CN117556305A - Network node identification method, system and electronic equipment - Google Patents
Network node identification method, system and electronic equipment Download PDFInfo
- Publication number
- CN117556305A CN117556305A CN202311599043.1A CN202311599043A CN117556305A CN 117556305 A CN117556305 A CN 117556305A CN 202311599043 A CN202311599043 A CN 202311599043A CN 117556305 A CN117556305 A CN 117556305A
- Authority
- CN
- China
- Prior art keywords
- node
- network
- nodes
- key
- identified
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 52
- 238000010586 diagram Methods 0.000 claims abstract description 63
- 238000012360 testing method Methods 0.000 claims abstract description 39
- 230000004931 aggregating effect Effects 0.000 claims abstract description 7
- 230000006870 function Effects 0.000 claims description 68
- 238000013507 mapping Methods 0.000 claims description 27
- 239000013598 vector Substances 0.000 claims description 19
- 230000008569 process Effects 0.000 claims description 18
- 239000011159 matrix material Substances 0.000 claims description 16
- 238000004590 computer program Methods 0.000 claims description 10
- 230000002776 aggregation Effects 0.000 claims description 8
- 238000004220 aggregation Methods 0.000 claims description 8
- 238000010276 construction Methods 0.000 claims description 7
- 238000009826 distribution Methods 0.000 claims description 5
- 230000003993 interaction Effects 0.000 claims description 5
- 230000001186 cumulative effect Effects 0.000 claims description 3
- 238000012163 sequencing technique Methods 0.000 claims description 3
- 238000003860 storage Methods 0.000 description 8
- 238000012549 training Methods 0.000 description 8
- 238000012545 processing Methods 0.000 description 5
- 230000009471 action Effects 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 4
- 238000005259 measurement Methods 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000005070 sampling Methods 0.000 description 3
- 238000012935 Averaging Methods 0.000 description 2
- 230000007613 environmental effect Effects 0.000 description 2
- 239000004973 liquid crystal related substance Substances 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000002787 reinforcement Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000008901 benefit Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000009044 synergistic interaction Effects 0.000 description 1
Abstract
The embodiment of the invention provides a network node identification method, a network node identification system and electronic equipment, and belongs to the field of Internet. The method comprises the following steps: constructing a network node diagram data set to be identified; inputting a network node diagram data set to be identified into a pre-constructed network key node identification model for testing, and obtaining a test score corresponding to each node, wherein the network key node identification model is constructed based on a Q function corresponding to a multi-layer parameterized quantum circuit for aggregating neighbor node information; and determining key nodes in the network node diagram data set to be identified according to the test scores corresponding to the nodes. The network node graph data set to be identified is tested through the network key node identification model, the test score of each node in the graph is obtained, the key nodes in the graph are determined based on the test scores of each node, the key nodes can be determined quickly, the accuracy is higher, and the application range of the built network key node identification model is wider.
Description
Technical Field
The invention relates to the technical field of Internet, in particular to a network node identification method, a network node identification system and electronic equipment.
Background
With the development of internet technology, people have been living in a world that is full of various complex networks, such as biological networks, hydroelectric networks, securities trading networks, social networks, etc. Mining research on key nodes in a network is one of main research directions in a complex network, and aims to discover nodes playing a key role in the processes of network structure, information transmission and the like.
The number of critical nodes is typically very small, but its impact can quickly reach most nodes in the network.
The measurement of the influence of the prior art on the nodes is mainly based on network topology information such as position, neighborhood, path and the like. However, determining key nodes in a large-scale network faces the dilemma that "accuracy and high efficiency" cannot be achieved, and some sorting methods based on network global information are too time-complex to popularize and use.
Disclosure of Invention
An object of an embodiment of the present invention is to provide a method, a system and an electronic device for identifying a network node, which are used for solving all or at least part of the technical problems existing in the prior art.
In order to achieve the above object, an embodiment of the present invention provides a method for identifying a network key node, including: constructing a network node diagram data set to be identified;
inputting the network node diagram data set to be identified into a pre-constructed network key node identification model for testing, and obtaining a test score corresponding to each node, wherein the network key node identification model is constructed based on a Q function corresponding to a multi-layer parameterized quantum circuit for aggregating neighbor node information;
and determining key nodes in the network node diagram data set to be identified according to the test scores corresponding to the nodes.
Optionally, the constructing a network node map data set to be identified includes:
generating a network node diagram according to the power law distribution, and dividing each network node diagram into a network node diagram data set to be identified containing a specified number of nodes according to breadth priority.
Optionally, the construction process of the network key node identification model includes:
mapping nodes in the network node diagram data set to be identified into quantum states based on a mapping quantum circuit diagram;
constructing a multi-layer parameterized quantum circuit based on aggregation neighbor node information based on a multi-layer message passing network;
and constructing a Q function corresponding to the multi-layer parameterized quantum circuit, and interacting with the network node diagram to be identified to obtain a network key node identification model.
Optionally, the construction of the mapping quantum circuit diagram includes:
setting initial characteristics of the nodes according to the statistical rule of the nodes, and randomly setting initial rotation parameter vectors;
constructing an initial mapping quantum circuit diagram based on the initial rotation parameter vector and the initial characteristic;
respectively determining a Euclidean distance correlation matrix of a network node diagram and a Hilbert space distance correlation matrix based on quantum state mapping, and constructing a loss function based on the Euclidean distance correlation matrix and the Hilbert space distance correlation matrix;
adjusting initial rotation parameters according to the loss function, and determining a target rotation parameter vector when the loss function meets a preset requirement;
and determining a mapping quantum circuit diagram based on the target rotation parameter vector and the initial characteristics of the nodes, wherein one mapping quantum circuit corresponds to one node.
Optionally, a multi-layer parameterized quantum circuit is constructed according to the following formula:
in the method, in the process of the invention,representing the amount of node v at layer tSub-state feature representation, U 1 Representing the quantum gate parameters used by node v,quantum state characteristic representation of the representation node v at the t-1 layer,>quantum state characteristic representation of neighbor node mu of representing node v in t-1 layer, U 2 Quantum gate parameters representing node μ use, operator +.>Representing tensor product operations.
Optionally, constructing a Q function corresponding to the multi-layer parameterized quantum circuit, and interacting with a network node diagram to be identified to obtain a network key node identification model, including:
respectively determining a first Q value corresponding to each line in the multi-layer parameterized quantum circuit, sequencing all nodes according to the first Q value corresponding to each line, and determining key nodes, wherein the larger the Q value is, the more key the corresponding node is represented;
removing the key nodes, and redefining a second Q value of each line corresponding to each node except the key nodes until the network connectivity metric index reaches a preset value after the nodes are removed;
constructing a multi-layer parameterized quantum circuit with the same initial parameters as the Q function corresponding to the multi-layer parameterized quantum circuit as a target Q function;
in the interaction process, the parameters of the target Q function are updated with the parameters of the Q function every certain round, wherein a loss function is constructed by the Q value of the Q function and the Q value of the target Q function.
Optionally, the removed node is determined to be a key node according to the following formula:
wherein R represents the cumulative connectivity, N is the number of nodes,v i representing the ith removed node, σ is a connectivity metric function.
Optionally, the error formula is as follows:
in the formula Θ E ＝{U 1 ，U 2 }，Θ D ＝{U y ，U z }，Θ ED ＝{Θ E ，Θ D }，s t The environment state at time t is represented, and the Q value of the target Q function is represented asWherein r is t Represents the prize value obtained at time t, gamma E [0,1 ]]Representing a parameter for measuring importance of future rewards,(s) t ，a t ，r t，t+n ，s t，t+n ) U (B) represents randomly taking samples from the buffer.
In another aspect, the present invention provides a system for identifying a network key node, including:
the construction unit is used for constructing a network node diagram data set to be identified;
the testing unit is used for inputting the network node diagram data set to be identified into a pre-constructed network key node identification model for testing, and obtaining a test score corresponding to each node, wherein the network key node identification model is constructed based on a Q function corresponding to a multi-layer parameterized quantum circuit for aggregating neighbor node information;
and the identification unit is used for determining key nodes in the network node diagram data set to be identified according to the test scores corresponding to the nodes.
In another aspect, the present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor implements the steps of the identification method described above when the processor executes the program.
The network key node identification model constructed by the Q function corresponding to the multi-layer parameterized quantum circuit based on the aggregated neighbor node information is used for testing the network node graph data set to be identified, so that the test score of each node in the graph is obtained, the key nodes in the graph are determined based on the test scores of each node, the key nodes can be rapidly determined, the accuracy is higher, and the application range of the constructed network key node identification model is wider.
Additional features and advantages of embodiments of the invention will be set forth in the detailed description which follows.
Drawings
The accompanying drawings are included to provide a further understanding of embodiments of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain, without limitation, the embodiments of the invention. In the drawings:
fig. 1 is a flowchart of an implementation of a method for identifying a network key node according to an embodiment of the present invention;
FIG. 2 is a quantum circuit diagram featuring a quantum state for a mapping node according to an embodiment of the present invention;
FIG. 3 is a quantum circuit for aggregating first layer neighbor information provided by an embodiment of the present invention;
FIG. 4 is a diagram of a decoding quantum circuit provided by an embodiment of the present invention;
FIG. 5 is a schematic diagram of a process for interacting with an environment according to an embodiment of the present invention;
fig. 6 is a flowchart of an implementation of another method for identifying network key nodes according to an embodiment of the present invention;
fig. 7 is a schematic structural diagram of an identification system of a network key node according to an embodiment of the present invention.
Detailed Description
The following describes the detailed implementation of the embodiments of the present invention with reference to the drawings. It should be understood that the detailed description and specific examples, while indicating and illustrating the invention, are not intended to limit the invention.
Referring to fig. 1, a flowchart of an implementation of a method for identifying a network key node according to an embodiment of the present invention includes the following implementation steps:
step 100: and constructing a network node diagram data set to be identified.
In some embodiments, the constructing a network node map data set to be identified includes:
generating a network node diagram according to the power law distribution, and dividing each network node diagram into a network node diagram data set to be identified containing a specified number of nodes according to breadth priority.
It should be noted that the specified number may be set according to a specific application scenario, for example, 12, which is not limited herein.
In some embodiments, a training graph dataset is generated from a power law distribution and each graph is preferentially divided in breadth into sub-graphs of 12 nodes: first, 100 networks are generated which conform to the power law distribution, and the network scale is 30-50 nodes. An initial feature representation is set for each node, e.g. the dimension of the initial feature is set to (1, 5), each component in turn representing: the centrality of the nodes, the centrality of the feature vector, the centrality of the medium number, the centrality of the proximity and the clustering coefficient are set, the network G (V, E) comprises N nodes, the nodes are randomly sampled in each generated graph, and the sampling number is thatWherein N is the number of nodes of the graph. Then, the breadth first extraction is carried out on 11 neighbors by taking the sampling node as the center, and the 11 neighbors and the sampling node are combined into a sub-graph containing 12 nodes.
In some implementations, the node's centrality (Degree Centrality) is considered to be more important, the more neighbors they own, the node i's centrality (DC i ) The definition is as follows:
wherein k is i The degree of the node i is given, and N is the number of nodes.
In some embodiments, the more important a node's feature vector centrality (Eigenvector Centrality) considers a connected neighbor node, the more important that node is, the feature vector centrality (EC i ) The definition is as follows:
wherein c is a proportionality constant, a ij Is the element of row j of the network adjacency matrix.
In some embodiments, the betweenness (Betweenness Centrality) of a node considers that if a node is on the shortest path between many other nodes, that node is important, the betweenness (BC i ) The definition is as follows:
wherein delta st Expressed as the total number of shortest paths from node s to node t, delta st (v i ) Representing the transit of node v in these shortest paths i Is provided for the number of paths of the network.
In some embodiments, proximity centrality of nodes (Closeness Centrality) if the shortest path length of a node to all other nodes is the more important this node is, node v is calculated first i Average distance to all other nodes in the network:
where N is the number of nodes in the network, d (v i ，v j ) Representing the shortest distance between nodes i and j. Smaller node i indicates closer proximity to other nodes in the network, and thus proximity centrality (CC i ) Defined as D a v g (v i ) Is the reciprocal of (2):
in some embodiments, the greater the proportion of nodes' neighbors interconnected by the Clustering coefficient (Clustering) of a node, the more important that node. Let the degree of node i be k i Its cluster coefficient (C i ) The definition is as follows:
wherein e i Is the number of edges between node i neighbors.
From this, the initial characteristics of node i can be obtained
Step 101: and inputting the network node diagram data set to be identified into a pre-constructed network key node identification model for testing, and obtaining a test score corresponding to each node.
The network key node identification model is constructed based on Q functions corresponding to multi-layer parameterized quantum circuits for aggregating neighbor node information.
In some embodiments, the process of constructing the network key node identification model includes the steps of:
s1: and mapping the nodes in the network node diagram data set to be identified into quantum states based on the mapping quantum circuit diagram.
In some embodiments, the step of constructing the map quantum wire map includes:
SA: setting initial characteristics of the nodes according to the statistical rule of the nodes, and randomly setting initial rotation parameter vectors;
SB: and constructing an initial mapping quantum circuit diagram based on the initial rotation parameter vector and the initial characteristic.
SC: and respectively determining a Euclidean distance correlation matrix of the network node diagram and a Hilbert space distance correlation matrix based on quantum state mapping, and constructing a loss function based on the Euclidean distance correlation matrix and the Hilbert space distance correlation matrix.
SD: and adjusting initial rotation parameters according to the loss function, and determining a target rotation parameter vector when the loss function meets a preset requirement.
In some embodiments, the initial rotation parameter vector is set randomlyThe dimension and the initial characteristic of the node are the same as (1, 5), let the initial characteristic of the node i be +.>The quantum circuit diagram of the node i mapping section is shown in fig. 2, in which the input of the circuit is quantum state |0>RX represents the rotation gate around X-axis in the quantum gate,>and->The i-th component of the initial characteristic of the node i and the i-th component of the initial rotation parameter are respectively represented, and the output is the quantum state mapping of the node i>Calculating Euclidean distance correlation matrix D of graph, setting +.>For the initial feature of node i, the calculation formula is as follows:
in the method, in the process of the invention,for the initial feature of node i +.>Representing the initial characteristics of node j, dij representing the Euclidean distance correlation matrix of node i and node j,/>Two norms representing the initial characteristics of node j, +.>Representing the two norms of the node i initial characteristics.
Calculating a Hilbert space (Hilbert) distance correlation matrix D' based on quantum state mapping, and settingFor the quantum state mapping of node i, the calculation formula is as follows:
in the method, in the process of the invention,quantum state mapping for node i initial feature, +.>And (5) quantum state mapping of initial characteristics of the node j.
Calculation of loss adjustment initial rotation parameter vectorThe loss function is defined as l=sum (|d-D' |), where sum is a per-element summation function, using a derivative-free optimization method based on interpolation model UOBYQA to find the optimal rotation parameter vector ++>(i.e., the target rotation parameter vector) such that the loss function is minimized.
SE: and determining a mapping quantum circuit diagram based on the target rotation parameter vector and the initial characteristics of the nodes, wherein one mapping quantum circuit corresponds to one node.
S2: based on the multi-layer message passing network, a multi-layer parameterized quantum circuit based on the information of the aggregation neighbor nodes is constructed.
In some embodiments, a multi-layer messaging network is constructed on a quantum wire, node characteristic representation and whole graph representation in Hilbert space after generating aggregated neighbor information, and according to the multi-layer messaging network, the quantum circuit is constructed to realize the aggregated neighbor information, and the mathematical expression of the part is as follows:
wherein, the liquid crystal display device comprises a liquid crystal display device,representation of quantum state characteristics of node v at t layer, U 1 Representing the quantum gate parameters used by node v,quantum state characteristic representation of neighbor node mu of representing node v in t-1 layer, U 2 Quantum gate parameters representing node μ use, operator +.>Representing tensor product operations. U (U) 1 And U 2 Together constitute training parameters of the coding part.
In some embodiments, (1) a quantum wire that aggregates first layer neighbor information is shown in FIG. 3, where U init Part of the circuit corresponds to fig. 2, each line represents a node, and the nodes of all lines share parametersRX, RY and RZ are quantum gates rotating around the X axis, the Y axis and the Z axis respectively, and the rotating gate parameter corresponding to the node v is U 1 The rotation gate parameter corresponding to the rest neighbor node mu epsilon N (v) is U 2 。U ent The operation gate in (a) is a CNOT gate, so that the nodes in the system are entangled. Fig. 3 constitutes a quantum system corresponding to a sub-graph of node v and its one-node neighbors. The output of the system is node v to aggregate the first layer neighborsAnd representing quantum state characteristics after information. (2) And generating the input of a second layer aggregation line according to the quantum state of the node v and the neighbor mu E N (v) after the aggregation of the first layer information. First layer input of node v>The out dimension is the reference, if the neighboring node muThe dimension is smaller than node v, then in +.>Then 0 is added to obtain->Make->And->Is the same in dimension; if +.>The dimension is larger than the dimension of the node v, then dividing the dimension of the node μ by the dimension of the node v to obtain a division d, and then +.>Dividing into several equal-length blocks according to interval size d, averaging the blocks, and averaging each block to form a block with +.>New tensor ∈of equal length>If->Dimension and->Identical->Is unchanged. All->Or->mu.epsilon.N (v), and ∈N (v)>Cumulatively sum to get +.>The third layer is the same as the input of the second layer aggregation line. (3) And measuring each line along the Z axis on the last layer of aggregation line, wherein the measurement result is used as the identification attribute of the corresponding node. (4) Global nodes are added to obtain a quantum state representation of the whole graph. Creating a global node which is connected with all nodes in the graph, but does not comprise global nodes in the neighbor set of other nodes, carrying out neighbor aggregation on the global node, repeating the steps (1) and (2), and outputting the aggregated multi-layer neighbor nodes->Dimension ratio graph (total 12 nodes) is 2 more 1 The output dimension is reduced by a factor of 12 in the manner of (2) so that the dimension corresponds to 12 quantum wires, i.e., 12 nodes.
S3: and constructing a Q function corresponding to the multi-layer parameterized quantum circuit, and interacting with the network node diagram to be identified to obtain a network key node identification model.
In some embodiments, when step S3 is performed, the following steps may be specifically performed:
s30: and respectively determining a first Q value corresponding to each line in the multi-layer parameterized quantum circuit, sequencing all nodes according to the first Q value corresponding to each line, and determining key nodes, wherein the larger the Q value is, the more key the corresponding node is represented.
S31: and removing the key nodes, and redetermining the second Q value of each line corresponding to each node except the key nodes until the network connectivity metric index reaches a preset value after the nodes are removed.
In some implementations, the node ordering problem is modeled as a reinforcement learning solvable problem, where reinforcement learning consists of an environmental State (State) S, an Action (Action) A taken according to the environment, and a Reward (Reward) R obtained after taking the Action. The quantum state characteristic of the rest network is used as an environmental state S in the node ordering problem, the quantum state characteristic of the node to be removed is used as an action space A which can be taken, the reduction amount of the accumulated connectivity of the network after the node is removed is used as a reward R, and the removed node is determined to be a key node according to the following formula:
wherein R represents the cumulative connectivity, N is the number of nodes, v i Representing the ith removed node, σ is a connectivity metric function, with larger values of R representing more important nodes.
In some embodiments, the circuit design of the multi-layer parameterized quantum circuit approximates a Q function is as follows: this part is a decoding circuit, i.e. the quantum states are decoded into scalar quantities, the scalar quantities are used for ordering, the decoding quantum circuit diagram is shown in figure 4, and the input part is thatAs the input of decoding quantum circuit, using node identification attribute as the parameter of initializing part RX gate of every circuit to establish one-to-one correspondence between circuit and node and construct three-layer hidden layer network formed from RY and RZ gates, the parameter of RY gate is formed from U y Controlling parameters of RZ gate by U z Control U y And U z Together constitute training parameters of the decoding section. U in layer after hidden layer network construction is completed ent The nodes are entangled. And measuring each line along the X-axis direction in sequence, wherein the measurement result is the Q value in Q-learning, and the nodes are ordered according to the result.
S32: constructing a multi-layer parameterized quantum circuit with the same initial parameters as the Q function corresponding to the multi-layer parameterized quantum circuit as a target Q function;
s33: in the interaction process, the parameters of the target Q function are updated with the parameters of the Q function every certain round, wherein a loss function is constructed by the Q value of the Q function and the Q value of the target Q function.
Specifically, the loss function is as follows:
in the formula Θ E ＝{U 1 ，U 2 }，Θ D ＝{U y ，U z }，Θ ED ＝{Θ E ，Θ D }，s t The environment state at time t is represented, and the Q value of the target Q function is represented asWherein r is t Represents the prize value obtained at time t, gamma E [0,1 ]]Representing a parameter for measuring importance of future rewards,(s) t ，a t ，r t，t+n ，s t，t+n ) U (B) represents randomly taking samples from the buffer.
In some embodiments, the trainable parameter is divided into two parts, the trainable parameter of the encoded part is Θ E ＝{U 1 ，U 2 Trainable parameters of the decoding part are Θ D ＝{U y ，U z }. Errors generated by the code part are measured by node identification attributes, and the nodes with connected edges are closer to each other in the identification attributes. The error of the decoding part is generated by a delay updating mechanism of the depth Q function, namely, another network with the same structure as that of fig. 4 and the same initial parameters is constructed as a target Q function, the initial parameters are the same as the Q function updated in each step, and the parameters of the target Q function are updated by the parameters of the Q function at regular step length. The Q value produced after the Q function measurement is expressed as Q (s t ，a t ) Wherein s is t Expressed as the state of the environment at time t, i.e. the target Q value produced by the target Q function is expressed asWherein r is t Marking the prize value obtained at time t, gamma e 0,1]Is a parameter for measuring the importance of future rewards. Thus, the formula for the overall error for a round of training is as follows:
wherein,(s) t ，a t ，r t，t+n ，s t，t+n ) U (B) represents randomly extracting samples from the buffer, thereby reducing the correlation of the samples, y i Representing the measured identity attribute of node i.
In some embodiments, the schematic diagram of the interaction process with the environment is shown in fig. 5, and after removing the node with the largest q value, the interaction process is stopped after reaching the termination condition. The above steps are repeated on all training chart datasets. After training is finished, using the obtained theta E And theta (theta) D And assigning values to the test Q-network, testing the test chart data, and giving out node importance ranking.
Step 102: and determining key nodes in the network node diagram data set to be identified according to the test scores corresponding to the nodes.
In some embodiments, referring to fig. 6, a flowchart of another method for identifying a network key node according to an embodiment of the present invention is shown:
s600: generating a training graph data set and dividing the graph into subgraphs with 12 nodes according to breadth first;
s601: mapping the node into a quantum state according to the initial characteristics of the node;
s602: making a characteristic representation and a graph representation of the multi-layer message transmission generation node on the quantum circuit;
s603: constructing a multi-layer parameterized quantum circuit approximate Q function;
s604: the Q function approximated by the quantum circuit is used for interactively adjusting network parameters with the original network;
s605: repeating the operations of steps S601-S604 for all training graph datasets;
s606: testing is conducted on the test chart data set, and nodes are ordered according to the test scores.
The network key node identification model constructed by the Q function corresponding to the multi-layer parameterized quantum circuit based on the aggregated neighbor node information is used for testing the network node graph data set to be identified, so that the test score of each node in the graph is obtained, the key nodes in the graph are determined based on the test scores of each node, the key nodes can be rapidly determined, the accuracy is higher, and the application range of the constructed network key node identification model is wider.
Referring to fig. 7, a schematic structural diagram of an identification system of a network key node according to an embodiment of the present invention includes:
a construction unit 700 for constructing a network node map data set to be identified.
The testing unit 701 is configured to input the network node diagram data set to be identified to a pre-constructed network key node identification model for testing, and obtain a test score corresponding to each node, where the network key node identification model is constructed based on a Q function corresponding to a multi-layer parameterized quantum circuit that aggregates neighbor node information.
And the identifying unit 702 is configured to determine key nodes in the network node graph data set to be identified according to the test scores corresponding to the nodes.
Through the mutual synergistic interaction among the construction unit, the testing unit and the identification unit, the identification system of the network key node has higher processing speed and higher accuracy, and further the application range of the system is widened.
In another aspect, an embodiment of the present invention further provides an electronic device, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor, where the steps of the identification method according to any one of the foregoing embodiments are implemented when the processor executes the program.
It will be appreciated by those skilled in the art that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
In one typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.
The memory may include volatile memory in a computer-readable medium, random Access Memory (RAM) and/or nonvolatile memory, etc., such as Read Only Memory (ROM) or flash RAM. Memory is an example of a computer-readable medium.
Computer readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of storage media for a computer include, but are not limited to, phase change memory (PRAM), static Random Access Memory (SRAM), dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium, which can be used to store information that can be accessed by a computing device. Computer-readable media, as defined herein, does not include transitory computer-readable media (transmission media), such as modulated data signals and carrier waves.
It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article or apparatus that comprises an element.
The foregoing is merely exemplary of the present application and is not intended to limit the present application. Various modifications and changes may be made to the present application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc. which are within the spirit and principles of the present application are intended to be included within the scope of the claims of the present application.
Claims (10)
1. The method for identifying the network key node is characterized by comprising the following steps of:
constructing a network node diagram data set to be identified;
inputting the network node diagram data set to be identified into a pre-constructed network key node identification model for testing, and obtaining a test score corresponding to each node, wherein the network key node identification model is constructed based on a Q function corresponding to a multi-layer parameterized quantum circuit for aggregating neighbor node information;
and determining key nodes in the network node diagram data set to be identified according to the test scores corresponding to the nodes.
2. The method of claim 1, wherein constructing the network node map data set to be identified comprises:
generating a network node diagram according to the power law distribution, and dividing each network node diagram into a network node diagram data set to be identified containing a specified number of nodes according to breadth priority.
3. The identification method according to claim 1, wherein the process of constructing the network key node identification model includes:
mapping nodes in the network node diagram data set to be identified into quantum states based on a mapping quantum circuit diagram;
constructing a multi-layer parameterized quantum circuit based on aggregation neighbor node information based on a multi-layer message passing network;
and constructing a Q function corresponding to the multi-layer parameterized quantum circuit, and interacting with the network node diagram to be identified to obtain a network key node identification model.
4. The method of claim 3, wherein the constructing of the map quantum wire map comprises:
setting initial characteristics of the nodes according to the statistical rule of the nodes, and randomly setting initial rotation parameter vectors;
constructing an initial mapping quantum circuit diagram based on the initial rotation parameter vector and the initial characteristic;
respectively determining a Euclidean distance correlation matrix of a network node diagram and a Hilbert space distance correlation matrix based on quantum state mapping, and constructing a loss function based on the Euclidean distance correlation matrix and the Hilbert space distance correlation matrix;
adjusting initial rotation parameters according to the loss function, and determining a target rotation parameter vector when the loss function meets a preset requirement;
and determining a mapping quantum circuit diagram based on the target rotation parameter vector and the initial characteristics of the nodes, wherein one mapping quantum circuit corresponds to one node.
5. A method of identification as claimed in claim 3, wherein the multi-layer parameterized quantum circuit is constructed according to the following formula:
in the method, in the process of the invention,representation of quantum state characteristics of node v at t layer, U 1 Quantum gate parameters used for representing node v, < +.>Quantum state characteristic representation of the representation node v at the t-1 layer,>quantum state characteristic representation of neighbor node mu of representing node v in t-1 layer, U 2 Representing node muQuantum gate parameters used, operator->Representing tensor product operations.
6. The identification method of claim 3, wherein constructing a Q function corresponding to the multi-layer parameterized quantum circuit and interacting with a network node diagram to be identified to obtain a network key node identification model comprises:
respectively determining a first Q value corresponding to each line in the multi-layer parameterized quantum circuit, sequencing all nodes according to the first Q value corresponding to each line, and determining key nodes, wherein the larger the Q value is, the more key the corresponding node is represented;
removing the key nodes, and redefining a second Q value of each line corresponding to each node except the key nodes until the network connectivity metric index reaches a preset value after the nodes are removed;
constructing a multi-layer parameterized quantum circuit with the same initial parameters as the Q function corresponding to the multi-layer parameterized quantum circuit as a target Q function;
in the interaction process, the parameters of the target Q function are updated with the parameters of the Q function every certain round, wherein a loss function is constructed by the Q value of the Q function and the Q value of the target Q function.
7. The identification method of claim 6, wherein the removed node is determined to be a key node according to the following formula:
wherein R represents the cumulative connectivity, N is the number of nodes, v i Representing the ith removed node, σ is a connectivity metric function.
8. The identification method of claim 6, wherein the loss function is as follows:
in the formula Θ E ＝{U 1 ,U 2 }，Θ D ＝{U y ,U z }，Θ E2 ＝{Θ E ,Θ D }，s t The environment state at time t is represented, and the Q value of the target Q function is represented asWherein r is t Represents the prize value obtained at time t, gamma E [0,1 ]]Representing a parameter for measuring importance of future rewards,(s) t ,a t ,r t,t+n ,s t,t+n ) U (B) represents randomly taking samples from the buffer.
9. A system for identifying network key nodes, comprising:
the construction unit is used for constructing a network node diagram data set to be identified;
the testing unit is used for inputting the network node diagram data set to be identified into a pre-constructed network key node identification model for testing, and obtaining a test score corresponding to each node, wherein the network key node identification model is constructed based on a Q function corresponding to a multi-layer parameterized quantum circuit for aggregating neighbor node information;
and the identification unit is used for determining key nodes in the network node diagram data set to be identified according to the test scores corresponding to the nodes.
10. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the steps of the identification method according to any one of claims 1-8 when the program is executed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311599043.1A CN117556305A (en) | 2023-11-27 | 2023-11-27 | Network node identification method, system and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311599043.1A CN117556305A (en) | 2023-11-27 | 2023-11-27 | Network node identification method, system and electronic equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117556305A true CN117556305A (en) | 2024-02-13 |
Family
ID=89810749
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311599043.1A Pending CN117556305A (en) | 2023-11-27 | 2023-11-27 | Network node identification method, system and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117556305A (en) |
-
2023
- 2023-11-27 CN CN202311599043.1A patent/CN117556305A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR101940029B1 (en) | Anomaly detection | |
| KR102295805B1 (en) | Method for managing training data | |
| CN112567396A (en) | Computing excited state properties of molecular systems with hybrid classical-quantum computing systems | |
| CN109905269A (en) | The method and apparatus for determining network failure | |
| CN116245015A (en) | Data change trend prediction method and system based on deep learning | |
| CN109086291B (en) | Parallel anomaly detection method and system based on MapReduce | |
| CN114021728B (en) | Quantum data measuring method and system, electronic device, and medium | |
| CN112580733B (en) | Classification model training method, device, equipment and storage medium | |
| AU2022263561A1 (en) | Quantum measurement device calibration method and apparatus, electronic device and medium | |
| CN113516246A (en) | Parameter optimization method, quantum chip control method and device | |
| US8327196B2 (en) | Identifying an optimized test bit pattern for analyzing electrical communications channel topologies | |
| JP2022088556A (en) | Calibration method and device of quantum measurement apparatus, electronic apparatus and medium | |
| CN113448821B (en) | Method and device for identifying engineering defects | |
| JP2022549407A (en) | Methods and systems for identification and analysis of regime shifts | |
| CN115865607A (en) | Distributed training computing node management method and related device | |
| CN117556305A (en) | Network node identification method, system and electronic equipment | |
| CN114900435B (en) | Connection relation prediction method and related equipment | |
| CN115730663A (en) | Quantum computing task mapping method and quantum computer operating system | |
| US11734484B2 (en) | Method for automating semiconductor design based on artificial intelligence | |
| WO2022199261A1 (en) | Model recommendation method and apparatus, and computer device | |
| Huo | Analysis of the Weighted Graph Shortest Path Problem: Algorithms, Applications, and Challenges | |
| CN116187464B (en) | Blind quantum computing processing method and device and electronic equipment | |
| WO2022021199A1 (en) | Neural network model construction method and device therefor | |
| WO2023212576A2 (en) | Local low-rank response imputation for automatic configuration of contextualized artificial intelligence | |
| Petrov | Understanding the Performance of Hyperbolic Graph Neural Networks |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |