be achieved, timely reflecting the latest status of the node in the cyberspace, and improving the update efficiency and accuracy of the cyberspace geographic map.

[0006] Accordingly, according to one aspect of the present application, a cyberspace geographic map monitoring system based on big data is provided, which includes:

[0007] An initial node description acquisition module is used to extract the initial node description of the first node in the cyberspace geographic map;

[0008] an initial node description semantic encoding module, configured to perform semantic encoding on the initial node description of the first node to obtain a semantic encoding feature vector of the first node initial description;

[0009] a multi-source heterogeneous data acquisition module, configured to acquire a time series of multi-source heterogeneous data of the first node collected by the sensor network, wherein the multi-source heterogeneous data includes network traffic data, device status data, and user behavior data;

[0010] A multi-source heterogeneous data time series encoding module is used to perform parameter sample-level time series encoding on the time series of the multi-source heterogeneous data to obtain a network traffic time series associated implicit feature vector, a device state time series associated implicit feature vector, and a user behavior time series associated implicit feature vector;

[0011] A power grid node state analysis module, configured to fuse the network traffic time series associated implicit feature vector, the device state time series associated implicit feature vector, and the user behavior time series associated implicit feature vector to obtain a multimodal power grid node state semantic encoding feature vector;

[0012] A node description updating module is configured to determine an updated node description of the first node based on an adaptive fusion feature of the multimodal power grid node state semantic coding feature vector and the first node initial description semantic coding feature vector.

[0013] In the above-mentioned big data-based cyberspace geographic map monitoring system, the initial node description semantic encoding module is used to: pass the initial node description of the first node through a semantic encoder based on the Bert model to obtain the semantic encoding feature vector of the first node initial description.

[0014] In the above-mentioned big data-based cyberspace geographic map monitoring system, the multi-source heterogeneous data time series encoding module includes: a data regularization unit, which is used to regularize the time series of the multi-source heterogeneous data according to the parameter sample dimension to obtain the time series of network traffic data, the time series of device status data and the time series of user behavior data; a time series encoding unit, which is used to perform time series analysis on the time series of the network traffic data, the time series of the device status data and the time series of the user behavior data respectively to obtain the network traffic time series associated implicit feature vector, the device status time series associated implicit feature vector and the user behavior time series associated implicit feature vector.

[0015] In the above-mentioned big data-based cyberspace geographic map monitoring system, the time series encoding unit is used to: pass the time series of the network traffic data, the time series of the device status data and the time series of the user behavior data through a sequence encoder based on the RNN model to obtain the network traffic time series associated implicit feature vector, the device status time series associated implicit feature vector and the user behavior time series associated implicit feature vector.

[0016] In the above-mentioned big data-based cyberspace geographic map monitoring system, the power grid node status analysis module is used to: input the network traffic time series associated implicit feature vector, the device state time series associated implicit feature vector and the user behavior time series associated implicit feature vector into the power grid node status analyzer based on the Bayesian probability network to obtain the multimodal power grid node status semantic encoding feature vector.

[0017] In the above-mentioned big data-based cyberspace geographic map monitoring system, the node description update module includes: an adaptive fusion unit, which is used to input the multimodal power grid node state semantic coding feature vector and the first node initial description semantic coding feature vector into the balanced threshold feature vector adaptive fusion network to obtain the node state update semantic coding feature vector; a node description generation unit, which is used to input the node state update semantic coding feature vector into the decoder-based node description updater to obtain the updated node description of the first node.

[0018] In the above-mentioned big data-based cyberspace geographic map monitoring system, the adaptive fusion unit includes: a multidimensional fusion subunit, which is used to pass the multimodal power grid node state semantic coding feature vector and the first node initial description semantic coding feature vector through a multidimensional fusion module to obtain a first power grid node state-initial description semantic fusion feature vector, a second power grid node state-initial description semantic fusion feature vector and a third power grid node state-initial description semantic fusion feature vector; a balanced threshold value calculation subunit, which is used to respectively calculate the balanced threshold values ​​of the first power grid node state-initial description semantic fusion feature vector, the second power grid node state-initial description semantic fusion feature vector and the third power grid node state-initial description semantic fusion feature vector to obtain a first balanced threshold value, a second balanced threshold value and a third balanced threshold value; a balanced fusion subunit, which is used to perform weighted fusion of the multimodal power grid node state semantic coding feature vector and the first node initial description semantic coding feature vector based on the first balanced threshold value, the second balanced threshold value and the third balanced threshold value to obtain the node state update semantic coding feature vector.

[0019] In the above-mentioned big data-based cyberspace geographic map monitoring system, the multi-dimensional fusion sub-unit is used to: cascade the multimodal power grid node state semantic coding feature vector and the first node initial description semantic coding feature vector to obtain the first power grid node state-initial description semantic fusion feature vector; add the multimodal power grid node state semantic coding feature vector and the first node initial description semantic coding feature vector by position to obtain the second power grid node state-initial description semantic fusion feature vector; and multiply the multimodal power grid node state semantic coding feature vector and the first node initial description semantic coding feature vector by position point to obtain the third power grid node state-initial description semantic fusion feature vector.

[0020] In the above-mentioned big data-based cyberspace geographic map monitoring system, the balance threshold value calculation subunit is used to: multiply the first power grid node state-initial description semantic fusion feature vector by a first predetermined transformation vector to obtain a first threshold scoring coefficient; add the first threshold scoring coefficient and the first bias parameter and then pass the sigmoid activation function to obtain the first balance threshold value.

[0021] In the above-mentioned big data-based cyberspace geographic map monitoring system, the balancing fusion subunit is used to: determine a first weight parameter and a second weight parameter based on the first balancing threshold value, the second balancing threshold value and the third balancing threshold value, wherein the first weight parameter is the average of the first balancing threshold value, the second balancing threshold value and the third balancing threshold value, and the second weight parameter is the difference between one and the first weight parameter; perform position-by-position weighting on the multimodal power grid node state semantic coding feature vector with the first weight parameter to obtain a weighted multimodal power grid node state semantic coding feature vector, and perform position-by-position weighting on the first node initial description semantic coding feature vector with the second weight parameter to obtain a weighted first node initial description semantic coding feature vector; and perform element-by-element addition on the weighted multimodal power grid node state semantic coding feature vector and the weighted first node initial description semantic coding feature vector to obtain the node state update semantic coding feature vector.

[0022] Compared to existing technologies, the big data-based cyberspace geographic map monitoring system provided by this application uses deep learning-based artificial intelligence technology to perform semantic understanding of the node description of the first node in the cyberspace geographic map, extracting the semantic feature representation of the node description. It also performs time series analysis on the multi-source heterogeneous data of the first node, mining the status information of the power grid node from it, and then combining the initial node description with the status information of the power grid node to update the node description. In this way, it is possible to achieve real-time monitoring of the node status in the cyberspace geographic map and real-time updating of the map, promptly reflecting the latest status of the node in the cyberspace, and improving the updating efficiency and accuracy of the cyberspace geographic map. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The above and other purposes, features, and advantages of the present application will become more apparent through a more detailed description of the embodiments of the present application in conjunction with the accompanying drawings. The accompanying drawings are intended to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the drawings, the same reference numerals generally represent the same components or steps.

[0024] Figure 1 This is a block diagram of a cyberspace geographic map monitoring system based on big data according to an embodiment of the present application.

[0025] Figure 2 Schematic diagram of the architecture of a cyberspace geographic map monitoring system based on big data according to an embodiment of the present application.

[0026] Figure 3 This is a block diagram of a multi-source heterogeneous data time series encoding module in a big data-based cyberspace geographic map monitoring system according to an embodiment of the present application.

[0027] Figure 4 This is a block diagram of a node description update module in a big data-based cyberspace geographic map monitoring system according to an embodiment of the present application.

[0028] Figure 5 This is a block diagram of an adaptive fusion unit in a big data-based cyberspace geographic map monitoring system according to an embodiment of the present application. DETAILED DESCRIPTION

[0029] Below, the embodiments of the present application will be described in more detail with reference to the accompanying drawings, and the above-mentioned and other purposes, features, and advantages of the present application will become more apparent. Obviously, the described embodiments are only some embodiments of the present application, rather than all embodiments of the present application, and it should be understood that the present application is not limited to the example embodiments described herein.

[0030] As mentioned in the background technology above, cyberspace geographic maps, as a key means of revealing the structure and properties of cyberspace, have significant value in many fields such as network security management, optimized configuration, and fault diagnosis. Among them, real-time monitoring and updating of cyberspace geographic maps are important links to ensure the accuracy and timeliness of map information. However, in terms of node updates in cyberspace geographic maps, there is currently no efficient node update method, making it difficult to present the latest status of nodes in the network in real time. To address the above technical problems, the technical concept of this application is to use deep learning-based artificial intelligence technology to perform semantic understanding on the node description of the first node in the cyberspace geographic map, extract the semantic feature representation of the node description, and simultaneously perform time series analysis on the multi-source heterogeneous data of the first node to mine the status information of the power grid node, thereby combining the initial node description and the power grid node status information to update the node description. In this way, it is possible to achieve real-time monitoring of the node status in the cyberspace geographic map and real-time updating of the map, timely reflecting the latest status of the node in the cyberspace, and improving the update efficiency and accuracy of the cyberspace geographic map.

[0031] Figure 1 This is a block diagram of a cyberspace geographic map monitoring system based on big data according to an embodiment of the present application. Figure 2 Schematic diagram of the architecture of the network space geographic map monitoring system based on big data according to the embodiment of the present application. Figure 1 and Figure 2As shown, according to the embodiment of the present application, the network space geographic map monitoring system 100 based on big data includes: an initial node description acquisition module 110, which is used to extract the initial node description of the first node in the network space geographic map; an initial node description semantic encoding module 120, which is used to semantically encode the initial node description of the first node to obtain the semantic encoding feature vector of the first node initial description; a multi-source heterogeneous data acquisition module 130, which is used to obtain the time series of multi-source heterogeneous data of the first node collected by the sensor network, wherein the multi-source heterogeneous data includes network traffic data, device status data and user behavior data; a multi-source heterogeneous data time series encoding module 140, which is used to encode the multi-source heterogeneous data. The time series of heterogeneous data is subjected to parameter sample level time series encoding to obtain network traffic time series associated implicit feature vectors, device state time series associated implicit feature vectors and user behavior time series associated implicit feature vectors; a power grid node state analysis module 150 is used to fuse the network traffic time series associated implicit feature vectors, the device state time series associated implicit feature vectors and the user behavior time series associated implicit feature vectors to obtain a multimodal power grid node state semantic coding feature vector; a node description update module 160 is used to determine an updated node description of the first node based on the adaptive fusion features of the multimodal power grid node state semantic coding feature vector and the first node initial description semantic coding feature vector.

[0032] In the above-mentioned big data-based cyberspace geographic map monitoring system 100, the initial node description acquisition module 110 is used to extract the initial node description of the first node in the cyberspace geographic map. It should be understood that the initial node description records the original state of the node at a certain moment, including basic information such as the node's identification, location, type, and function. In a complex and changeable network environment, the state of the node may change due to various factors. By extracting the initial node description as a benchmark for subsequent node status updates, a clear background and basis is provided for the update of the node description, which helps to improve the system's perception and understanding of node status changes in the cyberspace geographic map, and ensure the accuracy and reliability of the node description update.

[0033] In the above-mentioned cyberspace geographic map monitoring system 100 based on big data, the initial node description semantic encoding module 120 is used to semantically encode the initial node description of the first node to obtain the semantic encoding feature vector of the first node initial description. In a specific example of the present application, the encoding method for semantically encoding the initial node description of the first node is to pass the initial node description of the first node through a semantic encoder based on the Bert model to obtain the semantic encoding feature vector of the initial description of the first node. That is, in order to convert the initial node description of the first node into a vector form that can be understood by a computer and understand its content meaning, a semantic encoder based on the Bert model is used to process the initial node description of the first node. Those skilled in the art should know that the Bert model is a pre-trained language representation model that is based on the Transformer architecture and uses an attention mechanism to process natural language to generate high-quality word embeddings. In the technical solution of the present application, the Bert model adopts bidirectional encoding technology. When processing each word in the initial node description of the first node, it will simultaneously consider the contextual information on its left and right sides, so that the model can more comprehensively understand the semantic information and contextual relationship in the node description, and provide an in-depth understanding and accurate representation of the initial node description for subsequent node description update tasks.

[0034] In the above-mentioned big data-based cyberspace geographic map monitoring system 100, the multi-source heterogeneous data acquisition module 130 is used to obtain the time series of multi-source heterogeneous data of the first node collected by the sensor network, wherein the multi-source heterogeneous data includes network traffic data, device status data and user behavior data. It should be understood that the multi-source heterogeneous data of the first node reflects the real-time status and behavior characteristics of the first node in the cyberspace, and is an important basis for updating the node status. Among them, the network traffic data reflects the communication status and load status of the node, the device status data reflects the hardware performance and health status of the node, and the user behavior data reflects the usage of the node and user needs. Therefore, by continuously monitoring and comprehensively analyzing the multi-source heterogeneous data of the first node, the state changes and behavior patterns of the node can be comprehensively and accurately portrayed, providing strong data support for subsequent node description updates.

[0035] In the above-mentioned cyberspace geographic map monitoring system 100 based on big data, the multi-source heterogeneous data time series encoding module 140 is used to perform parameter sample-level time series encoding on the time series of the multi-source heterogeneous data to obtain network traffic time series associated implicit feature vectors, device status time series associated implicit feature vectors, and user behavior time series associated implicit feature vectors. Figure 3FIG is a block diagram of a multi-source heterogeneous data temporal coding module in a cyberspace geographic map monitoring system based on big data according to an embodiment of the present application. Figure 3 As shown, the multi-source heterogeneous data time series encoding module 140 includes: a data regularization unit 141, which is used to regularize the time series of the multi-source heterogeneous data according to the parameter sample dimension to obtain the time series of network traffic data, the time series of device status data and the time series of user behavior data; a time series encoding unit 142, which is used to perform time series analysis on the time series of the network traffic data, the time series of the device status data and the time series of the user behavior data respectively to obtain the network traffic time series associated implicit feature vector, the device status time series associated implicit feature vector and the user behavior time series associated implicit feature vector.

[0036] Specifically, the data regularization unit 141 is used to regularize the time series of the multi-source heterogeneous data according to the parameter sample dimension to obtain the time series of network traffic data, the time series of device status data and the time series of user behavior data. It should be understood that, considering that the network traffic data, the device status data and the user behavior data respectively reflect the different dimensional information of the node, in order to more accurately perform the node status analysis, in the technical solution of the present application, the time series of the multi-source heterogeneous data is split into the corresponding time series of network traffic data, the time series of device status data and the time series of user behavior data according to the parameter sample dimension, and ensure that the network traffic data, the device status data and the user behavior data are aligned in time, and unify the data of different sources and formats into the same format and dimension, so as to facilitate the subsequent data timing analysis and node status update.

[0037] Specifically, the time series encoding unit 142 is used to perform time series analysis on the time series of the network traffic data, the time series of the device status data, and the time series of the user behavior data to obtain the network traffic time series associated implicit feature vector, the device status time series associated implicit feature vector, and the user behavior time series associated implicit feature vector. In a specific example of the present application, the encoding method for performing time series analysis on the time series of the network traffic data, the time series of the device status data, and the time series of the user behavior data is to pass the time series of the network traffic data, the time series of the device status data, and the time series of the user behavior data through a sequence encoder based on an RNN model to obtain the network traffic time series associated implicit feature vector, the device status time series associated implicit feature vector, and the user behavior time series associated implicit feature vector. It should be understood that the RNN (recurrent neural network) model is a neural network structure specifically used to process sequence data. It can capture the time dependency in the sequence by iteratively processing each element in the sequence, thereby extracting implicit features from the time series data. In the technical solution of the present application, a sequence encoder based on an RNN model is used to perform time series analysis on the time series of network traffic data, device status data, and user behavior data, so as to explore the potential laws and trends of network traffic changes, device status changes, and user behavior changes, thereby revealing the state changes and behavior patterns of nodes, and providing an in-depth understanding of the real-time status and behavior patterns of nodes for subsequent node status updates.

[0038] In the above-mentioned big data-based cyberspace geographic map monitoring system 100, the power grid node state analysis module 150 is used to fuse the network traffic time series associated implicit feature vector, the device state time series associated implicit feature vector, and the user behavior time series associated implicit feature vector to obtain a multimodal power grid node state semantic encoding feature vector. In a specific example of the present application, the encoding method of fusing the network traffic time series associated implicit feature vector, the device state time series associated implicit feature vector, and the user behavior time series associated implicit feature vector is to input the network traffic time series associated implicit feature vector, the device state time series associated implicit feature vector, and the user behavior time series associated implicit feature vector into a power grid node state analyzer based on a Bayesian probability network to obtain the multimodal power grid node state semantic encoding feature vector. It should be understood that since network traffic data, device state data, and user behavior data respectively reflect different aspects of the power grid node state, by fusing the three features, the node state can be described more comprehensively and accurately, thereby obtaining a more accurate node description update result. At the same time, considering the complex associations and mutual influences between the network traffic data, device status data and user behavior data of the node, simple data splicing or weighted fusion may not be able to effectively mine the potential associations and deep features between the three. Therefore, in the technical solution of the present application, a power grid node state analyzer based on a Bayesian probability network is used to fuse the three. Specifically, a Bayesian probability network is a machine learning algorithm based on a probabilistic graph model, which can handle uncertainty and complex associations, and discover the dependencies between variables through learning and reasoning. Here, the device state time series association implicit feature vector is used as a prior probability, the network traffic time series association implicit feature vector is used as a conditional probability, and the user behavior time series association implicit feature vector is used as an evidence probability, so as to make full use of the potential association information between network traffic data, device status data and user behavior data to construct a Bayesian probability network model of the power grid node, effectively integrating the deep features of data from different sources, thereby obtaining a more accurate and comprehensive description of the node state.

[0039] In the above-mentioned big data-based cyberspace geographic map monitoring system 100, the node description updating module 160 is used to determine the updated node description of the first node based on the adaptive fusion feature of the multimodal power grid node state semantic encoding feature vector and the first node initial description semantic encoding feature vector. Figure 4 FIG. 1 is a block diagram of a node description update module in a network space geographic map monitoring system based on big data according to an embodiment of the present application. Figure 4As shown, the node description update module 160 includes: an adaptive fusion unit 161, which is used to input the multimodal power grid node state semantic coding feature vector and the first node initial description semantic coding feature vector into the balanced threshold feature vector adaptive fusion network to obtain a node state update semantic coding feature vector; a node description generation unit 162, which is used to input the node state update semantic coding feature vector into the decoder-based node description updater to obtain an updated node description of the first node.

[0040] Specifically, the adaptive fusion unit 161 is used to input the multimodal power grid node state semantic coding feature vector and the first node initial description semantic coding feature vector into the balanced threshold feature vector adaptive fusion network to obtain the node state update semantic coding feature vector. It should be understood that the multimodal power grid node state semantic coding feature vector provides key information about the node state extracted from the perspectives of network traffic, device status and user behavior; and the first node initial description semantic coding feature vector contains the initial basic information of the node. By fusing the two features, it is helpful to obtain a more complete and rich node state representation. However, since different modal data have different characteristics and expressions, direct merging may not be able to effectively utilize the information in various data. Therefore, in the technical solution of the present application, a balanced threshold feature vector adaptive fusion network is used to fuse the two. Specifically, the balanced threshold feature vector adaptive fusion network can adaptively adjust the fusion weight between the multimodal power grid node state semantic encoding feature vector and the first node initial description semantic encoding feature vector by learning and understanding the intrinsic relationship between the two, so as to balance and fuse information from different sources, thereby retaining and fusing effective information in different modal data to the greatest extent, making full use of the complementarity between multimodal data, reducing the impact of redundancy and noise, and improving the accuracy and reliability of node state updates.

[0041] Figure 5 FIG is a block diagram of an adaptive fusion unit in a network space geographic map monitoring system based on big data according to an embodiment of the present application. Figure 5As shown, the adaptive fusion unit 161 includes: a multi-dimensional fusion subunit 1611, which is used to pass the multi-dimensional fusion module of the multimodal power grid node state semantic coding feature vector and the first node initial description semantic coding feature vector to obtain a first power grid node state-initial description semantic fusion feature vector, a second power grid node state-initial description semantic fusion feature vector and a third power grid node state-initial description semantic fusion feature vector; a balanced threshold value calculation subunit 1612, which is used to respectively calculate the balanced threshold values ​​of the first power grid node state-initial description semantic fusion feature vector, the second power grid node state-initial description semantic fusion feature vector and the third power grid node state-initial description semantic fusion feature vector to obtain a first balanced threshold value, a second balanced threshold value and a third balanced threshold value; a balanced fusion subunit 1613, which is used to perform weighted fusion of the multimodal power grid node state semantic coding feature vector and the first node initial description semantic coding feature vector based on the first balanced threshold value, the second balanced threshold value and the third balanced threshold value to obtain the node state update semantic coding feature vector.

[0042] Specifically, the multi-dimensional fusion subunit 1611 is used to: cascade the multimodal power grid node state semantic coding feature vector and the first node initial description semantic coding feature vector to obtain the first power grid node state-initial description semantic fusion feature vector; add the multimodal power grid node state semantic coding feature vector and the first node initial description semantic coding feature vector by position to obtain the second power grid node state-initial description semantic fusion feature vector; and perform position point multiplication on the multimodal power grid node state semantic coding feature vector and the first node initial description semantic coding feature vector to obtain the third power grid node state-initial description semantic fusion feature vector.

[0043] Specifically, the balance threshold calculation subunit 1612 is configured to: multiply the first grid node state-initial description semantic fusion feature vector by a first predetermined transformation vector to obtain a first threshold scoring coefficient; and add the first threshold scoring coefficient and a first bias parameter, and then apply the resultant sum to a sigmoid activation function to obtain the first balance threshold. Similarly, the process of obtaining the second balance threshold based on the second predetermined transformation vector and the second bias parameter, and the process of obtaining the third balance threshold based on the third predetermined transformation vector and the third bias parameter, are similar, and will not be further elaborated upon herein.

[0044] Specifically, the balanced fusion subunit 1613 is used to: determine a first weight parameter and a second weight parameter based on the first balanced threshold value, the second balanced threshold value and the third balanced threshold value, wherein the first weight parameter is the average of the first balanced threshold value, the second balanced threshold value and the third balanced threshold value, and the second weight parameter is the difference between one and the first weight parameter; perform position-by-position weighting on the multimodal power grid node state semantic coding feature vector with the first weight parameter to obtain a weighted multimodal power grid node state semantic coding feature vector, and perform position-by-position weighting on the first node initial description semantic coding feature vector with the second weight parameter to obtain a weighted first node initial description semantic coding feature vector; and perform element-by-element addition of the weighted multimodal power grid node state semantic coding feature vector and the weighted first node initial description semantic coding feature vector to obtain the node state update semantic coding feature vector.

[0045] More specifically, the adaptive fusion unit 161 is configured to process the multimodal power grid node state semantic coding feature vector and the first node initial description semantic coding feature vector using the following adaptive fusion formula to obtain the node state update semantic coding feature vector, wherein the adaptive fusion formula is:

[0046]

[0047] Wherein, v1 is the semantic encoding feature vector of the multimodal power grid node state, v2 is the semantic encoding feature vector of the initial description of the first node, W1, W2 and W3 are the first predetermined transformation vector, the second predetermined transformation vector and the third predetermined transformation vector respectively, b1, b2 and b3 are the first bias parameter, the second bias parameter and the third bias parameter respectively; sigmoid represents the activation function, concat(·,·) represents the cascade processing, represents positional addition, ⊙ represents positional multiplication, t1, t2, and t3 are the first, second, and third equalization thresholds, respectively, and t1, t2, t3∈[0,1], v c A semantic encoding feature vector is updated for the node state.

[0048] Specifically, the node description generation unit 162 is used to input the node state update semantic coding feature vector into the decoder-based node description updater to obtain the updated node description of the first node. It should be understood that the decoder is a network structure that can recover the original data or generate new data from the coding features. In the technical solution of the present application, a decoder-based node description updater is used to process the node state update semantic coding feature vector, and the node state and behavior pattern information in the node state update semantic coding feature vector is decoded based on the powerful regression mapping capability of the decoder to generate corresponding node description information, thereby realizing the update of the first node state, so as to provide information basis for tasks such as power grid monitoring and fault prediction.

[0049] In particular, in the technical solution of the present application, the first node initial description semantic coding feature vector is used to represent the textual semantic features of the initial node description of the first node in the cyberspace geographic map, and the multimodal power grid node state semantic coding feature vector is used to represent the sample cross-temporal correlation features of multi-source heterogeneous data collected by the sensor network deployed at the first node. Taking into account the significant differences between feature modalities and feature dimensions, when the multimodal power grid node state semantic coding feature vector and the first node initial description semantic coding feature vector are input into the balanced threshold feature vector adaptive fusion network, local feature overflow will occur during the adaptive fusion process, thereby affecting the accuracy of the updated node description of the first node obtained by the node state update semantic coding feature vector through the decoder-based node description updater.

[0050] Based on this, in a preferred example of the present application, the node state update semantic coding feature vector is passed through a decoder-based node description updater to obtain an updated node description of the first node, including the following steps: calculating the node state update semantic coding and representation matrix and the node state update semantic coding difference representation matrix of the node state update semantic coding feature vector, wherein the value of each position of the node state update semantic coding and representation matrix is ​​the mean of a pair of eigenvalues ​​of the two positions of the node state update semantic coding feature vector corresponding to the position coordinates respectively, and the value of each position of the node state update semantic coding difference representation matrix is ​​the absolute value of the difference between a pair of eigenvalues ​​of the two positions of the node state update semantic coding feature vector corresponding to the position coordinates respectively; matrix multiplying the transpose vector of the node state update semantic coding feature vector with the node state update semantic coding and representation matrix to obtain the node state update semantic coding and representation vector, and multiplying the node state update semantic coding difference representation matrix with the node state update semantic coding feature vector. The node state update semantic coding feature vector is matrix multiplied to obtain the node state update semantic coding difference representation vector, wherein the node state update semantic coding feature vector is a column vector; the point sum of the node state update semantic coding and representation vector and the transpose vector of the node state update semantic coding difference representation vector is calculated to obtain the first node state update semantic coding and difference representation vector; the matrix product of the node state update semantic coding and representation matrix and the node state update semantic coding difference representation matrix is ​​calculated, and the transpose vector of the node state update semantic coding feature vector is matrix multiplied with the matrix product to obtain the second node state update semantic coding and difference representation vector; the point sum of the first node state update semantic coding and difference representation vector and the second node state update semantic coding and difference representation vector is calculated to obtain the corrected node state update semantic coding feature vector; the corrected node state update semantic coding feature vector is passed through the decoder-based node description updater to obtain the updated node description of the first node.

[0051] That is, by using the sum representation matrix and difference representation matrix of the group aggregation statistical evaluation of the local distribution of the eigenvalue granularity of the node state update semantic coding feature vector as the retrieval and response distribution enhancement of the sequence of the node state update semantic coding feature vector, and constructing a reference-free distribution retrieval response framework under an open domain based on the group aggregation sum-difference feature distribution of the node state update semantic coding feature vector, the distribution response redundancy caused by the local overflow characteristics of the node state update semantic coding feature vector is avoided through response superposition, so as to realize the fidelity constraint of the self-aggregation statistical correlation of the response of the node state update semantic coding feature vector to the target decoding regression domain, and improve the accuracy of the updated node description of the first node obtained by the node state update semantic coding feature vector through the decoder-based node description updater.

[0052] In summary, the cyberspace geographic map monitoring system based on big data according to the embodiment of the present application is explained, which uses artificial intelligence technology based on deep learning to perform semantic understanding of the node description of the first node in the cyberspace geographic map, extracts the semantic feature representation of the node description, and simultaneously performs time series analysis on the multi-source heterogeneous data of the first node, from which the status information of the power grid node is mined, thereby combining the initial node description and the status information of the power grid node to update the node description. In this way, it is possible to achieve real-time monitoring of the node status in the cyberspace geographic map and real-time updating of the map, timely reflect the latest status of the node in the cyberspace, and improve the updating efficiency and accuracy of the cyberspace geographic map.

[0053] The basic principles of the present invention have been described above in conjunction with specific embodiments. However, it should be noted that the advantages, strengths, and effects mentioned in the present invention are merely illustrative and non-limiting, and should not be construed as necessarily possessed by each embodiment of the present invention. Furthermore, the specific details of the above embodiments are provided for illustrative purposes and to facilitate understanding, and are not intended to be limiting. These details do not necessarily limit the present invention to being implemented using these specific details.

[0054] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, please refer to the relevant description of other embodiments. In the several embodiments provided by the present invention, it should be understood that the disclosed system and method can be implemented in other ways. For example, the system embodiment described above is only schematic. For example, the unit division is only a logical function division, and there may be other division methods in actual implementation. The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical subunits, that is, they may be located in one place, or they may be distributed on multiple network subunits. Some or all of the units may be selected according to actual needs to achieve the purpose of the scheme of this embodiment.

[0055] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims, not the foregoing description, and all variations within the meaning and range of equivalents of the claims are intended to be encompassed therein. Any reference to a figure in a claim should not be construed as limiting the claim to which it relates.

[0056] Finally, it should be noted that the above description has been provided for purposes of illustration and description. Furthermore, the above embodiments are intended only to illustrate the technical solutions of the present invention and are not intended to be limiting. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art will appreciate that the technical solutions of the present invention may be modified or replaced with equivalents without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A cyberspace geographic map monitoring system based on big data, characterized by: include: An initial node description acquisition module is used to extract the initial node description of the first node in the cyberspace geographic map; an initial node description semantic encoding module, configured to perform semantic encoding on the initial node description of the first node to obtain a semantic encoding feature vector of the first node initial description; a multi-source heterogeneous data acquisition module, configured to acquire a time series of multi-source heterogeneous data of the first node collected by the sensor network, wherein the multi-source heterogeneous data includes network traffic data, device status data, and user behavior data; A multi-source heterogeneous data time series encoding module is used to perform parameter sample-level time series encoding on the time series of the multi-source heterogeneous data to obtain a network traffic time series associated implicit feature vector, a device state time series associated implicit feature vector, and a user behavior time series associated implicit feature vector; A power grid node state analysis module, configured to fuse the network traffic time series associated implicit feature vector, the device state time series associated implicit feature vector, and the user behavior time series associated implicit feature vector to obtain a multimodal power grid node state semantic encoding feature vector; A node description updating module is configured to determine an updated node description of the first node based on an adaptive fusion feature of the multimodal power grid node state semantic coding feature vector and the first node initial description semantic coding feature vector.

2. The cyberspace geographic map monitoring system based on big data according to claim 1 is characterized in that: The initial node description semantic encoding module is used to: The initial node description of the first node is passed through a semantic encoder based on the Bert model to obtain a semantic encoding feature vector of the initial description of the first node.

3. The cyberspace geographic map monitoring system based on big data according to claim 2 is characterized in that: The multi-source heterogeneous data temporal encoding module includes: A data regularization unit, configured to regularize the time series of the multi-source heterogeneous data according to a parameter sample dimension to obtain a time series of network traffic data, a time series of device status data, and a time series of user behavior data; A time series encoding unit is used to perform time series analysis on the time series of the network traffic data, the time series of the device status data, and the time series of the user behavior data to obtain the network traffic time series associated implicit feature vector, the device status time series associated implicit feature vector, and the user behavior time series associated implicit feature vector.

4. The cyberspace geographic map monitoring system based on big data according to claim 3 is characterized in that: The time series encoding unit is used to: The time series of the network traffic data, the time series of the device status data and the time series of the user behavior data are respectively passed through a sequence encoder based on an RNN model to obtain the network traffic time series associated implicit feature vector, the device status time series associated implicit feature vector and the user behavior time series associated implicit feature vector.

5. The cyberspace geographic map monitoring system based on big data according to claim 4 is characterized in that: The power grid node status analysis module is used to: The network traffic time series associated implicit feature vector, the device state time series associated implicit feature vector and the user behavior time series associated implicit feature vector are input into a power grid node state analyzer based on a Bayesian probability network to obtain the multimodal power grid node state semantic encoding feature vector.

6. The cyberspace geographic map monitoring system based on big data according to claim 5 is characterized in that: The node description update module includes: an adaptive fusion unit, configured to input the multimodal power grid node state semantic coding feature vector and the first node initial description semantic coding feature vector into a balanced threshold feature vector adaptive fusion network to obtain a node state update semantic coding feature vector; A node description generating unit is configured to input the node state updated semantic encoding feature vector into a decoder-based node description updater to obtain an updated node description of the first node.

7. The cyberspace geographic map monitoring system based on big data according to claim 6 is characterized in that: The adaptive fusion unit includes: a multi-dimensional fusion subunit, configured to pass the multi-modal power grid node state semantic coding feature vector and the first node initial description semantic coding feature vector through a multi-dimensional fusion module to obtain a first power grid node state-initial description semantic fusion feature vector, a second power grid node state-initial description semantic fusion feature vector, and a third power grid node state-initial description semantic fusion feature vector; a balancing threshold value calculation subunit, configured to respectively calculate the balancing threshold values ​​of the first power grid node state-initial description semantic fusion feature vector, the second power grid node state-initial description semantic fusion feature vector, and the third power grid node state-initial description semantic fusion feature vector to obtain a first balancing threshold value, a second balancing threshold value, and a third balancing threshold value; A balancing fusion subunit is configured to perform weighted fusion on the multimodal power grid node state semantic coding feature vector and the first node initial description semantic coding feature vector based on the first balancing threshold value, the second balancing threshold value, and the third balancing threshold value to obtain the node state update semantic coding feature vector.

8. The cyberspace geographic map monitoring system based on big data according to claim 7 is characterized in that: The multi-dimensional fusion subunit is used to: Cascading the multimodal power grid node state semantic encoding feature vector and the first node initial description semantic encoding feature vector to obtain the first power grid node state-initial description semantic fusion feature vector; Adding the multimodal power grid node state semantic encoding feature vector and the first node initial description semantic encoding feature vector by position to obtain the second power grid node state-initial description semantic fusion feature vector; The multimodal power grid node state semantic encoding feature vector and the first node initial description semantic encoding feature vector are multiplied by position point to obtain the third power grid node state-initial description semantic fusion feature vector.

9. The cyberspace geographic map monitoring system based on big data according to claim 8 is characterized in that: The equalization threshold value calculation subunit is used to: Multiplying the first power grid node state-initial description semantic fusion feature vector by a first predetermined transformation vector to obtain a first threshold scoring coefficient; The first equalization threshold value is obtained by adding the first threshold scoring coefficient and the first bias parameter and then performing a sigmoid activation function.

10. The cyberspace geographic map monitoring system based on big data according to claim 9 is characterized in that: The balanced fusion subunit is used to: determining a first weight parameter and a second weight parameter based on the first equalization threshold value, the second equalization threshold value, and the third equalization threshold value, wherein the first weight parameter is an average of the first equalization threshold value, the second equalization threshold value, and the third equalization threshold value, and the second weight parameter is a difference between one and the first weight parameter; Performing position-wise weighting on the multimodal power grid node state semantic coding feature vector using the first weight parameter to obtain a weighted multimodal power grid node state semantic coding feature vector, and performing position-wise weighting on the first node initial description semantic coding feature vector using the second weight parameter to obtain a weighted first node initial description semantic coding feature vector; The node state update semantic coding feature vector is obtained by adding corresponding elements of the weighted multimodal power grid node state semantic coding feature vector and the weighted first node initial description semantic coding feature vector.

Images