5g message development template generation method, device, storage medium and product

Abstract

The application discloses a 5G message development template generation method, comprising the following steps: acquiring a 5G message demand data set and a 5G message development template data set based on the industry type of a 5G message new creation request; inputting each demand information in the 5G message demand data set into an LSTM encoding module of a coding and decoding neural network for processing to obtain a context vector, and inputting the context vector into an LSTM decoding module of the coding and decoding neural network for processing to obtain trained development template information; and determining a target 5G message development template based on each development template information and the trained development template information. The application further discloses a 5G message development template generation device, a storage medium and a computer program product. The application can automatically generate a corresponding 5G message development template according to a 5G message new creation request, automatically generate a 5G message development template, improve the production efficiency of the 5G message development template, and reduce the labor cost in the process of generating the 5G message development template.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of mobile communication, and particularly relates to a 5G message development template generation method, device, storage medium and product. BACKGROUND

[0002] At present, 5G technology is an important part of new generation infrastructure construction, and 5G message is one of the applications of 5G technology. Compared with the traditional short message with single function, 5G message not only broadens the breadth of information transmission and supports users to use multimedia content such as text, audio, video, card and location, but also extends the depth of interactive experience. Users can complete service search, discovery, interaction, payment and other businesses in the message window, and build an information window of one-stop service. 5G message has the characteristics of message (C2C) for individual user interaction and message (B2C) for interaction between industry customers and individual users. It can not only facilitate the transmission of rich media information such as voice, picture, video, card and file between users, but also support enterprises to provide interactive services in the form of chat robots on the 5G message platform.

[0003] At present, 5G message applications are generally developed by industry templates provided by 5G message opening platform. However, 5G message industry templates are mainly made by artificial, and 5G message is aimed at thousands of industries. The artificial way is time-consuming and laborious, and the efficiency is low.

[0004] The above content is only used to assist in understanding the technical solutions of the present application, and does not represent the acknowledgement of the above content as prior art. SUMMARY

[0005] The main purpose of the present application is to provide a 5G message development template generation method, device, storage medium and product, which aims to solve the technical problem of low efficiency of 5G message development template making.

[0006] To achieve the above purpose, the present application provides a 5G message development template generation method, which comprises the following steps:

[0007] Based on the industry type of the new 5G message request, the historical new 5G message corresponding 5G message demand data set and 5G message development template data set are obtained;

[0008] Each demand information in the 5G message demand data set is respectively input into the LSTM encoding module of the coding and decoding neural network for processing to obtain a context vector, and the context vector is input into the LSTM decoding module of the coding and decoding neural network for processing to obtain the trained development template information;

[0009] Determine the optimal weight of the encoding-decoding neural network based on each development template information of the development template information set and the trained development template information, and determine the target 5G message development template based on the optimal weight.

[0010] Further, the step of inputting each requirement information in the 5G message requirement data set into the LSTM encoding module of the encoding-decoding neural network for processing comprises:

[0011] Input each requirement information into the embedding layer of the LSTM encoding module respectively to obtain the requirement vector corresponding to each requirement information.

[0012] Input each requirement vector into the LSTM encoding layer of the corresponding LSTM encoding module respectively to obtain the context requirement vector corresponding to each requirement vector.

[0013] Input the context requirement vector corresponding to each requirement information into the merging layer to obtain the context vector.

[0014] Further, the step of inputting the context vector output by the LSTM encoding module into the LSTM decoding module of the encoding-decoding neural network for processing to obtain the trained development template information comprises:

[0015] Input the context vector into the LSTM decoding layer of the LSTM decoding module to obtain the intermediate data output by the LSTM decoding layer.

[0016] Input the intermediate data into the fully connected layer of the LSTM decoding module to obtain the trained development template information.

[0017] Further, the step of determining the optimal weight of the encoding-decoding neural network based on each development template information of the development template information set and the trained development template information comprises:

[0018] Determine the loss function based on each development template information of the development template information set and the trained development template information.

[0019] Determine the optimal weight of the encoding-decoding neural network based on the loss function.

[0020] Further, the step of obtaining the 5G message requirement data set and the 5G message development template data set corresponding to the historical newly-built 5G message comprises:

[0021] Obtain historical requirement information and industry message development templates corresponding to historical newly-built 5G messages.

[0022] The historical demand information and the industry message development template are respectively subjected to data processing operations to obtain a 5G message demand data set and a 5G message development template data set.

[0023] Further, the step of respectively subjecting the historical demand information and the industry message development template to data processing operations to obtain a 5G message demand data set and a 5G message development template data set comprises:

[0024] The historical demand information is subjected to text cleaning and text serialization processing to obtain the 5G message demand data set;

[0025] The industry message development template is subjected to text cleaning and text serialization processing to obtain the 5G message development template data set.

[0026] Further, before the step of obtaining the 5G message demand data set and the 5G message development template data set corresponding to the historical newly-built 5G message based on the industry type of the 5G message new building request, the method further comprises:

[0027] Upon receiving the 5G message new building request, obtaining new 5G message demand information corresponding to the 5G message new building request;

[0028] Based on the new 5G message demand information, determining whether a target 5G message development template corresponding to the 5G message new building request exists at present;

[0029] If the target 5G message development template does not exist, obtaining the industry type of the 5G message new building request.

[0030] In addition, to achieve the above-mentioned purposes, the present application also provides a 5G message development template generation device, which comprises a memory, a processor, and a 5G message development template generation program stored in the memory and executable on the processor, wherein the 5G message development template generation program implements the steps of the aforementioned 5G message development template generation method when executed by the processor.

[0031] In addition, to achieve the above-mentioned purposes, the present application also provides a storage medium having a 5G message development template generation program stored thereon, wherein the 5G message development template generation program implements the steps of the aforementioned 5G message development template generation method when executed by a processor.

[0032] In addition, to achieve the above-mentioned purposes, the present application also provides a computer program product comprising a computer program, wherein the computer program implements the steps of the aforementioned 5G message development template generation method when executed by a processor.

[0033] This invention obtains a dataset of historical 5G message requirements and a dataset of 5G message development templates based on the industry type of new 5G message requests. Then, each requirement information in the 5G message requirement dataset is input into the LSTM encoding module of a neural network for processing to obtain a context vector. This context vector is then input into the LSTM decoding module of the neural network for processing to obtain trained development template information. Based on the development template information and the trained development template information, the optimal weights of the neural network are determined, and the target 5G message development template is determined based on these optimal weights. This allows for the automatic generation of corresponding 5G message development templates based on new 5G message requests, improving the efficiency of 5G message development template creation and reducing labor costs. Simultaneously, the LSTM encoding / decoding module processes 5G message requirement datasets from similar industries to extract common requirements from different new 5G messages within the same industry. Furthermore, the 5G message development templates are optimized based on the 5G message development template datasets from similar industries, thereby improving the universality of the 5G message development templates across similar industries. Attached Figure Description

[0034] Figure 1 This is a schematic diagram of the structure of the 5G message development template generation device in the hardware operating environment involved in the embodiments of the present invention;

[0035] Figure 2 This is a flowchart illustrating the first embodiment of the 5G message development template generation method of the present invention;

[0036] Figure 3 This is a schematic diagram of the structure of the attention mechanism encoder and decoder of the present invention;

[0037] Figure 4 This is a schematic diagram of the structure of the encoding and decoding neural network in the 5G message development template generation method of the present invention;

[0038] Figure 5 This is a schematic diagram of the functional modules of an embodiment of the 5G message development template generation device of the present invention.

[0039] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0040] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0041] like Figure 1 As shown, Figure 1It is a structural schematic diagram of the 5G message development template generation device in the hardware running environment related to the embodiment scheme of the application.

[0042] The 5G message development template generation device can be a PC, a smart phone or a tablet computer.

[0043] As shown in the figure, Figure 1 The 5G message development template generation device can include a processor 1001 such as a CPU, a network interface 1004, a user interface 1003, a memory 1005 and a communication bus 1002. The communication bus 1002 is used to realize the connection and communication between the components. The user interface 1003 can include a display screen (Display), an input unit such as a keyboard (Keyboard), and the optional user interface 1003 can also include a standard wired interface, a wireless interface. The network interface 1004 can optionally include a standard wired interface, a wireless interface (such as a WI-FI interface). The memory 1005 can be a high-speed RAM memory or a stable memory (non-volatile memory) such as a disk memory. The memory 1005 can also be an independent storage device from the aforementioned processor 1001.

[0044] Optionally, the 5G message development template generation device can also include a camera, an RF (Radio Frequency, Radio Frequency) circuit, a sensor, an audio circuit, a WiFi module, etc. The sensor can be a light sensor, a motion sensor and other sensors.

[0045] Those skilled in the art can understand, Figure 1 The terminal structure shown in the figure does not constitute a limitation on the 5G message development template generation device, and can include more or fewer components than the figure, or combine certain components, or different component arrangements.

[0046] As shown in the figure, Figure 1 The memory 1005 as a computer storage medium can include an operating system, a network communication module, a user interface module and a 5G message development template generation program.

[0047] In the 5G message development template generation device shown in the figure, Figure 1 The network interface 1004 is mainly used to connect the background server and communicate data with the background server; the user interface 1003 is mainly used to connect the client (user end) and communicate data with the client; and the processor 1001 can be used to call the 5G message development template generation program stored in the memory 1005.

[0048] In the embodiment, the 5G message development template generation device comprises a memory 1005, a processor 1001 and a 5G message development template generation program stored in the memory 1005 and executable on the processor 1001, wherein the processor 1001 invokes the 5G message development template generation program stored in the memory 1005 and performs the steps of the 5G message development template generation method in each embodiment.

[0049] The application further provides a 5G message development template generation method, which refers to Figure 2 , Figure 2 The flowchart of the first embodiment of the method of the application is shown.

[0050] The 5G message development template generation method of the embodiment is applied to a 5G message open platform, which is located between an industry client and a MaaP (Messaging as a Platform) system to help the industry client to realize A2P communication in multiple scenarios on demand. An enterprise (industry client) can quickly complete the deployment of a message application through the platform without complex code development, so that the industry client can simply and conveniently create its own 5G message application, i.e., develop a 5G message application through a 5G message development template provided by the 5G message open platform. The MaaP system comprises a MaaP platform management module and a MaaP platform. The MaaP system is a core network element of an industry 5G message service, which will provide 5G commercial message service access and message uplink and downlink capabilities for the industry client, and provide industry chat robot search, detail query, message uplink and downlink and other functions for users.

[0051] The 5G message development template generation method comprises the following steps in the embodiment:

[0052] In step S101, the 5G message demand data set and the 5G message development template data set are obtained based on the industry type of the 5G message new creation request.

[0053] In the embodiment, the user (industry client) can submit a 5G message new creation request to the 5G message open platform through the industry client. After receiving the 5G message new creation request, the industry type based on the 5G message new creation request is determined, and the historical new 5G message corresponding to the industry type is obtained. The historical demand information of the historical new 5G message and the industry message development template are used to process the historical demand information and the industry message development template to obtain the 5G message demand data set and the 5G message development template data set.

[0054] The number of historical newly-built 5G messages matches the number of LSTM encoding layers of the LSTM encoding module of the coding-decoding neural network. Specifically, the training set and the test set are obtained through the historical newly-built 5G messages. The 5G message demand data set and the 5G message development template data set of the embodiment are both training sets. The number of demand information in the 5G message demand data set is equal to the number of LSTM encoding layers, and the number of demand information is 50% to 80% of the number of historical newly-built 5G messages. For example, if the number of demand information is 80% of the number of historical newly-built 5G messages, the demand information obtained from the other 20% of the historical newly-built 5G messages is the demand data test set and the development template data test set.

[0055] In step S102, each demand information in the 5G message demand data set is input into the LSTM encoding module of the coding-decoding neural network for processing to obtain a context vector, and the context vector is input into the LSTM decoding module of the coding-decoding neural network for processing to obtain the trained development template information.

[0056] The coding-decoding neural network is an LSTM (long short-term memory) neural network, which includes an LSTM encoding module and an LSTM decoding module. The LSTM encoding module includes N LSTM layer encodings. LSTM is a special type of recurrent neural network. The so-called recurrent neural network is that the same neural network is repeatedly used. LSTM can learn long-term dependency information and remember long-term information by controlling the time value saved in the cache, and is suitable for long sequence learning. Each neuron has four inputs and one output, and each neuron has a Cell to store memory values. Each LSTM neuron contains three gates: a forget gate, an input gate, and an output gate. Long short-term memory neural network has good effect on long sequence learning.

[0057] The LSTM uses an attention mechanism to solve the limitations of the encoder-decoder structure. It provides more rich context obtained from the encoder to the decoder. The encoder transmits more data to the decoder. In the traditional model, the encoder only transmits the last hidden state of the encoding stage, while in the attention mechanism model, the encoder transmits all hidden states to the decoder. At the same time, attention provides a learning mechanism that allows the decoder to learn where to focus in a more rich context when predicting the sequence of outputs at each time step. The attention network assigns an attention weight to each input. The closer the input is to the current operation, the closer the attention weight is to 1, and vice versa. These attention weights are recalculated at each output step.

[0058] Referring to Figure 3 , Figure 3 In the formula, T x is the number of input time steps, T y is the number of output time steps, attention i is the attention weight at the output time step i; c i is the context at the output time step i. The attention weight attention i = softmax(dense(x, y i-1 )), the weight length is Tx, and the sum of all weights is 1; the context is the sum of the attention weight and the input, that is, the formula of the context is:

[0059]

[0060] In this embodiment, after obtaining the 5G message requirement data set and the 5G message development template data set, each requirement information in the 5G message requirement data set is input into the LSTM encoding module of the encoding and decoding neural network to obtain the output context vector of each LSTM layer in the LSTM encoding module, that is, the context vector corresponding to each requirement information, and then the context vectors corresponding to each requirement information are merged through the LSTM encoding module to obtain a context merging vector, that is, the context vector output by the LSTM encoding module. Finally, the context merging vector is input into the LSTM decoding module of the encoding and decoding neural network to take the output data of the LSTM decoding module as the trained development template information.

[0061] In step S103, based on each development template information in the development template information set and the trained development template information, the optimal weight of the encoding and decoding neural network is determined, and the target 5G message development template is determined based on the optimal weight.

[0062] In this embodiment, after obtaining the trained development template information, the optimal weight of the encoding and decoding neural network is determined according to each development template information in the development template information set and the trained development template information. Specifically, first, the loss function is determined according to each development template information in the development template information set and the trained development template information, and the minimum value of the loss function is determined by using the gradient descent algorithm. The weight corresponding to the minimum value is taken as the optimal weight of the encoding and decoding neural network. The target 5G message development template is determined according to the optimal weight, that is, the optimal weight is taken as the weight of the 5G message development template to obtain the target 5G message development template, and then the target 5G message development template is stored in the 5G message open platform.

[0063] It should be noted that when the 5G message open platform stores the target 5G message development template, the 5G message open platform sends the information that the template has been created to the terminal corresponding to the 5G message new request.

[0064] The 5G message development template generation method provided in the embodiment, by the industry type based on the 5G message new request, obtains the 5G message demand data set and the 5G message development template data set corresponding to the historical new 5G message; then each demand information in the 5G message demand data set is input into the LSTM encoding module of the encoding-decoding neural network respectively, processed to obtain the context vector, and the context vector is input into the LSTM decoding module of the encoding-decoding neural network for processing to obtain the trained development template information; then based on each development template information of the development template information set and the trained development template information, the optimal weight of the encoding-decoding neural network is determined, and the target 5G message development template is determined based on the optimal weight, which can automatically generate the corresponding 5G message development template according to the 5G message new request, to automatically generate the 5G message development template, improve the production efficiency of the 5G message development template, and reduce the labor cost in the process of the 5G message development template.

[0065] At the same time, the LSTM encoding-decoding module is used to process the 5G message demand data set of the same industry to extract the common demand of different new 5G messages in the same industry, and the 5G message development template is optimized according to the 5G message development template data set of the same industry, thereby improving the universality of the 5G message development template in the same industry.

[0066] Based on the first embodiment, a second embodiment of the 5G message development template generation method of the application is provided, in which the step S102 comprises:

[0067] In step S101, each demand information is input into the embedding layer of the LSTM encoding module to obtain the demand vector corresponding to each demand information.

[0068] In step S102, each demand vector is input into the LSTM encoding layer of the corresponding LSTM encoding module to obtain the context demand vector corresponding to each demand vector.

[0069] In step S103, the context demand vector corresponding to each demand information is input into the merging layer to obtain the context vector.

[0070] Reference Figure 4 , Figure 4 The model structure diagram of the encoding-decoding neural network in the embodiment is shown in the following figure:

[0071] The first layer is an input layer, which inputs N pieces of demand information in the 5G message demand dataset, and each information sequence has a length of L. Therefore, the shape of the output data of this layer is (None, L).

[0072] The second layer is an embedding layer, which converts each word in the demand information into a demand information vector through word embedding. The input data dimension is need_vocab_size, and the output is set to convert the word into a 128-dimensional spatial vector. The input sequence length is L. Therefore, the shape of the output data of this layer is (None, L, 128). The function of this embedding layer is to map the words in the demand information into vectors, and convert the index of each word into a fixed shape vector of 128 dimensions.

[0073] The third layer is an LSTM encoding layer, which includes N LSTM layers in parallel, each LSTM layer containing 128 LSTM neurons (encoder LSTM), and the activation function is set to “relu”. The shape of the output data of this layer is (None, L, 128), and N demand information vectors are encoded into N context demand vectors of fixed length.

[0074] The fourth layer is a merge layer (concatenate): the context demand vectors are concatenated and merged into a context vector h of fixed length according to the column dimension, which is the context vector output by the LSTM encoding module.

[0075] The fifth layer is an attention LSTM decoding layer, which contains 128 LSTM neurons (attention encoder LSTM), and the activation function is set to “relu”. The shape of the output data of this layer is (None, L, 128); it converts the context vector output by the LSTM encoding module into intermediate data.

[0076] The sixth layer is a fully connected (Dense) layer (output layer): it contains Dense fully connected neurons, and the number of template_vocab_size, and the activation function is set to “softmax”. The softmax output result is sent to the multi-class cross-entropy loss function. The shape of the output data of this fully connected layer is (None, template_vocab_size), and the output data of the fully connected layer is the developed template information after training.

[0077] That is, the LSTM encoding module includes the second layer, the third layer, and the fourth layer.

[0078] In the embodiment, after obtaining the 5G message requirement data set and the 5G message development template data set, each requirement information of the 5G message requirement data set is input into an embedding layer of an LSTM encoding module, and each word in the requirement information is mapped into a vector through the embedding layer, so that the index of each word in the requirement information is converted into a vector, to obtain a requirement vector corresponding to each requirement information. Then, each requirement vector is input into an LSTM encoding layer of a corresponding LSTM encoding module, that is, each requirement vector is input into an LSTM layer, and the output of each LSTM layer is used as a context requirement vector corresponding to the requirement vector, and then N requirement information vectors are encoded into N context requirement vectors with a fixed length. Finally, each context requirement vector is input into a merging layer to splice and merge the context requirement vectors in the column dimension, to obtain a context vector with a fixed length, which is the output of the LSTM encoding module.

[0079] The 5G message development template generation method provided in the embodiment encodes the 5G message requirement data set into a context vector through the LSTM encoding module, and then decodes the context vector through the LSTM decoding module, to automatically generate the 5G message development template, so that the production efficiency of the 5G message development template is further improved.

[0080] Based on the second embodiment, a third embodiment of the 5G message development template generation method is provided, in which step S102 includes:

[0081] In step S301, the context vector is input into an LSTM decoding layer of the LSTM decoding module, to obtain intermediate data output by the LSTM decoding layer.

[0082] In step S302, the intermediate data is input into a full connection layer of the LSTM decoding module, to obtain the trained development template information.

[0083] Referring to Figure 4 The LSTM decoding module includes a fifth layer and a sixth layer.

[0084] In the embodiment, after the context vector is obtained, the context vector is input into an LSTM decoding layer of the LSTM decoding module, the context vector is converted into intermediate data by the LSTM decoding layer, and the intermediate data is input into a full connection layer to obtain output data of the full connection layer, i.e., the trained development template information, and then the automatic generation of the 5G message development template is realized, and the production efficiency of the 5G message development template is improved.

[0085] The 5G message development template generation method provided in the embodiment can automatically generate a corresponding 5G message development template according to a context vector corresponding to a 5G message new creation request, to automatically generate the 5G message development template, and further improve the production efficiency of the 5G message development template and reduce the labor cost in the process of the 5G message development template.

[0086] Based on the first embodiment, a fourth embodiment of the 5G message development template generation method is provided, and in the fourth embodiment, step S103 comprises:

[0087] In step S401, a loss function is determined based on each development template information in the development template information set and the trained development template information.

[0088] In step S402, an optimal weight of the coding-decoding neural network is determined based on the loss function.

[0089] In the embodiment, after the trained development template information is obtained, a loss function is determined based on each development template information in the development template information set and the trained development template information. Specifically, the trained development template information is compared with each development template information in the development template information set to calculate the loss function, and the loss function is cross-entropy loss.

[0090] Further, in an embodiment, step S402 comprises:

[0091] The optimal loss of the loss function is determined based on the gradient descent algorithm, and the optimal weight is determined based on the optimal loss.

[0092] In the embodiment, the gradient descent optimization algorithm selects an adam optimizer to improve the learning speed of the traditional gradient descent. The coding-decoding neural network can find the optimal weight value that minimizes the loss function by the gradient descent algorithm, and the coding-decoding neural network can autonomously learn the optimal weight by training.

[0093] It should be noted that after the optimal weight is obtained, the target 5G message development template is determined according to the optimal weight, and then it is judged again whether the target 5G message development template converges. Specifically, the demand data test set is input into the target 5G message development template to obtain the loss function value of the target 5G message development template. If the loss function value is less than a preset threshold, it is determined that the target 5G message development template converges, and then the target 5G message development template is stored in the 5G message open platform.

[0094] The 5G message development template generation method provided in the embodiment determines the loss function based on each development template information in the development template information set and the trained development template information, and then determines the optimal weight of the coding and decoding neural network based on the loss function. The optimal weight can be accurately obtained according to the development template information set and the trained development template information, and the production efficiency of the 5G message development template is further improved.

[0095] Based on the first embodiment, a fifth embodiment of the 5G message development template generation method of the present application is provided. In the present embodiment, step S101 comprises:

[0096] In step S501, historical demand information corresponding to historical newly created 5G messages and industry message development templates are obtained.

[0097] In step S502, data processing operations are respectively performed on the historical demand information and the industry message development templates to obtain a 5G message demand data set and a 5G message development template data set.

[0098] In the present embodiment, a user (industry client) can submit a 5G message new creation request to the 5G message open platform through an industry client. After receiving the 5G message new creation request, the industry type based on the 5G message new creation request is determined, and the historical demand information corresponding to the historical newly created 5G messages and the industry message development templates corresponding to the industry type are obtained. Then, data processing operations are performed on the historical demand information and the industry message development templates to obtain a 5G message demand data set and a 5G message development template data set.

[0099] Specifically, step S502 comprises:

[0100] In step a, text cleaning and text serialization processing are respectively performed on the historical demand information to obtain the 5G message demand data set.

[0101] In step b, text cleaning and text serialization processing are respectively performed on the industry message development templates to obtain the 5G message development template data set.

[0102] The historical demand information is the function required to be implemented by a historical newly-built 5G message, for example, a certain education industry message is: to collect and analyze the daily health status report information of students and teachers, and to issue epidemic prevention health guidance to students and teachers. The i-th historical demand information can be represented as {S1 i , S2 i , S3 i , …, SLi}. The industry message development template includes card size, card style, button number, floating menu information, fixed menu information, AI interaction model information, and pre-integrated network capability information, and is presented in the form of XML code. The industry message development template can be represented as {P1, P2, P3, …, P M}.

[0103] In this embodiment, text cleaning and text serialization processing are performed on each historical demand information to remove all punctuation marks in each historical demand information. If the historical demand information is in Chinese, the text is segmented, if the text is in English, the letters are uniformly converted to lowercase, and each word is tokenized, so that each piece of text is converted into a piece of index number. Then the maximum text length in the processed historical demand information is determined, and the processed historical demand information that does not reach the maximum text length is zero-padded to obtain the 5G message demand data set, and the size of the dictionary is need_vocab_size.

[0104] At the same time, text cleaning and text serialization processing are performed on each industry message development template to remove all punctuation marks in each industry message development template. If the industry message development template is in Chinese, the text is segmented, if the text is in English, the letters are uniformly converted to lowercase, and each word is tokenized, so that each piece of text is converted into a piece of index number. Then the maximum text length in the processed historical demand information is determined, and the processed historical demand information that does not reach the maximum text length is zero-padded to obtain the 5G message development template data set, and the size of the dictionary is template_vocab_size.

[0105] The 5G message development template generation method proposed in this embodiment can accurately obtain the 5G message demand data set and the 5G message development template data set according to the historical demand information and the industry message development template by obtaining the historical demand information corresponding to the historical newly-built 5G message and the industry message development template, and then performing data processing operations on the historical demand information and the industry message development template to obtain the 5G message demand data set and the 5G message development template data set, thereby further improving the production efficiency of the 5G message development template.

[0106] Based on the above various embodiments, a sixth embodiment of the 5G message development template generation method is proposed, in which, before step S101, the 5G message development template generation method further comprises:

[0107] Step S601, upon receiving a 5G message new request, obtaining new 5G message requirement information corresponding to the 5G message new request;

[0108] Step S602, based on the new 5G message requirement information, determining whether the target 5G message development template corresponding to the 5G message new request exists currently;

[0109] Step S603, if the target 5G message development template does not exist, obtaining the industry type of the 5G message new request.

[0110] In this embodiment, upon receiving a 5G message new request, the new 5G message requirement information corresponding to the 5G message new request is obtained, and based on the new 5G message requirement information, it is determined whether the target 5G message development template corresponding to the 5G message new request exists currently; if the target 5G message development template does not exist, the industry type of the 5G message new request is obtained, so as to facilitate the making of the target 5G message development template, avoid the repeated development of the target 5G message development template, and improve the resource utilization rate. If the target 5G message development template exists, the 5G message corresponding to the 5G message new request is generated through the target 5G message development template.

[0111] The 5G message development template generation method proposed in this embodiment, by obtaining the new 5G message requirement information corresponding to the 5G message new request upon receiving the 5G message new request, then based on the new 5G message requirement information, determining whether the target 5G message development template corresponding to the 5G message new request exists currently, and then if the target 5G message development template does not exist, obtaining the industry type of the 5G message new request, by judging whether the target 5G message development template exists, the accuracy of developing the target 5G message development template is improved, and the resource utilization rate is improved.

[0112] The application also provides a 5G message development template generation device, referring to Figure 5 , the 5G message development template generation device comprises:

[0113] The acquisition module 10 is used for acquiring the 5G message requirement data set and the 5G message development template data set corresponding to the historical new 5G message based on the industry type of the 5G message new request;

[0114] The training module 20 is configured to input each requirement information in the 5G message requirement data set into an LSTM encoding module of the encoding-decoding neural network respectively for processing to obtain a context vector, and input the context vector into an LSTM decoding module of the encoding-decoding neural network for processing to obtain the trained development template information.

[0115] The generation module 30 is configured to determine an optimal weight of the encoding-decoding neural network based on each development template information in the development template information set and the trained development template information, and determine a target 5G message development template based on the optimal weight.

[0116] The method performed by each program unit can refer to each embodiment of the 5G message development template generation method, and will not be described here.

[0117] The application further provides a storage medium, i.e., a computer readable storage medium. The 5G message development template generation program is stored on the storage medium of the application, and the 5G message development template generation program is executed by the processor to realize the steps of the 5G message development template generation method as described above.

[0118] The method realized by the 5G message development template generation program executed by the processor can refer to each embodiment of the 5G message development template generation method, and will not be described here.

[0119] In addition, the embodiment of the application further provides a computer program product, and the computer program product includes the 5G message development template generation program. The 5G message development template generation program is executed by the processor to realize the steps of the 5G message development template generation method as described above.

[0120] It should be noted that in this document, the term "comprising" or "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such a process, method, article or system. Without more limitations, the element defined by the statement "including a" does not exclude the presence of additional identical elements in the process, method, article or system including the element.

[0121] The above-mentioned embodiment numbers of the application are only for description, and do not represent the advantages and disadvantages of the embodiments.

[0122] Those skilled in the art can clearly understand the above-mentioned embodiment method can be realized by means of software and the necessary general hardware platform, of course, can also be through hardware, but in many cases the former is a better embodiment. Based on such understanding, the technical solutions of the present application essentially or say the part of the prior art contribution can be embodied in the form of software products, the computer software product is stored in a storage medium (such as ROM / RAM, magnetic disc, optical disc) as described above, including a number of instructions to make a terminal device (may be a mobile phone, computer, server, air conditioner, or network equipment, etc.) executes the method described in various embodiments of the present application.

[0123] The above is only the preferred embodiment of the present application, not therefore limit the patent scope of the present application, any equivalent structure or equivalent flow transformation made by using the content of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A method for generating 5G message development templates, characterized in that, The 5G message development template generation method comprises the following steps: Based on the industry type of the 5G message new request, the 5G message demand data set and the 5G message development template data set corresponding to the historical new 5G message are obtained; Each demand information in the 5G message demand data set is input into the LSTM encoding module of the encoding-decoding neural network for processing to obtain a context vector, and the context vector is input into the LSTM decoding module of the encoding-decoding neural network for processing to obtain the trained development template information, wherein the output data of the LSTM decoding module is used as the trained development template information; Based on each development template information of the development template information set and the trained development template information, the optimal weight of the encoding-decoding neural network is determined, and the target 5G message development template is determined based on the optimal weight. 2.The 5G message development template generation method of claim 1, wherein, The step of inputting each demand information in the 5G message demand data set into the LSTM encoding module of the encoding-decoding neural network for processing comprises: Each demand information is input into the embedding layer of the LSTM encoding module to obtain a demand vector corresponding to each demand information; Each demand vector is input into the LSTM encoding layer of the corresponding LSTM encoding module to obtain a context demand vector corresponding to each demand vector; The context demand vector corresponding to each demand information is input into the merging layer to obtain the context vector. 3.The 5G message development template generation method of claim 2, wherein, The step of inputting the context vector into the LSTM decoding module of the encoding-decoding neural network for processing to obtain the trained development template information comprises: The context vector is input into the LSTM decoding layer of the LSTM decoding module to obtain intermediate data output by the LSTM decoding layer; The intermediate data is input into the fully connected layer of the LSTM decoding module to obtain the trained development template information. 4.The method of claim 1, wherein, The step of determining the optimal weight of the encoding-decoding neural network based on each development template information of the development template information set and the trained development template information comprises: Based on each development template information of the development template information set and the trained development template information, a loss function is determined; Based on the loss function, the optimal weight of the encoding-decoding neural network is determined. 5.The method of claim 1, wherein, The step of obtaining the 5G message demand data set and the 5G message development template data set corresponding to the historical new 5G message comprises: Obtain historical demand information and industry message development templates corresponding to historical new 5G messages; Respectively perform data processing operations on the historical demand information and the industry message development templates to obtain the 5G message demand data set and the 5G message development template data set. 6.The method of claim 5, wherein, The step of respectively performing data processing operations on the historical demand information and the industry message development templates to obtain the 5G message demand data set and the 5G message development template data set comprises: Respectively perform text cleaning and text serialization processing on the historical demand information to obtain the 5G message demand data set; Text cleaning and text serialization are respectively performed on the industry message development template to obtain the 5G message development template dataset. 7.The method of claim 1 to 6, wherein, Before the step of obtaining the 5G message demand dataset and the 5G message development template dataset corresponding to the historical newly-built 5G message based on the industry type of the 5G message new building request, the method further comprises: Upon receiving the 5G message new building request, obtaining new 5G message demand information corresponding to the 5G message new building request; Based on the new 5G message demand information, determining whether a target 5G message development template corresponding to the 5G message new building request exists; If the target 5G message development template does not exist, obtaining the industry type of the 5G message new building request. 8.A 5G message development template generation device, characterized by comprising: The 5G message development template generation device comprises a memory, a processor, and a 5G message development template generation program stored on the memory and executable on the processor, and the 5G message development template generation program, when executed by the processor, implements the steps of the 5G message development template generation method according to any one of claims 1 to 7.

9. A storage medium, characterized by The storage medium stores a 5G message development template generation program, and the 5G message development template generation program, when executed by the processor, implements the steps of the 5G message development template generation method according to any one of claims 1 to 7.

10. A computer program product comprising a computer program, characterized in that, The computer program, when executed by the processor, implements the steps of the 5G message development template generation method according to any one of claims 1 to 7.

Images