CN117729545A - 5G network communication control method - Google Patents
5G network communication control method Download PDFInfo
- Publication number
- CN117729545A CN117729545A CN202410179104.7A CN202410179104A CN117729545A CN 117729545 A CN117729545 A CN 117729545A CN 202410179104 A CN202410179104 A CN 202410179104A CN 117729545 A CN117729545 A CN 117729545A
- Authority
- CN
- China
- Prior art keywords
- registration information
- feature vector
- semantic understanding
- understanding feature
- semantic
- 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.)
- Granted
Links
- 238000004891 communication Methods 0.000 title claims abstract description 60
- 238000000034 method Methods 0.000 title claims abstract description 57
- 238000011217 control strategy Methods 0.000 claims abstract description 24
- 239000013598 vector Substances 0.000 claims description 240
- 230000004927 fusion Effects 0.000 claims description 42
- 238000005457 optimization Methods 0.000 claims description 13
- 238000013527 convolutional neural network Methods 0.000 claims description 11
- 230000002457 bidirectional effect Effects 0.000 claims description 10
- 125000004122 cyclic group Chemical group 0.000 claims description 10
- 230000007774 longterm Effects 0.000 claims description 10
- 230000008569 process Effects 0.000 claims description 7
- 230000011218 segmentation Effects 0.000 claims description 7
- 230000001105 regulatory effect Effects 0.000 claims description 5
- 238000007499 fusion processing Methods 0.000 claims description 3
- 238000005516 engineering process Methods 0.000 abstract description 6
- 238000003058 natural language processing Methods 0.000 abstract description 6
- 238000013135 deep learning Methods 0.000 abstract description 5
- 230000006399 behavior Effects 0.000 description 9
- 230000006870 function Effects 0.000 description 3
- 238000013507 mapping Methods 0.000 description 3
- 238000011176 pooling Methods 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000001276 controlling effect Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 238000007726 management method Methods 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 230000002776 aggregation Effects 0.000 description 1
- 238000004220 aggregation Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks
Landscapes
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The application relates to the technical field of network communication, and particularly discloses a 5G network communication control method, which adopts a natural language processing technology based on deep learning to perform semantic understanding on registration information of a target user, captures semantic feature expression of the registration information of the target user, and digs out potential risk information in the registration information, so that a corresponding control strategy is intelligently selected to control the target user. Thus, intelligent control strategy selection and implementation can be realized, and effective decision support is provided for 5G network communication control.
Description
Technical Field
The present application relates to the field of network communications technologies, and in particular, to a 5G network communications policing method.
Background
With the popularity of 5G networks, network communication management is becoming an important issue. The 5G network connects a large number of terminal devices and application systems, and personal information and privacy of users are at higher risk. By controlling the 5G network user communication, malicious behaviors such as network attack, phishing, malicious software propagation and the like can be monitored and prevented, so that the user can be protected from being infringed by network crimes, and network security is maintained. In addition, through controlling user communication, the behavior of the user can be supervised and managed, improper behavior is prevented, and the user communication behavior is ensured to meet the requirements of laws and regulations and regulatory authorities.
In this regard, patent CN110740489a discloses a 5G network communication control method, which identifies a target user according to a control command and information such as a user identifier in a registration message by sending a re-registration message to the target user, so as to determine a control user that needs to be controlled in the target user in a control area, thereby realizing precise differentiation between a control user and a non-control user in the control area, and effectively improving communication control accuracy. Specifically, after confirming that the target user is a control user, the target user needs to be controlled by selecting a corresponding control policy according to a registration message of the control user. However, the conventional network communication control method is often formulated based on general rules and standards, and cannot fully consider specific situations of different users, and cannot perform personalized control according to information of specific users, which may cause excessive control in some situations, bring inconvenience and unnecessary limitation to users, and insufficient control in other situations, and cannot effectively restrict behaviors of users.
Therefore, an optimized 5G network communication policing method is desired.
Disclosure of Invention
The present application has been made in order to solve the above technical problems. The embodiment of the application provides a 5G network communication control method, which adopts a natural language processing technology based on deep learning to perform semantic understanding on registration information of a target user, captures semantic feature expression of the registration information of the target user, and digs out potential risk information in the registration information, thereby intelligently selecting a corresponding control strategy to control the target user. Thus, intelligent control strategy selection and implementation can be realized, and effective decision support is provided for 5G network communication control.
Accordingly, according to one aspect of the present application, there is provided a 5G network communication policing method comprising receiving a policing command, wherein the policing command comprises policing region information; determining a target user according to the control region information, wherein the target user is positioned in a control region corresponding to the control region information; transmitting a re-registration message to user equipment of the target user; receiving a registration message sent by the user equipment of the target user according to the re-registration message; determining whether the target user is a regulated user according to the regulation command and the registration message; responding to the target user as a control user, and routing a registration message of the control user to a control network element; the control network element selects a corresponding control policy to control the control user according to the registration message, wherein the control network element selects the corresponding control policy to control the control user according to the registration message, and the control network element comprises:
responding to the target user as a control user, and acquiring registration information of the target user;
performing embedded coding on the registration information to obtain a sequence of embedded coding vectors of the registration information word granularity;
respectively extracting short-term dependency semantic relation and long-term dependency semantic relation of the sequence of the registration information word granularity embedded coding vector to obtain a first registration information semantic understanding feature vector and a second registration information semantic understanding feature vector;
carrying out projection fusion on the first registration information semantic understanding feature vector and the second registration information semantic understanding feature vector to obtain a multi-scale registration information semantic understanding feature vector;
and generating a control strategy label of the target user based on the multi-scale registration information semantic understanding feature vector.
In the above 5G network communication policing method, performing embedded encoding on the registration information to obtain a sequence of embedded encoding vectors of a registration information word granularity, including: after the registration information is subjected to word segmentation, a word granularity encoder based on a word embedding layer is used for obtaining a sequence of the registration information word granularity embedded encoding vector.
In the above 5G network communication management method, extracting the short-term dependency semantic relation and the long-term dependency semantic relation of the sequence of the registration information word granularity embedded encoding vector to obtain a first registration information semantic understanding feature vector and a second registration information semantic understanding feature vector, respectively, includes: carrying out local semantic association coding on the sequence of the registration information word granularity embedded coding vector to obtain the first registration information semantic understanding feature vector; and carrying out bidirectional cyclic semantic association coding on the sequence of the registration information word granularity embedded coding vector to obtain the second registration information semantic understanding feature vector.
In the above 5G network communication policing method, performing local semantic association encoding on the sequence of the registration information word granularity embedded encoding vector to obtain the first registration information semantic understanding feature vector, including: and embedding the registration information word granularity into a sequence of coding vectors, and obtaining the first registration information semantic understanding feature vectors through a first semantic encoder based on a text convolutional neural network model.
In the above 5G network communication policing method, performing bidirectional cyclic semantic association encoding on the sequence of the registration information word granularity embedded encoding vector to obtain the second registration information semantic understanding feature vector, including: and embedding the registration information word granularity into a sequence of coding vectors, and obtaining the second registration information semantic understanding feature vectors through a second semantic encoder based on a bi-directional gating circulating unit.
In the above 5G network communication policing method, performing projection fusion on the first registration information semantic understanding feature vector and the second registration information semantic understanding feature vector to obtain a multi-scale registration information semantic understanding feature vector, including: and carrying out semantic fusion on the first registration information semantic understanding feature vector and the second registration information semantic understanding feature vector by using a multi-scale semantic fusion device based on a projection layer so as to obtain the multi-scale registration information semantic understanding feature vector.
In the above 5G network communication policing method, performing semantic fusion on the first registration information semantic understanding feature vector and the second registration information semantic understanding feature vector by using a projection layer-based multi-scale semantic fusion device to obtain the multi-scale registration information semantic understanding feature vector, including: carrying out semantic fusion on the first registration information semantic understanding feature vector and the second registration information semantic understanding feature vector by using a projection fusion formula to obtain the multi-scale registration information semantic understanding feature vector; the projection fusion formula is as follows:
wherein,is the first registration information semantic understanding feature vector,/->Is the second registration information semantic understanding feature vector,/->Is the semantic understanding feature vector of the multi-scale registration information,>a projection fusion process is represented and is performed,representing a cascading process.
In the above 5G network communication policing method, generating the policing policy tag of the target user based on the multi-scale registration information semantic understanding feature vector includes: performing feature distribution optimization on the multi-scale registration information semantic understanding feature vector to obtain an optimized multi-scale registration information semantic understanding feature vector; and the optimized multi-scale registration information semantic understanding feature vector passes through a classifier to obtain a classification result, wherein the classification result is used for representing a control strategy label.
In the above 5G network communication policing method, performing feature distribution optimization on the multi-scale registration information semantic understanding feature vector to obtain an optimized multi-scale registration information semantic understanding feature vector, including: performing fusion optimization on the first registration information semantic understanding feature vector and the second registration information semantic understanding feature vector to obtain an optimized feature vector; and fusing the optimized feature vector and the multi-scale registration information semantic understanding feature vector to obtain the optimized multi-scale registration information semantic understanding feature vector.
In the above 5G network communication policing method, the classifying method includes that the optimized multi-scale registration information semantic understanding feature vector passes through a classifier to obtain a classification result, where the classification result is used to represent a policing policy tag, and includes: performing full-connection coding on the optimized multi-scale registration information semantic understanding feature vector by using a full-connection layer of the classifier to obtain a full-connection coding feature vector; inputting the fully-connected coding feature vector into a Softmax classification function of the classifier to obtain probability values of the optimized multi-scale registration information semantic understanding feature vector belonging to various classification labels, wherein the classification labels comprise all control strategy labels; and determining the classification label corresponding to the maximum probability value as the classification result.
Compared with the prior art, the 5G network communication control method provided by the application adopts a natural language processing technology based on deep learning to perform semantic understanding on registration information of a target user, captures semantic feature expression of the registration information of the target user, and digs out potential risk information in the registration information, so that a corresponding control strategy is intelligently selected to control the target user. Thus, intelligent control strategy selection and implementation can be realized, and effective decision support is provided for 5G network communication control.
Drawings
The foregoing and other objects, features and advantages of the present application will become more apparent from the following more particular description of embodiments of the present application, as illustrated in the accompanying drawings. The accompanying drawings are included to provide a further understanding of embodiments of the application and are incorporated in and constitute a part of this specification, illustrate the application and not constitute a limitation to the application. In the drawings, like reference numerals generally refer to like parts or steps.
Fig. 1 is a flowchart of a 5G network communication policing method according to an embodiment of the present application.
Fig. 2 is a schematic architecture diagram of a 5G network communication policing method according to an embodiment of the present application.
Fig. 3 is a flowchart of extracting short-term dependency semantic relationships and long-term dependency semantic relationships of sequences of the registration information word granularity embedded coding vectors to obtain a first registration information semantic understanding feature vector and a second registration information semantic understanding feature vector in a 5G network communication control method according to an embodiment of the present application.
Fig. 4 is a flowchart of generating a control policy tag of the target user based on the multi-scale registration information semantic understanding feature vector in the 5G network communication control method according to an embodiment of the present application.
Fig. 5 is a flowchart of a method for managing 5G network communications according to an embodiment of the present application, where the optimized multi-scale registration information semantic understanding feature vector is passed through a classifier to obtain a classification result.
Detailed Description
For an understanding of embodiments of the present invention, specific embodiments of the invention will be described in more detail below with reference to the drawings. While specific embodiments of the invention are shown in the drawings, it should be understood that the invention may be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art.
Patent CN110740489a discloses a 5G network communication control method, which specifically comprises: receiving a control command, wherein the control command comprises control region information; determining a target user according to the control region information, wherein the target user is positioned in a control region corresponding to the control region information; transmitting a re-registration message to user equipment of the target user; receiving a registration message sent by the user equipment of the target user according to the re-registration message; determining whether the target user is a regulated user according to the regulation command and the registration message; responding to the target user as a control user, and routing a registration message of the control user to a control network element; and the control network element selects a corresponding control strategy to control the control user according to the registration message.
Specifically, after confirming that the target user is a control user, the target user needs to be controlled by selecting a corresponding control policy according to a registration message of the control user. However, the conventional network communication control method is often formulated based on general rules and standards, and cannot fully consider specific situations of different users, and cannot perform personalized control according to information of specific users, which may cause excessive control in some situations, bring inconvenience and unnecessary limitation to users, and insufficient control in other situations, and cannot effectively restrict behaviors of users.
Aiming at the technical problems, the technical concept of the application is to adopt a natural language processing technology based on deep learning to perform semantic understanding on registration information of a target user, capture semantic feature expression of the registration information of the target user, and mine potential risk information in the registration information, so that a corresponding control strategy is intelligently selected to control the target user. Thus, intelligent control strategy selection and implementation can be realized, and effective decision support is provided for 5G network communication control.
Fig. 1 is a flowchart of a 5G network communication policing method according to an embodiment of the present application. Fig. 2 is a schematic architecture diagram of a 5G network communication policing method according to an embodiment of the present application. As shown in fig. 1 and 2, a 5G network communication policing method according to an embodiment of the present application includes the steps of: s110, responding to the target user as a control user, and acquiring registration information of the target user; s120, embedding and coding the registration information to obtain a sequence of registration information word granularity embedded coding vectors; s130, respectively extracting short-term dependency semantic relation and long-term dependency semantic relation of the sequence of the registration information word granularity embedded coding vector to obtain a first registration information semantic understanding feature vector and a second registration information semantic understanding feature vector; s140, carrying out projection fusion on the first registration information semantic understanding feature vector and the second registration information semantic understanding feature vector to obtain a multi-scale registration information semantic understanding feature vector; s150, generating a control strategy label of the target user based on the multi-scale registration information semantic understanding feature vector.
In the above 5G network communication control method, in step S110, in response to the target user being a control user, registration information of the target user is obtained. It should be understood that, through the user identifier, account information and the like in the registration information, the unique identity of the target user can be determined, so that the other users are prevented from being controlled in error, and the specified user is ensured to be targeted by the control measures. After confirming that the target user is a control user, the demand degree of the user on the network resource can be known through user personal information, geographic position, equipment information, user demand information and the like contained in the registration information, so that the control strategy is adjusted according to the demand degrees of different users, reasonable distribution of the resource is ensured, and the reasonable demand of the user is met. Meanwhile, the registration information also comprises information such as account credit, historical behavior record and the like of the user. By analyzing the credit status and behavior records of the user, whether the user has security risks, such as malicious attack, illegal behaviors and the like, can be judged, and corresponding control strategies, such as limiting the network access speed, limiting the access to certain specific websites or applications and the like, are formulated so as to prevent potential network security threats.
In the above 5G network communication control method, in step S120, the registration information is embedded and encoded to obtain a sequence of the registration information word granularity embedded encoding vector. In a specific example of the present application, the encoding manner of embedding and encoding the registration information to obtain the sequence of the registration information word granularity embedded encoding vector is to obtain the sequence of the registration information word granularity embedded encoding vector through a word granularity encoder based on a word embedding layer after performing word segmentation processing on the registration information. It should be understood that word segmentation is the process of dividing a continuous sequence of text into a series of discrete words or phrases, each representing a semantic unit. In the technical scheme of the application, the registration information is divided into a plurality of discrete words by word segmentation processing, so that semantic information mining with finer granularity can be carried out on the registration information. After the registration information is subjected to word segmentation, a word granularity encoder based on a word embedding layer is further used for mapping each word into an embedded coding vector representation form. Here, word embedding is a technique of representing words as real vectors, capable of encoding semantic information of words as positional relationships in a feature space. In the technical scheme of the application, each word in the registration information is respectively input into the word granularity encoder for embedded encoding, the word embedding layer of the word granularity encoder can capture the semantic relation among the words through the distributed representation of the learning words in the high-dimensional feature space, and then each word text is respectively converted into the corresponding registration information word granularity embedded encoding vector, so that a computer can conveniently analyze and process the semantic information of the registration information, and therefore intelligent control decision is realized.
In the above 5G network communication control method, step S130 extracts the short-term dependency semantic relationship and the long-term dependency semantic relationship of the sequence of the registration information word granularity embedded encoding vector to obtain the first registration information semantic understanding feature vector and the second registration information semantic understanding feature vector, respectively. Specifically, fig. 3 is a flowchart of extracting short-term dependency semantic relationships and long-term dependency semantic relationships of sequences of the registration information word granularity embedded coding vectors to obtain first registration information semantic understanding feature vectors and second registration information semantic understanding feature vectors in the 5G network communication control method according to the embodiment of the present application. As shown in fig. 3, the step S130 includes: s131, carrying out local semantic association coding on the sequence of the registration information word granularity embedded coding vector to obtain the first registration information semantic understanding feature vector; s132, performing bidirectional cyclic semantic association coding on the sequence of the registration information word granularity embedded coding vector to obtain the second registration information semantic understanding feature vector.
Specifically, in step S131, the sequence of the registration information word granularity embedded encoding vector is subjected to local semantic association encoding to obtain the first registration information semantic understanding feature vector. In a specific example of the present application, the encoding mode of performing local semantic association encoding on the sequence of the registration information word granularity embedded encoding vector to obtain the first registration information semantic understanding feature vector is that the sequence of the registration information word granularity embedded encoding vector passes through a first semantic encoder based on a text convolutional neural network model to obtain the first registration information semantic understanding feature vector. One of ordinary skill in the art will appreciate that a text convolutional neural network (Convolutional Neural Network, CNN) is a deep learning model commonly used to process sequence data. In natural language processing tasks, the text convolutional neural network model can effectively capture local associations and semantic features between words through convolution operations and pooling operations. In the technical scheme, the sequence of the registration information word granularity embedded coding vector is input into a first semantic encoder based on a text convolutional neural network model, and the first semantic encoder performs sliding convolution operation on the sequence of the registration information word granularity embedded coding vector in a local window by utilizing a convolution layer of the text convolutional neural network model to extract local semantic association features in the sequence, and performs feature reduction and abstraction by pooling operation, so that context semantic feature representation of the registration information is obtained. In addition, the first semantic encoder based on the text convolutional neural network model can gradually extract semantic features of higher layers through stacking of a plurality of convolutional layers and pooling layers so as to obtain first registration information semantic understanding feature vectors containing local context semantic association features of the registration information, and richer semantic representations are provided for subsequent control decision tasks.
Specifically, in step S132, the sequence of the registration information word granularity embedded encoding vector is subjected to bidirectional cyclic semantic association encoding to obtain the second registration information semantic understanding feature vector. In a specific example of the present application, the encoding mode of performing bi-directional cyclic semantic association encoding on the sequence of the registration information word granularity embedded encoding vector to obtain the second registration information semantic understanding feature vector is that the sequence of the registration information word granularity embedded encoding vector passes through a second semantic encoder based on a bi-directional gating cyclic unit to obtain the second registration information semantic understanding feature vector. It should be understood that, considering that the first semantic encoder based on the text convolutional neural network model models the local semantic association relationship between the words in the registration information, only short-distance dependency relationship between the words can be captured, and long-distance dependency relationship, that is, semantic dependency relationship between words in a far position, exists between the words in the registration information. Therefore, in order to further capture long-distance dependency relations among words in the registration information, a two-way gating circulation unit is used for carrying out long-distance semantic dependency coding on the sequence of the registration information word granularity embedded coding vector. One of ordinary skill in the art will appreciate that a bi-directional gating loop unit is a variant of a recurrent neural network that can take into account both forward and backward context information in the sequence data by running two independent gating loop units in both forward and backward directions. Compared with the traditional cyclic neural network, the bi-directional gating cyclic unit can better capture long-term dependency and bi-directional context information in the sequence. In the technical scheme of the application, the sequence of the registration information word granularity embedded coding vector is input into a second semantic encoder based on a bidirectional gating circulating unit, the bidirectional gating circulating unit can respectively process the sequence of the registration information word granularity embedded coding vector through a forward circulating computing unit and a backward circulating computing unit, hidden states of the two directions are spliced or fused, and the sequence of the registration information word granularity embedded coding vector is modeled by fully utilizing bidirectional context information so as to capture long-term dependency relationship among words in the registration information and the bidirectional context information, thereby providing more accurate and comprehensive semantic feature representation for subsequent control decisions.
In the above 5G network communication control method, in step S140, projection fusion is performed on the first registration information semantic understanding feature vector and the second registration information semantic understanding feature vector to obtain a multi-scale registration information semantic understanding feature vector. In a specific example of the present application, the implementation manner of performing projection fusion on the first registration information semantic understanding feature vector and the second registration information semantic understanding feature vector to obtain a multi-scale registration information semantic understanding feature vector is to perform semantic fusion on the first registration information semantic understanding feature vector and the second registration information semantic understanding feature vector by using a multi-scale semantic fusion device based on a projection layer to obtain the multi-scale registration information semantic understanding feature vector. It should be appreciated that the first registration information semantic understanding feature vector and the second registration information semantic understanding feature vector represent semantic information of different levels of the registration information, respectively. By means of semantic fusion of the two, long-distance semantic dependency and short-distance semantic dependency of the registration information can be comprehensively considered, so that multi-scale registration information semantic understanding feature vectors with diversified semantic representation capability are obtained, and richer and comprehensive semantic information support is provided for subsequent control decisions. Specifically, the multi-scale semantic fusion device based on the projection layer can map the first registration information semantic understanding feature vector and the second registration information semantic understanding feature vector with different scales into a shared semantic space through projection operation, so that the first registration information semantic understanding feature vector and the second registration information semantic understanding feature vector have the same dimension and the same representation mode, and semantic fusion is realized in the same semantic feature space.
Specifically, using a multi-scale semantic fusion device based on a projection layer to perform semantic fusion on the first registration information semantic understanding feature vector and the second registration information semantic understanding feature vector to obtain the multi-scale registration information semantic understanding feature vector, including: carrying out semantic fusion on the first registration information semantic understanding feature vector and the second registration information semantic understanding feature vector by using a projection fusion formula to obtain the multi-scale registration information semantic understanding feature vector; the projection fusion formula is as follows:
wherein,is the first registration information semantic understanding feature vector,/->Is the second registration information semantic understanding feature vector,/->Is the semantic understanding feature vector of the multi-scale registration information,>a projection fusion process is represented and is performed,representing a cascading process.
In the above 5G network communication control method, step S150 generates a control policy tag of the target user based on the multi-scale registration information semantic understanding feature vector. Specifically, fig. 4 is a flowchart of generating a control policy tag of the target user based on the multi-scale registration information semantic understanding feature vector in the 5G network communication control method according to the embodiment of the present application. As shown in fig. 4, the step S150 includes: s151, performing feature distribution optimization on the multi-scale registration information semantic understanding feature vector to obtain an optimized multi-scale registration information semantic understanding feature vector; and S152, enabling the optimized multi-scale registration information semantic understanding feature vector to pass through a classifier to obtain a classification result, wherein the classification result is used for representing a control strategy label.
Specifically, in step S151, feature distribution optimization is performed on the multi-scale registration information semantic understanding feature vector to obtain an optimized multi-scale registration information semantic understanding feature vector. In the above technical solution, the first registration information semantic understanding feature vector and the second registration information semantic understanding feature vector respectively express local convolution associated text semantic features and bidirectional circulation associated text semantic features of the word segmentation semantics of the registration information, so that variability in text semantic feature calculation dimensions due to text semantic feature extraction differences exists in the first registration information semantic understanding feature vector and the second registration information semantic understanding feature vector.
In this way, in order to promote the fusion effect of the first registration information semantic understanding feature vector and the second registration information semantic understanding feature vector under the classification judgment based on the classifier when the projection layer-based multi-scale semantic fusion device is used for carrying out semantic fusion on the first registration information semantic understanding feature vector and the second registration information semantic understanding feature vector, in the technical scheme of the application, fusion optimization is preferably carried out on the first registration information semantic understanding feature vector and the second registration information semantic understanding feature vector to obtain an optimized feature vector, and the optimized feature vector and the multi-scale registration information semantic understanding feature vector are fused to obtain the optimized multi-scale registration information semantic understanding feature vector.
In a specific example of the present application, the fusion optimization formula is used to perform fusion optimization on the first registration information semantic understanding feature vector and the second registration information semantic understanding feature vector to obtain the optimized feature vector, where the fusion optimization formula is:
wherein,and->The first registration information semantic understanding feature vector +.>And said second registration information semantic understanding feature vector +.>Characteristic value of>And->Respectively representing the square of one norm of the feature vector and the square root of two norms of the feature vector, said first feature vector +.>And a second feature vector->Having the same eigenvector length->，/>Expressed as +.>An exponential operation as a base, and +.>Is a weight superparameter,/->Is the eigenvalue of the optimized eigenvector.
Here, the above-described optimization fusion semantically understands the feature vector of the first registration information based on the correspondence at the feature value granularityAnd said second registration information semantic understanding feature vector +.>Dividing foreground manifold and background manifold based on vector scale to stack the first registration information semantic understanding feature vector +_ under feature corresponding channel super manifold aggregation mechanism>And said second registration information semantic understanding feature vector +.>Is associated with a dynamic feature value of the first registration information semantic understanding feature vector, thereby marking the first registration information semantic understanding feature vector +.>And said second registration information semantic understanding feature vector +.>Inter-sequence variation feature semantic information between the first registration information, realizing semantic understanding of feature vector +.>And said second registration information semantic understanding feature vector +.>And (3) similar full-connection type stacking fusion of the variability of semantic content among different computing dimensions to obtain an optimized feature vector. Then, feature fusion is carried out on the optimized feature vector and the multi-scale registration information semantic understanding feature vector so as to promote the first registration information semantic understanding feature vector +.>And said second registration information semantic understanding feature vector +.>The semantic fusion effect of the multi-scale registration information semantic understanding feature vector is improved, and the accuracy of a classification result obtained by the classifier is improved.
Specifically, in step S152, the optimized multi-scale registration information semantic understanding feature vector is passed through a classifier to obtain a classification result, where the classification result is used to represent a regulatory policy tag. It should be appreciated that a classifier is a machine learning model that is capable of learning the mapping from input data (the optimized multi-scale registration information semantic understanding feature vector) to output categories (policing policy labels). Firstly, in a model training stage, a classifier is trained and optimized by using marked registration information samples, and parameters and weights in the classifier are adjusted, so that the classifier can learn the mapping relation between input features and preset labels, and further control strategy labels corresponding to different registration information can be accurately predicted. Once the classifier training is completed, the optimized multi-scale registration information semantic understanding feature vector of the unknown label is input into the classifier to be classified, a corresponding control strategy label is output, and then the control strategy of the target user is determined according to the classification result, so that effective decision support is provided for 5G network communication control.
Fig. 5 is a flowchart of a method for managing 5G network communications according to an embodiment of the present application, where the optimized multi-scale registration information semantic understanding feature vector is passed through a classifier to obtain a classification result. As shown in fig. 5, the step S152 includes: s1521, performing full-connection coding on the optimized multi-scale registration information semantic understanding feature vector by using a full-connection layer of the classifier to obtain a full-connection coding feature vector; s1522, inputting the fully-connected coding feature vector into a Softmax classification function of the classifier to obtain probability values of the optimized multi-scale registration information semantic understanding feature vector belonging to each classification label, wherein the classification labels comprise all control strategy labels; and S1523, determining the classification label corresponding to the maximum probability value as the classification result.
In summary, the 5G network communication control method according to the embodiments of the present application is set forth, which performs semantic understanding on registration information of a target user by using a natural language processing technology based on deep learning, captures semantic feature expression of registration information of the target user, and discovers potential risk information in the registration information, so as to intelligently select a corresponding control policy to control the target user. Thus, intelligent control strategy selection and implementation can be realized, and effective decision support is provided for 5G network communication control.
In the foregoing embodiments, the descriptions of the embodiments are emphasized, and in part, not described or illustrated in any particular embodiment, reference is made to the related descriptions of other embodiments. In the several embodiments provided by the present invention, it should be understood that the disclosed apparatus, system and method may be implemented in other manners. For example, the system embodiments described above are merely illustrative, and for example, the module division is merely a logical function division, and other manners of division may be implemented in practice. The modules described as separate components may or may not be physically separate, and components shown as modules may or may not be physical units, may be located in one place, or may be distributed over multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
In addition, each functional module in the embodiments of the present invention may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units can be realized in a form of hardware or a form of hardware and a form of software functional modules.
It will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential characteristics thereof. The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Any reference signs in the claims shall not be construed as limiting the claim concerned.
Furthermore, it is evident that the word "comprising" does not exclude other elements or steps, and that the singular does not exclude a plurality. The units recited in the system claims may also be implemented by means of software or hardware.
Finally, it should be noted that the above-mentioned embodiments are merely for illustrating the technical solution of the present invention and not for limiting the same, and although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that modifications and equivalents may be made to the technical solution of the present invention without departing from the spirit and scope of the technical solution of the present invention.
Claims (10)
1. A5G network communication policing method includes receiving a policing command, wherein the policing command includes policing area information; determining a target user according to the control region information, wherein the target user is positioned in a control region corresponding to the control region information; transmitting a re-registration message to user equipment of the target user; receiving a registration message sent by the user equipment of the target user according to the re-registration message; determining whether the target user is a regulated user according to the regulation command and the registration message; responding to the target user as a control user, and routing a registration message of the control user to a control network element; the control network element selects a corresponding control policy to control the control user according to the registration message, wherein the control network element selects the corresponding control policy to control the control user according to the registration message, and the control network element is characterized by comprising:
responding to the target user as a control user, and acquiring registration information of the target user;
performing embedded coding on the registration information to obtain a sequence of embedded coding vectors of the registration information word granularity;
respectively extracting short-term dependency semantic relation and long-term dependency semantic relation of the sequence of the registration information word granularity embedded coding vector to obtain a first registration information semantic understanding feature vector and a second registration information semantic understanding feature vector;
carrying out projection fusion on the first registration information semantic understanding feature vector and the second registration information semantic understanding feature vector to obtain a multi-scale registration information semantic understanding feature vector;
and generating a control strategy label of the target user based on the multi-scale registration information semantic understanding feature vector.
2. The method of claim 1, wherein embedding the registration information to obtain a sequence of registration information word granularity embedded encoding vectors comprises:
after the registration information is subjected to word segmentation, a word granularity encoder based on a word embedding layer is used for obtaining a sequence of the registration information word granularity embedded encoding vector.
3. The 5G network communication policing method of claim 2, wherein extracting short-term and long-term dependency semantic relationships of the sequence of registration information word granularity embedded encoding vectors to obtain a first registration information semantic understanding feature vector and a second registration information semantic understanding feature vector, respectively, comprises:
carrying out local semantic association coding on the sequence of the registration information word granularity embedded coding vector to obtain the first registration information semantic understanding feature vector;
and carrying out bidirectional cyclic semantic association coding on the sequence of the registration information word granularity embedded coding vector to obtain the second registration information semantic understanding feature vector.
4. A 5G network communication policing method according to claim 3, characterized in that performing local semantic association coding on the sequence of registration information word granularity embedded coding vectors to obtain the first registration information semantic understanding feature vector, comprising:
and embedding the registration information word granularity into a sequence of coding vectors, and obtaining the first registration information semantic understanding feature vectors through a first semantic encoder based on a text convolutional neural network model.
5. The 5G network communication policing method of claim 4, wherein bi-directionally cyclic semantic association encoding the sequence of registration information word granularity embedded encoding vectors to obtain the second registration information semantic understanding feature vector, comprising:
and embedding the registration information word granularity into a sequence of coding vectors, and obtaining the second registration information semantic understanding feature vectors through a second semantic encoder based on a bi-directional gating circulating unit.
6. The 5G network communication policing method of claim 5, wherein projection fusing the first registration information semantic understanding feature vector and the second registration information semantic understanding feature vector to obtain a multi-scale registration information semantic understanding feature vector, comprising:
and carrying out semantic fusion on the first registration information semantic understanding feature vector and the second registration information semantic understanding feature vector by using a multi-scale semantic fusion device based on a projection layer so as to obtain the multi-scale registration information semantic understanding feature vector.
7. The 5G network communication policing method of claim 6, wherein semantically fusing the first registration information semantically understood feature vector and the second registration information semantically understood feature vector using a projection layer-based multi-scale semantic fusion engine to obtain the multi-scale registration information semantically understood feature vector, comprising:
carrying out semantic fusion on the first registration information semantic understanding feature vector and the second registration information semantic understanding feature vector by using a projection fusion formula to obtain the multi-scale registration information semantic understanding feature vector; the projection fusion formula is as follows:
wherein,is the first registration information semantic understanding feature vector,/->Is the second registration information semantic understanding feature vector,/->Is the semantic understanding feature vector of the multi-scale registration information,>a projection fusion process is represented and is performed,representing a cascading process.
8. The 5G network communication policing method of claim 7, wherein generating the policing policy tag for the target user based on the multi-scale registration information semantic understanding feature vector comprises:
performing feature distribution optimization on the multi-scale registration information semantic understanding feature vector to obtain an optimized multi-scale registration information semantic understanding feature vector;
and the optimized multi-scale registration information semantic understanding feature vector passes through a classifier to obtain a classification result, wherein the classification result is used for representing a control strategy label.
9. The 5G network communication policing method of claim 8, wherein performing feature distribution optimization on the multi-scale registration information semantic understanding feature vector to obtain an optimized multi-scale registration information semantic understanding feature vector, comprising:
performing fusion optimization on the first registration information semantic understanding feature vector and the second registration information semantic understanding feature vector to obtain an optimized feature vector;
and fusing the optimized feature vector and the multi-scale registration information semantic understanding feature vector to obtain the optimized multi-scale registration information semantic understanding feature vector.
10. The 5G network communication policing method of claim 9, wherein passing the optimized multi-scale registration information semantic understanding feature vector through a classifier to obtain a classification result, the classification result being used to represent a policing policy tag, comprising:
performing full-connection coding on the optimized multi-scale registration information semantic understanding feature vector by using a full-connection layer of the classifier to obtain a full-connection coding feature vector;
inputting the fully-connected coding feature vector into a Softmax classification function of the classifier to obtain probability values of the optimized multi-scale registration information semantic understanding feature vector belonging to various classification labels, wherein the classification labels comprise all control strategy labels;
and determining the classification label corresponding to the maximum probability value as the classification result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410179104.7A CN117729545B (en) | 2024-02-18 | 2024-02-18 | 5G network communication control method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410179104.7A CN117729545B (en) | 2024-02-18 | 2024-02-18 | 5G network communication control method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117729545A true CN117729545A (en) | 2024-03-19 |
| CN117729545B CN117729545B (en) | 2024-05-03 |
Family
ID=90200191
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410179104.7A Active CN117729545B (en) | 2024-02-18 | 2024-02-18 | 5G network communication control method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117729545B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108073677A (en) * | 2017-11-02 | 2018-05-25 | 中国科学院信息工程研究所 | A kind of multistage text multi-tag sorting technique and system based on artificial intelligence |
| WO2019037604A1 (en) * | 2017-08-21 | 2019-02-28 | 华为技术有限公司 | Communication method, related device and system |
| CN109902175A (en) * | 2019-02-20 | 2019-06-18 | 上海方立数码科技有限公司 | A kind of file classification method and categorizing system based on neural network structure model |
| CN109948149A (en) * | 2019-02-28 | 2019-06-28 | 腾讯科技（深圳）有限公司 | A kind of file classification method and device |
| CN110740489A (en) * | 2019-10-21 | 2020-01-31 | 中国联合网络通信集团有限公司 | 5G network communication control method, device and communication system |
| JP2022144272A (en) * | 2021-03-18 | 2022-10-03 | ヤフー株式会社 | Distribution apparatus, distribution method, and distribution program |
| CN116225950A (en) * | 2023-03-15 | 2023-06-06 | 北京赋乐科技有限公司 | Identification method and system of fraud APP based on multi-mode fusion |
-
2024
- 2024-02-18 CN CN202410179104.7A patent/CN117729545B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2019037604A1 (en) * | 2017-08-21 | 2019-02-28 | 华为技术有限公司 | Communication method, related device and system |
| CN108073677A (en) * | 2017-11-02 | 2018-05-25 | 中国科学院信息工程研究所 | A kind of multistage text multi-tag sorting technique and system based on artificial intelligence |
| CN109902175A (en) * | 2019-02-20 | 2019-06-18 | 上海方立数码科技有限公司 | A kind of file classification method and categorizing system based on neural network structure model |
| CN109948149A (en) * | 2019-02-28 | 2019-06-28 | 腾讯科技（深圳）有限公司 | A kind of file classification method and device |
| CN110740489A (en) * | 2019-10-21 | 2020-01-31 | 中国联合网络通信集团有限公司 | 5G network communication control method, device and communication system |
| JP2022144272A (en) * | 2021-03-18 | 2022-10-03 | ヤフー株式会社 | Distribution apparatus, distribution method, and distribution program |
| CN116225950A (en) * | 2023-03-15 | 2023-06-06 | 北京赋乐科技有限公司 | Identification method and system of fraud APP based on multi-mode fusion |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117729545B (en) | 2024-05-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105844239B (en) | It is a kind of that video detecting method is feared based on CNN and LSTM cruelly | |
| Yi et al. | Pedestrian behavior understanding and prediction with deep neural networks | |
| Wu et al. | Attentive prototype few-shot learning with capsule network-based embedding | |
| Fan et al. | Federated generative adversarial learning | |
| Sharma et al. | Analysis of machine learning techniques based intrusion detection systems | |
| CN110855648B (en) | Early warning control method and device for network attack | |
| Qu et al. | Statistics-enhanced direct batch growth self-organizing mapping for efficient DoS attack detection | |
| JP2022531974A (en) | Dealing with rare training data for artificial intelligence | |
| CN117729545B (en) | 5G network communication control method | |
| KR102021138B1 (en) | Method and program of malicious domain classification based on artificial intelligence | |
| Li et al. | Fast key-frame image retrieval of intelligent city security video based on deep feature coding in high concurrent network environment | |
| CN115983275A (en) | Named entity identification method, system and electronic equipment | |
| CN115797701A (en) | Target classification method and device, electronic equipment and storage medium | |
| CN112926368B (en) | Method and device for identifying obstacle | |
| Cao et al. | Traffic condition classification model based on traffic-net | |
| CN113626854A (en) | Image data privacy protection method based on localized differential privacy | |
| CN113592341A (en) | Measurement loss function, sector complexity evaluation method and system | |
| KR20220102012A (en) | Method and apparatus for data genration | |
| Ma et al. | Video-based person re-identification by semi-supervised adaptive stepwise learning | |
| Madhavi et al. | Bird Species Recognition Using Deep Transfer Learning | |
| Li et al. | Pre-trained language model-enhanced conditional generative adversarial networks for intrusion detection | |
| CN116932766B (en) | Object classification method, device, apparatus, storage medium, and program product | |
| SrirangamSridharan et al. | Doc2img: A new approach to vectorization of documents | |
| CN110300097A (en) | Information based on Incorporate network transmits data format Uniform Name method | |
| CN114385892B (en) | Article grade identification method, device, server and storage medium |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |