An advertising copy generation method and system

Abstract

The present specification provides a copywriting method and system, by periodically acquiring description information of a target applet, extracting keywords based on the description information of the target applet, and automatically generating a target copywriting matched with the target applet according to the extracted keywords, the generated target copywriting can well present the function of the applet to attract users to click and use.

Description

Technical Field

[0001] This specification relates to the field of data processing technology, and in particular to a method and system for generating text. Background Technology

[0002] A badge is an advertising space within an application (APP) that displays and recommends relevant internal business activities or product features and services to users. Badge text can showcase the functions of a mini-program and attract user clicks. However, currently only a few mini-programs have corresponding badge text, and configuring badge text is time-consuming and labor-intensive.

[0003] Therefore, there is a need to provide a more efficient method and system for generating copy. Summary of the Invention

[0004] The main purpose of this manual is to provide a method and system for generating copy.

[0005] Firstly, this specification provides a text generation method for placing badges on mini-programs in an application interface, comprising: obtaining descriptive information of a target mini-program according to a preset period; determining at least one keyword based on the descriptive information, wherein the at least one keyword reflects the characteristic information of the target mini-program; determining the target text of the target mini-program based on the at least one keyword; and outputting the target text.

[0006] In some embodiments, the descriptive information includes at least one of the following: the name of the target mini-program, the slogan of the target mini-program, or the functional description of the target mini-program.

[0007] In some embodiments, the keyword has three or fewer characters; and the part of speech of the keyword includes at least one of verb, noun, or compound phrase.

[0008] In some embodiments, determining the target text of the target mini-program based on the at least one keyword includes: for each of the at least one keyword: determining the part-of-speech and number of characters of the keyword, determining that the keyword meets preset conditions based on the part-of-speech and the number of characters, and using the keyword as a first candidate text of the target mini-program, and determining a first confidence score of the first candidate text, the first confidence score reflecting the credibility of the first candidate text; and determining the target text of the target mini-program based on at least one first candidate text and at least one first confidence score.

[0009] In some embodiments, determining the target text of the target mini-program based on the at least one keyword includes: for each of the at least one keyword: determining the part-of-speech tag and number of characters of the keyword; determining that the keyword does not meet the preset condition based on the part-of-speech tag and the number of characters; performing text matching on the keyword based on a text generation model to determine a second alternative text of the target mini-program; and determining a second confidence score for the second alternative text, the second confidence score reflecting the credibility of the second alternative text; and determining the target text of the target mini-program based on at least one second alternative text and at least one second confidence score.

[0010] In some embodiments, the step of performing text matching on the keywords based on a text generation model to determine a second candidate text for the target mini-program, and determining a second confidence score for the second candidate text, includes: predicting verbs that match the keywords and the probability values ​​of the verbs based on the text generation model, wherein the verbs are one character; and combining the verbs and the keywords to form the second candidate text, and using the probability values ​​of the verbs as the second confidence score of the second candidate text.

[0011] In some embodiments, the preset conditions include: the keyword is a verb and the keyword has two characters; or the keyword has three characters.

[0012] In some embodiments, determining the target text of the target mini-program based on at least one first alternative text and at least one first confidence score includes: randomly selecting a first alternative text from the at least one first alternative text as the target text, and using the first confidence score of the first alternative text as the target confidence score of the target text.

[0013] In some embodiments, determining the target text of the target mini-program based on at least one second alternative text and at least one second confidence score includes: selecting the second alternative text with the highest confidence score from the at least one second alternative text as the target text, and using the second confidence score of the second alternative text as the target confidence score of the target text.

[0014] In some embodiments, the first confidence score is higher than the second confidence score.

[0015] In some embodiments, outputting the target text includes: obtaining a first review result of the target text, the first review result including: the target text is in an available state or the target text is in an unavailable state; and outputting the target text based on the first review result.

[0016] In some embodiments, the step of outputting the target text based on the first review result includes: determining that the first review result indicates that the target text is in an available state, using the target text as the badge text of the target mini-program, and outputting the badge text.

[0017] In some embodiments, after determining that the first review result indicates that the target document is available, the method further includes: saving the target document.

[0018] In some embodiments, the step of outputting the target text based on the first review result includes: determining that the first review result indicates that the target text is in an unavailable state, obtaining manually configured text, using the manually configured text as the badge text of the target mini-program, and outputting the badge text.

[0019] In some embodiments, after obtaining the first review result of the target text, the method further includes: obtaining the feedback result of the target text, wherein the feedback result includes a negative feedback result corresponding to the target text being in an unavailable state.

[0020] In some embodiments, after obtaining the first review result of the target text, the method further includes: adjusting the parameters of the text generation model based on the negative feedback result to update the target text.

[0021] In some embodiments, the copy generation method further includes: obtaining a second review result for the at least one keyword; and updating the at least one keyword based on the second review result.

[0022] Secondly, this disclosure provides a copywriting generation system, comprising: at least one storage medium including at least one instruction set for implementing and analyzing a copywriting generation method; and at least one processor communicatively connected to the at least one storage medium, wherein, when the system is running, the at least one processor reads the at least one instruction set and executes the method described above according to the instructions of the at least one instruction set.

[0023] As can be seen from the above technical solutions, the copywriting generation method and the system that execute this method provided in this specification periodically acquire the description information of the target mini-program, extract keywords based on the description information, and automatically generate target copywriting that matches the target mini-program based on the extracted keywords. The generated target copywriting can effectively present the functions of the mini-program to attract users to click and use it. The method and system can not only achieve large-scale and rapid matching of target copywriting for mini-programs and improve the efficiency of mini-program badge copywriting generation, but also enable periodic updates of the target files of the target mini-program, and timely release of corresponding target copywriting in different operating cycles.

[0024] Other functions of the copy generation methods and systems provided in this specification will be partially listed in the following description. The figures and examples described below will be obvious to those skilled in the art. The inventive aspects of the copy generation methods and systems provided in this specification can be fully explained through practice or use of the methods, apparatus, and combinations described in the detailed examples below. Attached Figure Description

[0025] To more clearly illustrate the technical solutions in the embodiments of this specification, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0026] Figure 1 A schematic diagram illustrating an application scenario of a copywriting generation method system according to some embodiments of this specification is shown.

[0027] Figure 2 A schematic diagram showing the interface of an application on a client according to some embodiments of this specification is shown;

[0028] Figure 3 This document illustrates a badge delivery scenario in a client application according to some embodiments of this specification.

[0029] Figure 4 This document illustrates a badge delivery scenario in a client application according to some embodiments of this specification.

[0030] Figure 5 A schematic diagram of the structure of a computing device provided according to some embodiments of this specification is shown;

[0031] Figure 6 A flowchart illustrating a text generation method according to some embodiments of this specification is shown.

[0032] Figure 7 The diagram illustrates a process for generating target text according to some embodiments of this specification; and

[0033] Figure 8 A schematic diagram illustrating the working principle of a text generation model provided according to some embodiments of this specification is shown. Detailed Implementation

[0034] The following description provides specific application scenarios and requirements for this specification, intended to enable those skilled in the art to make and use the contents of this specification. Various partial modifications to the disclosed embodiments will be apparent to those skilled in the art, and the general principles defined herein can be applied to other embodiments and applications without departing from the spirit and scope of this specification. Therefore, this specification is not limited to the embodiments shown, but rather to the widest scope consistent with the claims.

[0035] The terminology used herein is for the purpose of describing particular exemplary embodiments only and is not restrictive. For example, unless the context clearly indicates otherwise, the singular forms “a,” “an,” and “the” used herein may also include the plural forms. When used in this specification, the terms “comprising,” “including,” and / or “containing” mean that the associated integers, steps, operations, elements, and / or components are present, but do not exclude the presence of one or more other features, integers, steps, operations, elements, components, and / or groups, or that other features, integers, steps, operations, elements, components, and / or groups may be added to the system / method.

[0036] Considering the following description, these and other features of this specification, as well as the operation and function of the related components of the structure, and the economy of assembly and manufacture of the parts, can be significantly improved. All of these form part of this specification with reference to the accompanying drawings. However, it should be clearly understood that the drawings are for illustrative and descriptive purposes only and are not intended to limit the scope of this specification. It should also be understood that the drawings are not drawn to scale.

[0037] The flowcharts used in this specification illustrate operations implemented according to some embodiments of this specification. It should be clearly understood that the operations in the flowcharts may not be implemented in a sequential order. Instead, the operations may be implemented in reverse order or simultaneously. Furthermore, one or more additional operations may be added to the flowcharts. One or more operations may be removed from the flowcharts.

[0038] For ease of description, the terms that will appear in the following descriptions will be explained as follows.

[0039] Application: Generally refers to mobile phone software, the English name is Mobile APP (Application). It mainly refers to the software installed on smartphones to improve the shortcomings of the original system and personalize it.

[0040] Mini Program: The English name is Mini Program (or Mini Application), which is an application that can be used without downloading or installing.

[0041] Figure 1 This diagram illustrates an application scenario of a copywriting generation method system 100 according to some embodiments of this specification. The copywriting generation method described in this specification is applied to the placement of badges in mini-programs within an application interface. The badges described in this specification are advertising resource positions within an application that display and recommend internal business-related activities or product feature services to users. The copywriting generation method system 100 (hereinafter referred to as system 100) in this specification may include a client 110, an integrated development platform server 130, a database 140, and an operations terminal 150.

[0042] Integrated Development Platform (IDE) 120 is mounted on IDE server 130. IDE 120, also known as Integrated Development Environment (IDE), is an application that provides a program development environment, generally including tools such as a code editor, compiler, debugger, and graphical user interface. Developers can write program code (i.e., program development) on IDE 120. IDE server 130 (hereinafter referred to as server 130) can be a computing device on IDE 120 specifically used to handle document generation methods.

[0043] Client 110 may include at least one mobile terminal device, which may be a mobile phone, tablet, or other device capable of installing and using applications. Users can use client 110 to download and use various applications. Figure 2 This diagram illustrates the user interface of an application on client 110 according to some embodiments of this specification. The applications described in this specification can be of various types. For example... Figure 2 As shown, the application interface can include multiple display grids (represented by dashed boxes), each containing different function buttons, such as recharge, bill payment, and express delivery. Generally, each display grid contains one function button, and each function button is displayed using an icon and text. For ease of description, Figure 2Section A represents icons, and Section B represents text. Section A can display the icon corresponding to the function button, and Section B can display the name of the function button.

[0044] In this specification, a display grid within the application can aggregate at least one mini-program that the user has favorited and / or recently used. For ease of description, the display grid that can aggregate at least one mini-program is referred to as the mini-program display grid. A mini-program display grid can display at least one mini-program; for example, it can display one, two, three, four, or more mini-programs. Each mini-program has a corresponding mini-program icon, which reflects the mini-program's functions and other characteristics for user identification. For example, Figure 2 In the grid display of mini-programs, area A contains four mini-program icons, labeled A1, A2, A3, and A4, with each icon corresponding to a mini-program.

[0045] It should be understood that Figure 2 The number of display grids can be one or more, and the size of each display grid can be the same or different. Figure 2 The display grid for the mini-program can be any display grid. Figure 2 The application interface diagrams shown do not limit the actual application interface. Figure 2 The number of mini-programs displayed in the mini-program display grid does not limit the actual number of mini-programs displayed.

[0046] Due to the wide variety of mini-programs, there are correspondingly many types and numbers of mini-program icons, making it difficult for users to accurately identify mini-programs directly from the icons. Especially when multiple mini-program icons are displayed simultaneously in a single display grid, the limited display area of ​​the grid causes the mini-program icons to be passively shrunk, hindering user identification. In this manual, server 130 can configure corresponding badge text for mini-programs and display them in the mini-program display grid using a combination of badge text and icon.

[0047] Figure 3 A schematic diagram of a badge delivery scenario in a client 110 application according to some embodiments of this specification is shown. In some embodiments, such as Figure 3 As shown, server 130 configures corresponding badge text for the mini-program, which can be displayed in area A of the mini-program display grid using badge text and an icon. Figure 3 In the mini-program display grid, section A includes a mini-program icon. A corresponding badge text (corresponding to) is added to the upper right corner of the mini-program icon in section A. Figure 3(C) The shaded area in the image. The badge text can showcase the features of the mini-program through text, helping users to identify the mini-program icon and thus attracting users to click.

[0048] Figure 4 A schematic diagram of a badge delivery scenario in a client 110 application according to some embodiments of this specification is shown. In some embodiments, such as Figure 4 As shown, in Figure 4 In the mini-program display grid, section A includes four mini-program icons: A1, A2, A3, and A4. The superscript text can be displayed in the upper right corner of section A (corresponding to...). Figure 4 (See shaded area C). At this time, the displayed badge text can be the badge text corresponding to any one of the four mini-program icons. For example, the displayed badge text could be the badge text corresponding to mini-program icon A1, the badge text corresponding to mini-program icon A2, the badge text corresponding to mini-program icon A3, and the badge text corresponding to mini-program icon A4. Furthermore, if the user stays on the current screen for a long time, the displayed badge text can switch at preset time intervals (e.g., 1 second, 1.5 seconds, 2 seconds, etc.). For example, if the currently displayed badge text is the badge text corresponding to mini-program icon A3, and the user stays on the current screen for a long time, after 1 second, the displayed badge text could be the badge text corresponding to one of the remaining three mini-program icons (A1, A2, or A4). In the mini-program display grid, the above display method can showcase the mini-program's functions through the mini-program icon and further through corresponding badge text, thereby attracting users to click and enter. To better present the mini-program's functions and attract users to click through the text, the badge text described in this manual can be limited to three characters or less, and the badge text should preferably use a verb-object structure (verb + noun combination).

[0049] It should be understood that the badge text can be placed not only in the upper right corner of the mini program icon, but also in other positions on the mini program icon, as long as both the mini program icon and the corresponding badge text are clearly displayed. Figure 3 , Figure 4 The placement of mini-program icons and corresponding badge text in the mini-program display grid does not limit the actual placement.

[0050] Client 110 may also include at least one application on a mobile terminal device.

[0051] The operation terminal 150 may include at least one mobile terminal device, wherein the at least one mobile terminal device may be a terminal device with computing capabilities such as a computer, mobile phone, or tablet. Operation personnel can use the operation terminal 150 to obtain the results obtained by the document generation method of the server 130 and provide feedback; the operation terminal 150 can return the feedback results provided by the operation personnel to the server 130.

[0052] Server 130 may store data or instructions for performing the document generation method described in this specification, and may execute or be used to execute said data and / or instructions. Server 130 may include hardware devices with data processing capabilities and the necessary programs required to drive the hardware devices. Of course, server 130 may also be merely a hardware device with data processing capabilities, or merely a program running on the hardware device. In some embodiments, server 130 may also be deployed as a plug-in on client 110. Alternatively, server 130 may also be deployed as a plug-in on operator 150.

[0053] Database 140 may store data and / or instructions. In some embodiments, database 140 may store data and / or instructions executed by server 130 or used to execute the document generation methods described herein. Client 110 and server 130 may have access to database 140, and client 110 and server 130 may access data or instructions stored in database 140 via a network. In some embodiments, database 140 may be directly connected to client 110 and server 130. In some embodiments, database 110 may be part of server 130. In some embodiments, database 140 may include mass storage, removable storage, volatile read-write memory, read-only memory (ROM), or similar content, or any combination thereof. Exemplary mass storage may include non-transitory storage media such as disks, optical discs, and solid-state drives. Exemplary removable storage may include flash drives, floppy disks, optical discs, memory cards, zip disks, magnetic tapes, etc. Typical volatile read-write memory may include random access memory (RAM). Example RAMs may include dynamic RAM (DRAM), dual date rate synchronous dynamic RAM (DDRSDRAM), static RAM (SRAM), thyristor RAM (T-RAM), and zero-capacitance RAM (Z-RAM), etc. Exemplary ROMs may include mask ROM (MROM), programmable ROM (PROM), virtual programmable ROM (PEROM), electronically programmable ROM (EEPROM), optical disc (CD-ROM), and digital multifunction disk ROM, etc.

[0054] It should be understood that Figure 1The number of clients 110 and servers 130 shown is merely illustrative. Depending on the implementation requirements, there can be any number of clients 110, servers 130, and operators 150.

[0055] It should be noted that the copy generation method can be executed entirely on the client 110; or entirely on the server 130; or entirely on the operations side 150; or partially on the client 110, partially on the server 130, and partially on the operations side 150.

[0056] For ease of description, the following descriptions will use the execution of the text generation method on server 130 as an example to describe the technical solutions involved in this specification.

[0057] Figure 5 This is a schematic diagram of the structure of a computing device 200 provided according to some embodiments of this specification. The computing device 200 can be a general-purpose computer or a special-purpose computer. For example, the computing device 200 can be a server, a personal computer, a portable computer (such as a laptop computer, tablet computer, etc.), or other electronic devices with computing capabilities. Of course, the computing device can be... Figure 1 The server 130 can also be a terminal device used by multiple developers to develop programs on an integrated development platform.

[0058] like Figure 5 As shown, the computing device 200 may include a COM port 250, which can be connected to or from a network to facilitate data communication. The computing device 200 may also include a processor 220 in the form of one or more processors, such as a central processing unit (CPU), for executing program instructions. The computing device 200 may also include an internal communication bus 210 and various forms of program storage media and data storage media, such as a disk 270 (non-transitory memory) and read-only memory (ROM) 230 or random access memory (RAM) 240, etc., for storing various data files to be processed and / or transmitted. The storage media may be local to the computing device 200 or shared by the computing device 200 (e.g., Figure 1 The computing device 200 may also include program instructions stored in ROM 230, RAM 240, and / or other types of non-transitory storage media to be executed by processor 220. The computing device 200 may also include I / O components 260 to support data communication with other computing devices in the distributed computing system 100. The computing device 200 may also receive programming and data via network communication.

[0059] For illustrative purposes only, only one processor 220 is described in the computing device 200. However, those skilled in the art will understand that the computing device 200 in this specification may also include multiple processors. Therefore, the methods / steps / operations performed by one processor as described in this specification may also be performed jointly or separately by multiple processors. For example, in this specification, the processor of the computing device 200 may simultaneously execute step A and step B. It should be understood that step A and step B may also be performed jointly by two different processors. For example, a first processor executes step A, a second processor executes step B, or a first processor and a second processor jointly execute steps A and B.

[0060] Figure 6 A flowchart 300 of a text generation method according to some embodiments of this specification is shown. The following will be combined with... Figure 6 This specification describes the technical solution. The entity implementing the described technical solution may be... Figure 1 The client 110, server 130, and operator terminal 150 are selected from the above. Specifically, the client 110, server 130, or operator terminal 150 may have the following characteristics: Figure 5 The aforementioned structure, namely, the client 110 and server 130 or operator terminal 150, can be a device for a copywriting generation method, comprising: at least one storage medium and at least one processor. The at least one storage medium includes at least one instruction set for the copywriting generation method. The at least one processor is communicatively connected to the at least one storage medium. When the system is running, the at least one processor can read the at least one instruction set and execute instructions according to the at least one instruction set. Figure 6 The method 300. For illustrative purposes only, this specification will describe the method 300 using server 130 as an example. The method 300 is applied to the display of badges in a mini-program within an application interface, and the method 300 may include:

[0061] S310: Obtain the description information of the target mini-program according to a preset cycle.

[0062] like Figure 2 , Figure 3As shown in this specification, various types of applications installed on client 110 (including smartphones, tablets, and other terminal devices) may have multiple display grids (represented by dashed boxes) on their display interfaces, with different display grids containing different function buttons. Furthermore, within a specific display grid of an application (the mini-program display grid), at least one mini-program that the user has favorited and / or recently used can be displayed. The mini-program display grid can show at least one mini-program, each with a corresponding mini-program icon. In this specification, server 130 can match corresponding target text for at least one mini-program, and the number of characters in the matched target text is controlled to within 3 characters, so that client 110 can simultaneously display the mini-program icon and the corresponding target text in the mini-program display grid. In this specification, the target text includes badge text.

[0063] In this manual, the target mini-program can be understood as the object on which server 130 matches the badge text. Server 130 can periodically match the corresponding target text for the target mini-program according to a preset cycle, that is, server 130 can obtain the description information of the target mini-program according to the preset cycle. In this manual, the description information of the target mini-program can be actively submitted to server 130 by the merchant, or it can be obtained by server 130 based on the configuration information, running information, etc. of the target mini-program. In this manual, the preset cycle can be 1 hour, 1 day, 1 week, or other time periods, which will not be elaborated here.

[0064] In this specification, the description information of the target mini-program can be understood as the source material for the server 130 to generate the target copy. The server 130 can match the target mini-program with the corresponding target copy based on the description information of the target mini-program. Therefore, the server 130 can periodically obtain the description information of the target mini-program according to a predetermined cycle; and, in order to accurately match the target mini-program with the corresponding target copy, the description information should encompass the characteristics of the target mini-program as much as possible.

[0065] In some embodiments, the description information may include at least one of the following: the name of the target mini-program, the slogan of the target mini-program, or the functional description of the target mini-program.

[0066] In this manual, the slogan of the target mini-program can be understood as its advertising slogan or promotional slogan. The functional introduction of the target mini-program may include its business scope, brand introduction, etc. For example, taking the target mini-program as "KDD" as an example, its slogan could be "Life is so beautiful," and its functional introduction could be "A fast food and fried chicken chain enterprise, selling fried chicken, hamburgers, fries, rice bowls, egg tarts, soft drinks, and other high-calorie fast food...". In this manual, the name, slogan, and functional introduction of the target mini-program serve as the source material for the target copy generated by server 130. These materials are often quite long and generally cannot be directly used as the target copy. In this case, server 130 can continue to extract key information from the numerous materials of the target mini-program to filter out unnecessary information and refine the materials.

[0067] S320, determine at least one keyword based on the description information, the at least one keyword reflecting the feature information of the target mini-program.

[0068] In this specification, at least one keyword is derived from the description information of the target mini-program. Compared to the description information of the target mini-program, keywords can accurately summarize or reflect the characteristic information of the target mini-program with fewer characters. The number of characters and part of speech of the keywords extracted by server 130 based on the description information of the target mini-program may vary. In this specification, the number of characters of the keyword may be less than or equal to three characters. For example, the number of characters of the keyword extracted by server 130 based on the description information of the target mini-program may be 1 character, 2 characters, or 3 characters. It should be understood that the keywords extracted by server 130 based on the description information of the target mini-program may include Chinese characters, as well as English letters, Arabic numerals, or other symbolic languages. When the keywords contain English letters, Arabic numerals, or other symbolic languages, each English letter, each Arabic numeral, or each other symbol represents one character. Taking the name of the target mini-program as "KDD" as an example, the keywords extracted by server 130 from the description information may include "chicken", "eat fried chicken", "eat hamburger", "chicken cutlet", "fast food", "KDD", etc.

[0069] In this specification, for short descriptive information within the target mini-program, server 130 can directly use them as keywords. For each target mini-program, server 130 can extract one or more keywords from its descriptive information.

[0070] In this specification, the part of speech of the keywords may include at least one of the following: verb, noun, or compound phrase. For example, the keywords "eat fried chicken" and "eat hamburger" are compound phrases consisting of a verb and a noun; the keywords "chicken," "chicken cutlet," and "fast food" are nouns; and the keyword "KDD" is a noun. In this specification, keywords with different word counts or parts of speech can serve as the basis for the target text generated by server 130.

[0071] S330, determine the target text of the target mini-program based on the at least one keyword.

[0072] In this specification, server 130 can generate target text corresponding to the target mini-program based on at least one keyword. The target text should be no more than three characters long and can use a verb-object structure (verb + noun combination). By keeping the generated target text to three characters or less and using a verb-object structure (verb + noun combination) as much as possible, the functionality of the mini-program can be better presented, attracting user clicks and increasing the frequency of mini-program usage.

[0073] As mentioned earlier, for each target mini-program, server 130 can extract one or more keywords from the description information of the target mini-program. In this specification, server 130 can determine one candidate copy based on each keyword; correspondingly, server 130 can determine multiple candidate copy based on multiple keywords. Server 130 then selects one candidate copy as the target copy from the multiple candidate copy.

[0074] In this specification, during the process of server 130 determining the target text of the target mini-program based on the at least one keyword, since the part of speech of each keyword may be different, or the number of characters of each keyword may also be different, server 130 needs to match the corresponding alternative text for the target mini-program according to the characteristics of each keyword (part of speech and / or number of characters) to determine the target text corresponding to the target mini-program.

[0075] Therefore, the copywriting generation method provided in this specification periodically acquires the description information of the target mini-program, extracts keywords based on the description information, and automatically generates target copywriting that matches the target mini-program based on the extracted keywords. The generated target copywriting can effectively present the mini-program's functions to attract users to click and use it. This copywriting generation method not only enables large-scale and rapid matching of target copywriting for mini-programs, improving the efficiency of mini-program badge copywriting generation, but also allows for periodic updates of the target files of the target mini-program, thus making ample preparations for the deployment of target copywriting in different operational cycles.

[0076] Figure 7A schematic diagram illustrating the generation process of target text according to some embodiments of this specification is shown. For example... Figure 7 As shown, in some embodiments, S330 may include:

[0077] For each of the at least one keywords: determine the part-of-speech tag and character count of the keyword; determine that the keyword meets preset conditions based on the part-of-speech tag and character count; and use the keyword as a first candidate text for the target mini-program, and determine a first confidence score for the first candidate text, the first confidence score reflecting the credibility of the first candidate text; and

[0078] The target text of the target mini-program is determined based on at least one of the first alternative texts and at least one of the first confidence scores.

[0079] In this specification, when server 130 determines a candidate text based on each keyword, it can simultaneously generate a confidence score for the candidate text. The confidence score of the candidate text reflects its credibility. The higher the confidence score of the candidate text, the higher the degree of matching between the candidate text and the target mini-program. Accordingly, when client 110 displays a candidate text with a higher confidence score, it is more likely to attract user clicks, thereby further increasing the frequency of mini-program usage.

[0080] As mentioned earlier, for each target mini-program, server 130 can extract one or more keywords from the mini-program's description information. In this specification, server 130 can determine a candidate copy and its confidence score based on each keyword. Correspondingly, server 130 can determine multiple candidate copies and their respective confidence scores based on multiple keywords. Server 130 then selects one candidate copy from these multiple candidate copies as the target copy. The method and system provided in this specification generate confidence scores for the candidate copies simultaneously. These confidence scores reflect the credibility of the candidate copies, greatly facilitating the management of candidate copies by operations personnel.

[0081] In this specification, server 130 can first identify each keyword to determine whether its part of speech and word count meet preset conditions. The method by which server 130 generates alternative text when the part of speech and word count of the keyword meet the preset conditions differs from the method by which server 130 generates alternative text when the part of speech and word count of the keyword do not meet the preset conditions. For ease of description, the alternative text generated by server 130 when the part of speech and word count of the keyword meet the preset conditions will be referred to as the first alternative text, and the confidence score of the alternative text generated by server 130 when the part of speech and word count of the keyword meet the preset conditions will be referred to as the first confidence score. The alternative text generated by server 130 when the part of speech and word count of the keyword do not meet the preset conditions will be referred to as the second alternative text, and the confidence score of the alternative text generated by server 130 when the part of speech and word count of the keyword do not meet the preset conditions will be referred to as the second confidence score.

[0082] In this specification, the preset conditions may include: the keyword is a verb and has two characters; or the keyword has three characters. That is, the keyword meeting the preset conditions may include either of the following two situations: (1) the keyword is a verb and has two characters; (2) the keyword has three characters. For example, taking the name of the target mini-program as "KDD", as mentioned above, the keywords extracted by the server 130 from the description information may include "eat fried chicken", "eat hamburger", "chicken cutlet", "fast food", "KDD", etc. Among them, the keywords "eat fried chicken", "eat hamburger", and "KDD" all have three characters, which meets the preset conditions. Although the keywords "chicken cutlet" and "fast food" have two characters, they are both nouns, so the keywords "chicken cutlet" and "fast food" do not meet the preset conditions. In addition, the keyword "chicken" has only one character, which also does not meet the preset conditions.

[0083] In this specification, when server 130 determines that the part-of-speech and character count of a keyword meet preset conditions, server 130 can directly use the keyword as the first candidate text for the target mini-program. At this time, server 130 does not need to process the keyword, indicating that the keyword can well reflect the characteristics of the target mini-program and that the keyword itself is concise and refined. Furthermore, server 130 can directly determine the first confidence score of the first candidate text as the highest confidence score. For example, if the confidence score is represented by 0-1, where 0 represents the lowest confidence score and 1 represents the highest confidence score, then server 130 can determine the first confidence score of the first candidate text as 1.

[0084] In this specification, when a target mini-program has multiple keywords that meet preset conditions, server 130 can determine multiple first candidate texts, and the first confidence score of each of the multiple first candidate texts is 1. In this case, server 130 can select one first candidate text from the multiple first candidate texts as the target text for the target mini-program. For example, as mentioned above, the keywords "eat fried chicken," "eat hamburger," and "KDD" all have three characters, meeting the preset conditions. In this case, server 130 can use the keyword "eat fried chicken" as a first candidate text, and server 130 can determine that the first confidence score of the first candidate text "eat fried chicken" is 1; server 130 can also use the keyword "eat hamburger" as a first candidate text, and server 130 can determine that the first confidence score of the first candidate text "eat hamburger" is 1; server 130 can also use the keyword "KDD" as a first candidate text, and server 130 can determine that the first confidence score of the first candidate text "KDD" is 1. That is, server 130 can identify three first candidate texts, namely "eat fried chicken", "eat hamburger" and "KDD", and determine that the first confidence score of each of the above three first candidate texts is the same, which is 1.

[0085] In some embodiments, determining the target text of the target mini-program based on at least one first alternative text and at least one first confidence score may include: randomly selecting a first alternative text from the at least one first alternative text as the target text, and using the first confidence score of the first alternative text as the target confidence score of the target text.

[0086] In this specification, if a keyword meets preset conditions, server 130 can directly use the keyword as the first candidate text. Even if the target mini-program has multiple keywords that meet the preset conditions, server 130 can determine multiple first candidate texts, and each of the multiple first candidate texts has the same first confidence score. In this case, server 130 can randomly select one first candidate text from the multiple first candidate texts as the target text for the target mini-program.

[0087] In this specification, server 130 can also generate a target confidence score for the target text while generating the target text. The target confidence score reflects the credibility of the target text. The higher the target confidence score of the target text, the higher the degree of matching between the target text and the target program. Accordingly, when client 110 uses target text with a higher confidence score for display, it is more likely to attract user clicks, thereby further increasing the frequency of use of the mini-program.

[0088] As mentioned earlier, for each target mini-program, server 130 can extract one or more keywords from the description information of the target mini-program. In this specification, server 130 can determine a candidate copy and its confidence score based on each keyword. Correspondingly, server 130 can determine multiple candidate copies and their confidence scores based on multiple keywords. Server 130 then selects one candidate copy from the multiple candidate copies as the target copy and uses the confidence score of the selected candidate copy as the target confidence score of the target copy.

[0089] If a keyword meets the preset conditions, server 130 can directly use the keyword as the first candidate text and determine the first confidence score of the first candidate text as the highest confidence score. Even if the target mini-program has multiple keywords that meet the preset conditions, server 130 can determine multiple first candidate texts, and each of the multiple first candidate texts has the same first confidence score. In this case, server 130 can randomly select one first candidate text from the multiple first candidate texts as the target text of the target mini-program, and use the first confidence score of the selected first candidate text as the target confidence score of the target text (i.e., determine the target text's target confidence score as the highest confidence score). For example, server 130 can randomly select "eat fried chicken" as the target text from three first candidate texts (including "eat fried chicken", "eat hamburger", and "KDD"). Since the first confidence score of "eat fried chicken" is 1, the target confidence score corresponding to the target text "eat fried chicken" is 1. Server 130 can also randomly select "eat hamburger" as the target text from three first candidate texts (including "eat fried chicken", "eat hamburger", and "KDD"). Since the first confidence score of "eat hamburger" is 1, the target confidence score corresponding to the target text "eat hamburger" is 1. Server 130 can also randomly select "eat fried chicken" as the target text from three first candidate texts (including "eat fried chicken", "eat hamburger", and "KDD"). Since the first confidence score of "KDD" is 1, the target confidence score corresponding to the target text "KDD" is 1.

[0090] In some embodiments, S330 may include: for each of the at least one keyword: determining the part-of-speech and number of characters of the keyword, determining that the keyword does not meet the preset condition based on the part-of-speech and the number of characters, and performing text matching on the keyword based on a text generation model to determine a second alternative copy for the target mini-program, and determining a second confidence score for the second alternative copy, the second confidence score reflecting the credibility of the second alternative copy; and determining the target copy for the target mini-program based on at least one second alternative copy and at least one second confidence score.

[0091] In this specification, when server 130 determines that the part-of-speech and character count of a keyword do not meet preset conditions, the keyword cannot be directly used as the first alternative copy for the target mini-program. In this case, server 130 needs to process the keyword. For each keyword, server 130 can generate a second alternative copy that matches the target mini-program and generate a second confidence score for the second alternative copy. For example, in the example where the target mini-program is named "KDD", the keywords "chicken cutlet", "fast food", and "chicken" do not meet the preset conditions. In this case, server 130 can perform text matching on the keywords "chicken cutlet", "fast food", and "chicken" respectively to determine three second alternative copies and determine the content of each second alternative copy. That is, server 130 can perform text matching on the keyword "chicken cutlet" to determine a second alternative copy and a second confidence score for the second alternative copy; server 130 can perform text matching on the keyword "fast food" to determine another second alternative copy and a second confidence score for the second alternative copy; server 130 can also perform text matching on the keyword "chicken" to determine yet another second alternative copy and a second confidence score for the second alternative copy.

[0092] In this specification, server 130 can perform text matching on keywords based on a text generation model to determine the second alternative copy and the second confidence score of the second alternative copy for the target mini-program. The text generation model in this disclosure can be a Transformer model, which uses a multi-head attention mechanism to capture the relationship between multiple subspaces in word vectors, thereby performing context prediction.

[0093] Figure 8 A schematic diagram illustrating the working principle of the training phase of a text generation model according to some embodiments of this specification is shown. Figure 8As shown, server 130 can take the text "take pictures of plants" as input to the text generation model. During or before input, the text has been segmented according to part-of-speech tags. The segmented text, arranged in sequence, includes "will," "plant," and "take pictures." Server 130 can perform one-hot encoding on the segmented text sequence. Server 130 can use the text generation model to perform word embeddings based on the vocabulary content and position of the input text sequence, mapping the text sequence to a high-dimensional space. For example, server 130 can use the word2vec network in the text generation model to perform word embedding on the text sequence, corresponding to embedding-1. Simultaneously, server 130 can use the text generation model to encode the position of words in the text sequence, corresponding to embedding-2. Then, server 130 can use the attention mechanism network, pooling network, and normalization network in the text generation model to predict the words following "take pictures" and determine the probability value of the predicted words. The higher the probability value of the predicted words, the higher the accuracy of the predicted words. For example, given the input sequence text "take pictures of plants", the server 130 uses a text generation model to predict words such as "upload", "take pictures", "know", "have", "use", "rent", "buy", and "eat". The sum of the probability values ​​of these predicted words is 1, and the probability values ​​of the predicted words decrease sequentially. That is, the predicted word "upload" has the highest probability value, indicating that the word most closely related to the input sequence text "take pictures of plants" is "upload".

[0094] It is important to note that during the training phase of the text generation model, the server 130 can compare the predicted words with the preset words each time, and continuously adjust the network parameters of the text generation model based on the comparison results to improve the prediction accuracy of the text generation model.

[0095] In this specification, server 130 can use keywords that do not meet preset conditions as input to a text generation model, generate predicted words that best match the keywords using the text generation model, combine the keywords and predicted words to obtain a second candidate text, and use the probability value of the predicted words as the second confidence score of the second candidate text. For example, Figure 8 If the probability value of the word "upload" obtained by predicting the input text is the highest, then the server 130 can combine "upload" with the input text. If the probability value of the predicted word "upload" is 89%, then the confidence score of the combined text is 89%.

[0096] It should be understood that Figure 8The text input to the text generation model, the predicted words output, and the probability values ​​of the predicted words output are only for illustrating the working principle of the text generation model and do not restrict the input or output of the actual text generation model.

[0097] In some embodiments, the step of performing text matching on the keywords based on a text generation model to determine a second candidate text for the target mini-program, and determining a second confidence score for the second candidate text, may include: predicting verbs that match the keywords and the probability values ​​of the verbs based on the text generation model, wherein the verbs are one character; and combining the verbs and the keywords to form the second candidate text, and using the probability values ​​of the verbs as the second confidence score of the second candidate text.

[0098] In this specification, considering that the word count requirement for the superscript text is no more than three characters, the server 130 can control the word count of the generated second alternative text by setting the word count of the predicted words in the text generation model. For example, the server 130 can set the word count of each predicted word in the text generation model to one character, and correspondingly, the word count of the second alternative text formed by the combination of predicted words and keywords will not exceed three characters. In this specification, the server 130 can further control the structure of the generated second alternative text by setting the part of speech of the predicted words in the text generation model. For example, the server 130 can set the part of speech of each predicted word in the text generation model to be a verb, and correspondingly, the second alternative text formed by the combination of predicted words and keywords will be a verb-object structure (verb + noun combination).

[0099] In the example where the target mini-program is named "KDD", for the keyword "chicken cutlet", server 130 can match a single-word verb for the keyword "chicken cutlet" based on a text generation model. Specifically, server 130 can input the keyword "chicken cutlet" into the text generation model, which can then obtain at least one predicted word. For example, the predicted word can include single-word verbs such as "eat", "show off", and "hit", among which "eat" has the highest probability value, with a probability value of 77%. At this time, server 130 can combine "eat" with the keyword "chicken cutlet" to obtain "eat chicken cutlet", and use "eat chicken cutlet" as the second candidate text. Correspondingly, the second confidence score of the second candidate text is 93%.

[0100] Since keywords that do not meet the preset conditions also include "fast food" and "chicken," server 130 can input the keyword "fast food" into the text generation model. The text generation model then determines the single-word verb that best matches the keyword "fast food," thereby determining a second alternative copy and its second confidence score. For example, the second alternative copy generated by server 130 based on text matching of the keyword "fast food" using the text generation model could be "buy fast food," with a corresponding second confidence score of 60%.

[0101] Server 130 can also input the keyword "chicken" into the text generation model, and use the text generation model to determine the single-word verb that best matches the keyword "chicken," thereby determining a second alternative copy and a second confidence score for the second alternative copy. For example, the second alternative copy generated by server 130 based on text matching of the keyword "chicken" using the text generation model could be "eat chicken," with a corresponding second confidence score of 80%.

[0102] In this specification, after the server 130 determines a second alternative copy for each keyword that does not meet the preset conditions and determines the second confidence score of each second alternative copy, the server 130 can determine one of the second alternative copies as the target copy based on the second confidence score of each second alternative copy.

[0103] It should be noted that in this specification, when the keyword is a single-character verb, the server 130 inputs the keyword into the text generation model. Since the text generation model can recognize the semantic information of the input keyword, and it is used to predict verbs that match the keyword, the contextual relationship between the two verbs differs from that of a noun + verb. In this case, the probability values ​​of each verb predicted by the server are not significantly different. Consequently, the absolute probability value of the predicted word with the highest probability value is relatively small, resulting in a lower confidence score for the generated second alternative text.

[0104] In this specification, determining the target text of the target mini-program based on at least one second alternative text and at least one second confidence score may include: selecting the second alternative text with the highest second confidence score from the at least one second alternative text as the target text, and using the second confidence score of the second alternative text as the target confidence score of the target text.

[0105] In this specification, server 130 can determine the target text by comparing the second confidence scores of multiple second candidate texts. Server 130 can select the text with the highest second confidence score from the multiple second candidate texts as the target text, and use the second confidence score corresponding to the selected second candidate text as the target confidence score of the target text. For example, as mentioned earlier, in the example where the target mini-program is named "KDD", the second confidence score corresponding to the second candidate text "eat chicken cutlet" is 93%, the second confidence score corresponding to the second candidate text "buy fast food" is 60%, and the second confidence score corresponding to the second candidate text "eat chicken" is 80%. In this case, server 130 can select the second candidate text "eat chicken cutlet" as the target text and determine the target confidence score of the target text to be 0.93 (equivalent to 93%).

[0106] In some embodiments, the first confidence score is higher than the second confidence score.

[0107] In this specification, server 130 directly uses keywords that meet preset conditions as the first candidate text for the target mini-program, and the first confidence score corresponding to the first candidate text is 1. For keywords that do not meet the preset conditions, server 130 performs text matching based on a text generation model to determine the second candidate text, and determines the second confidence score corresponding to the second candidate text based on the probability value of the predicted word. Generally, the probability value of the predicted word obtained by server 130 based on the text generation model is less than 1, that is, the second confidence score is less than 1. Therefore, the first confidence score is higher than the second confidence score. Correspondingly, for the target mini-program, if at least one keyword extracted by server 130 based on the description information of the target mini-program contains a keyword that meets the preset conditions, then server 130 can directly use the keyword that meets the preset conditions as the target text of the target mini-program and determine the target confidence score of the target text as 1; if none of the keywords extracted by server 130 based on the description information of the target mini-program meet the preset conditions, then server 130 can generate the target text based on the text generation model and determine the target confidence score of the target text based on the probability value of the predicted words obtained by the text generation model.

[0108] S340, Output the target text.

[0109] In this manual, after the server 130 determines the target text of the target mini-program, it can directly output the target text to the client 110, so that the client 110 can use the target text as the badge text of the target mini-program for distribution, thereby distributing the corresponding target text in a timely manner during different operating cycles.

[0110] In this specification, after determining the target text for the target mini-program, server 130 can directly output the target text to operations server 150 for review and feedback of the first review result. After receiving the first review result from operations server 150, server 130 can output the target text based on the review result, thereby deploying the corresponding target text in a timely manner during different operational cycles.

[0111] In this specification, server 130 can also pre-set a confidence score threshold. After server 130 determines the target text and target confidence score of the target mini-program, it can compare the target confidence score with the confidence score threshold and output the target text based on the comparison result. When the target confidence score of the target text is greater than the confidence score threshold, it indicates that the generated target text has high credibility. At this time, server 130 can output the target text to client 110, so that client 110 can use the target text as the badge text of the target mini-program for distribution. When the target confidence score of the target text is less than the confidence score threshold, server 130 can output the target text and target confidence score to operations server 150 for review and feedback of the first review result. After receiving the first review result from operations server 150, server 130 can output the target text based on the review result.

[0112] Specifically, S340 may include: obtaining a first review result for the target text, wherein the first review result includes: the target text being in an available state or the target text being in an unavailable state; and outputting the target text based on the first review result. In this specification, the first review result fed back by the operation terminal 150 may include two situations: approved or not approved. Approved means the target text is in an available state, while not approved means the target text is in an unavailable state.

[0113] When the target copy fails the review by the operations side 150, the operations staff can provide feedback on the rejection through the operations side 150. The server 130 can then adjust the parameters of the text generation model based on the rejection feedback, and can also adjust the keyword extraction strategy based on the rejection feedback. Furthermore, when the target copy fails the review by the operations side 150, the operations staff can also configure a new target copy online through the operations side 150 and submit the new target copy to the server 130.

[0114] In response to the two different situations mentioned above, server 130 can output the target text in different ways.

[0115] In some embodiments, outputting the target text based on the first review result may include: determining that the first review result indicates the target text is available, using the target text as the badge text of the target mini-program, and outputting the badge text. In this specification, when the server 130 receives a first review result from the operator 150 indicating that the review has passed, the server 130 can output the target text to the client 110, so that the client 110 can use the target text as the badge text of the target mini-program for distribution.

[0116] In some embodiments, after determining that the first review result indicates the target document is available, the method may further include: saving the target document.

[0117] In this manual, after the server 130 determines that the first review result of the target text by the operation terminal 150 is in an available state, the server 130 can not only output the target text to the client 110 so that the client 110 can use the target text as the badge text of the target mini program, but also save the currently deployed target text.

[0118] Furthermore, once server 130 determines that the first review result of the target copy by the operation terminal 150 is in an available state, for the target copy with a determined target confidence score, server 130 can not only output the target copy to client 110 for delivery, but also save the target copy and its corresponding target confidence score. Since server 130 obtains the description information of the target mini-program according to a preset period, extracts keywords based on the description information, and generates the target copy and target confidence score of the target mini-program based on the keywords, server 130 continuously generates target copy and target confidence scores according to a preset period. At this time, server 130 can compare the confidence scores of the currently delivered target copy with those generated in the next preset period, and select the target copy with the highest confidence score to output to client 110 for delivery.

[0119] If the target confidence score of the currently delivered target copy is higher than the target confidence score of the target copy generated in the next preset period, then server 130 can output the currently delivered target copy to client 110 for the next round of delivery. If the target confidence score of the currently delivered target copy is lower than the target confidence score of the target copy generated in the next preset period, then server 130 can output the target copy generated in the next preset period to client 110 for the next round of delivery. Furthermore, server 130 can permanently store the currently delivered target copy and use the stored target copy for subsequent confidence score comparisons. Therefore, the method and system provided in this specification generate a target confidence score for the target copy while generating the target copy. The target confidence score reflects the credibility of the target copy, which greatly facilitates the management of target copy by operations personnel.

[0120] In some embodiments, outputting the target text based on the first review result may further include: determining that the first review result indicates the target text is unavailable, obtaining manually configured text, using the manually configured text as the badge text of the target mini-program, and outputting the badge text. In this specification, when server 130 receives a first review result from operator 150 indicating a failed review, server 130 may simultaneously receive a new target text from operator 150. In this specification, the new target text from operator 150 is the manually configured text. Server 130 may output the new target text (manually configured text) to client 110, so that client 110 may use the new target text as the badge text of the target mini-program.

[0121] In some embodiments, after obtaining the first review result of the target text, the method further includes: obtaining the feedback result of the target text, wherein the feedback result includes a negative feedback result corresponding to the target text being in an unavailable state.

[0122] In this manual, when server 130 receives the initial review result from operator 150 as "not approved," server 130 can simultaneously receive both the new target document from operator 150 and the feedback result from operator 150. If the target document does not meet the requirements, operator 150 can send negative feedback to server 130. Negative feedback may include reasons why the target document failed the review, such as grammatical issues, sensitive words, or other sensitive phrases.

[0123] In this specification, the period at which the server 130 obtains feedback results of the target text can be the same as or different from the preset period.

[0124] In some embodiments, after obtaining the feedback result of the target text, the method may further include: adjusting the parameters of the text generation model based on the negative feedback result to update the target text. In this specification, the server 130 may adjust the parameters of the text generation model based on factors such as grammatical factors, sensitive word factors, sensitive phrase factors, or other factors included in the negative feedback result, thereby updating the target text.

[0125] In this specification, the frequency at which the server 130 adjusts the parameters of the text generation model based on the feedback results can be consistent with or inconsistent with the preset period.

[0126] In some embodiments, the method 300 may further include: S350, obtaining a second review result for the at least one keyword; and updating the at least one keyword based on the second review result.

[0127] As mentioned earlier, server 130 can adjust its keyword extraction strategy based on the rejection reasons in the first review result. Furthermore, in this specification, server 130 can also directly review keywords to obtain a second review result for at least one keyword. Specifically, server 130 can output keywords to operations terminal 150, where operations personnel can review the keywords and provide feedback on the second review result. At this point, server 130 can directly adjust its keyword extraction strategy based on the second review result, thereby improving the accuracy of the keywords in reflecting the target program's characteristic information, and consequently enhancing the reliability of the target copy.

[0128] The copy generation method and system provided in this manual periodically acquire the description information of target mini-programs, extract keywords based on the description information, and automatically generate target copy matching the target mini-program based on the extracted keywords. The generated target copy can effectively present the functions of the mini-program to attract users to click and use it. The method and system can not only achieve large-scale and rapid matching of target copy for mini-programs, improving the efficiency of mini-program badge copy generation, but also periodically update the target files of the target mini-program, and timely release corresponding target copy in different operating cycles.

[0129] In summary, after reading this detailed disclosure, those skilled in the art will understand that the foregoing detailed disclosure is presented by way of example only and is not restrictive. Although not explicitly stated herein, those skilled in the art will understand that this specification is intended to encompass various reasonable changes, improvements, and modifications to the embodiments. These changes, improvements, and modifications are intended to be made by this specification and are within the spirit and scope of the exemplary embodiments described herein.

[0130] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0131] Furthermore, certain terms in this specification have been used to describe embodiments of this specification. For example, "an embodiment," "an embodiment," and / or "some embodiments" mean that a particular feature, structure, or characteristic described in connection with that embodiment may be included in at least one embodiment of this specification. Therefore, it is to be emphasized and understood that two or more references to "an embodiment" or "an embodiment" or "alternative embodiment" in various parts of this specification do not necessarily refer to the same embodiment. Moreover, specific features, structures, or characteristics may be suitably combined in one or more embodiments of this specification.

[0132] It should be understood that in the foregoing description of the embodiments of this specification, for the purpose of simplifying the description and to aid in understanding a feature, various features are sometimes combined in a single embodiment, drawing, or description thereof. Alternatively, various features may be distributed across multiple embodiments of this specification. However, this does not mean that the combination of these features is necessary, and those skilled in the art, upon reading this specification, may extract some features as individual embodiments for understanding. That is, the embodiments in this specification can also be understood as an integration of multiple sub-embodiments. It is also valid when each sub-embodiment contains fewer than all the features of a single foregoing disclosed embodiment.

Claims

1. A text generation method, applied to the placement of badges in mini-programs within an application interface, comprising: Obtain the description information of the target mini-program according to a preset cycle; At least one keyword is determined based on the description information, and the at least one keyword reflects the feature information of the target mini-program; For each of the at least one keywords: determine the part-of-speech and number of characters of the keyword, and based on the relationship between the part-of-speech and the number of characters and preset conditions, determine the candidate text of the target mini-program and the confidence score of the candidate text, wherein, if the part-of-speech and the number of characters satisfy the preset conditions, the keyword is used as the candidate text of the target mini-program, or if the part-of-speech and the number of characters do not satisfy the preset conditions, text matching of the keyword is performed based on a text generation model to determine the candidate text of the target mini-program, and the confidence score is used to characterize the credibility of the candidate text; Based on at least one of the candidate texts and at least one of the confidence scores, the target text for the target mini-program is determined; and Output the target text.

2. The copywriting generation method as described in claim 1, wherein, The description information includes: The target mini-program's name, its slogan, or its feature description.

3. The copywriting generation method as described in claim 1, wherein, The number of characters in the keyword is less than or equal to three; as well as The part of speech of the keywords includes at least one of the following: verb, noun, or compound phrase.

4. The copywriting generation method as described in claim 1, wherein, The determination of the candidate text and the confidence score of the candidate text for the target mini-program based on the relationship between the part-of-speech tag and the number of characters and preset conditions includes: For each of the at least one keywords: based on the part of speech and the number of characters, determine that the keyword meets the preset condition, and use the keyword as the first candidate text of the target mini-program, and determine the first confidence score of the first candidate text, wherein the first confidence score reflects the credibility of the first candidate text; Determining the target copy of the target mini-program based on at least one of the candidate copy and at least one of the confidence scores includes: determining the target copy of the target mini-program based on at least one of the first candidate copy and at least one of the first confidence scores.

5. The copywriting generation method as described in claim 1, wherein, The determination of the candidate text and the confidence score of the candidate text for the target mini-program based on the relationship between the part-of-speech tag and the number of characters and preset conditions includes: For each of the at least one keywords: based on the part of speech and the number of characters, determine that the keyword does not meet the preset condition, and perform text matching on the keyword based on the text generation model to determine the second alternative copy of the target mini-program, and determine the second confidence score of the second alternative copy, wherein the second confidence score reflects the credibility of the second alternative copy; Determining the target copy of the target mini-program based on at least one of the candidate copy and at least one of the confidence scores includes: determining the target copy of the target mini-program based on at least one second candidate copy and at least one second confidence score.

6. The copywriting generation method as described in claim 5, wherein, The step of performing text matching on the keywords based on a text generation model to determine the second candidate text for the target mini-program, and determining the second confidence score of the second candidate text, includes: Based on the text generation model, the model predicts verbs that match the keywords and the probability values ​​of those verbs, wherein each verb consists of one character; and The verb and the keyword are combined to form the second candidate copy, and the probability value of the verb is used as the second confidence score of the second candidate copy.

7. The copywriting generation method as described in claim 4 or 5, wherein, The preset conditions include: The keyword must be a verb and must have at least two characters; or The keyword must be three characters long.

8. The copywriting generation method as described in claim 4, wherein, The step of determining the target text for the target mini-program based on at least one first candidate text and at least one first confidence score includes: Randomly select one first candidate text from the at least one first candidate text as the target text, and use the first confidence score of the first candidate text as the target confidence score of the target text.

9. The copywriting generation method as described in claim 5, wherein, The step of determining the target text for the target mini-program based on at least one second alternative text and at least one second confidence score includes: The second candidate text with the highest second confidence score is selected from the at least one second candidate text as the target text, and the second confidence score of the second candidate text is used as the target confidence score of the target text.

10. The copywriting generation method as described in claim 1, wherein, The output of the target text includes: The first review result of the target text is obtained, wherein the first review result includes: the target text is in an available state or the target text is in an unavailable state; and The target text is output based on the first review result.

11. The copywriting generation method as described in claim 10, wherein, The step of outputting the target text based on the first review result includes: If the first review result indicates that the target text is available, the target text is used as the badge text of the target mini-program, and the badge text is output.

12. The copywriting generation method as described in claim 11, wherein, After determining that the first review result indicates the target document is available, the method further includes: Save the target text.

13. The copywriting generation method as described in claim 10, wherein, The step of outputting the target text based on the first review result includes: If the first review result determines that the target text is unavailable, obtain the manually configured text, use the manually configured text as the badge text of the target mini-program, and output the badge text.

14. The copywriting generation method as described in claim 13, wherein, After obtaining the first review result of the target document, the process also includes: Obtain feedback results for the target text, including negative feedback results corresponding to the target text being in an unavailable state.

15. The copywriting generation method as described in claim 14, wherein, After obtaining the feedback result of the target copy, the process also includes: The parameters of the text generation model are adjusted based on the negative feedback results to update the target text.

16. The copy generation method as described in claim 1, wherein, Also includes: Obtain a second review result for at least one of the keywords; as well as The at least one keyword is updated based on the second review result.

17. A copywriting generation system, comprising: At least one storage medium, including at least one instruction set, for implementation analysis of the copy generation method; as well as At least one processor is communicatively connected to the at least one storage medium. When the system is running, the at least one processor reads the at least one instruction set and executes the method of any one of claims 1-16 according to the instructions of the at least one instruction set.

Images