Cboth recommendation method, system and device and storage medium

Abstract

The invention provides a copywriting recommendation method, system and device and a storage medium, and the method comprises the steps: responding to a received copywriting generation request input by a user, and obtaining topic data related to a target topic from a plurality of data sources according to the target topic matched with the copywriting generation request; generating target prompt information according to a preset prompt information generation mode; respectively inputting the target prompt information and the topic data into a plurality of copywriting generation models to obtain candidate copywriting respectively output by the plurality of copywriting generation models; and determining the candidate copywriting with the highest matching degree with the user and the target topic from the plurality of candidate copywriting as a target copywriting, and recommending the target copywriting to the user. Therefore, on the basis of generating sufficient high-quality candidate copywriting, the target copywriting with the optimal matching degree with the candidate copywriting can be finely distributed to the user, so that the technical defects that copywriting creation lacks diversity and copywriting distribution is not fine enough in the prior art are effectively overcome.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and more specifically, to a copywriting recommendation method, system, device, and storage medium. Background Technology

[0002] In social applications, after adding strangers as friends, users often need to send them an opening message. In current technology, users usually search for relevant templates online or generate relevant opening messages as reference text through AI (Artificial Intelligence) generation models.

[0003] However, the aforementioned existing technologies, on the one hand, lack the perception of external information and create copy based on existing prior knowledge. Therefore, the existing technologies have the defects of generating similar reference copy and lacking diversity in copy creation. On the other hand, after generating reference copy through AI generation models, the lack of refined distribution makes it difficult for AI generation models to distribute different reference copy to different users in a refined manner, even if they generate a sufficient number of reference copy. Summary of the Invention

[0004] In view of this, this application provides a copywriting recommendation method, system, device and storage medium, which can generate a sufficient amount of high-quality candidate copywriting and then distribute target copywriting with the best matching degree to users in a refined manner, thereby effectively overcoming the technical defects of existing technologies such as lack of diversity in copywriting creation and insufficient refinement in copywriting distribution.

[0005] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings.

[0006] In a first aspect, embodiments of this application provide a copywriting recommendation method, which is applied to a copywriting recommendation system corresponding to a target application. The copywriting recommendation method includes: In response to receiving a text generation request from a user, topic data related to the target topic is obtained from multiple data sources based on the target topic that matches the text generation request. According to a preset prompt information generation method, generate target prompt information to instruct the generation of text related to the target topic; The target prompt information and the topic data are respectively input into multiple copywriting generation models, and the copywriting output by the multiple copywriting generation models is used as the candidate copywriting corresponding to the target topic. From the multiple candidate texts, the candidate text with the highest degree of matching between the user and the target topic is determined as the target text, and the target text is recommended to the user.

[0007] Secondly, embodiments of this application provide a copywriting recommendation system, which corresponds to a target application, wherein the copywriting recommendation system includes: A data source agent is used to respond to a text generation request input by a user, and to obtain topic data related to the target topic from multiple data sources based on the target topic that matches the text generation request. A production intelligent agent is used to generate target prompt information according to a preset prompt information generation method, which is used to instruct the generation of text related to the target topic; The target prompt information and the topic data are respectively input into multiple copywriting generation models, and the copywriting output by the multiple copywriting generation models is used as the candidate copywriting corresponding to the target topic. A distribution agent is used to determine the candidate text with the highest matching degree between the candidate text and the user and the target topic from a plurality of candidate texts, and recommend the target text to the user.

[0008] Thirdly, embodiments of this application provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described copywriting recommendation method.

[0009] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program, which, when executed by a processor, performs the steps of the above-described copywriting recommendation method.

[0010] The technical solutions provided by the embodiments of this application may include the following beneficial effects: This application provides a copywriting recommendation method, system, device, and storage medium. In response to receiving a copywriting generation request from a user, it acquires topic data related to the target topic from multiple data sources based on the target topic matching the request; generates target prompt information to instruct the generation of copywriting related to the target topic according to a preset prompt information generation method; inputs the target prompt information and topic data into multiple copywriting generation models respectively, obtaining the copywriting output by each model as candidate copywriting corresponding to the target topic; and determines the candidate copywriting with the highest matching degree between the user and the target topic from among the multiple candidate copywritings as the target copywriting, and recommends the target copywriting to the user. Thus, this application can, based on generating a sufficient number of high-quality candidate copywritings, finely distribute the target copywriting with the best matching degree to the user, thereby effectively overcoming the technical defects of existing technologies such as lack of diversity in copywriting creation and insufficient fineness in copywriting distribution. Attached Figure Description

[0011] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 This paper shows a schematic diagram of the structure of a copywriting recommendation system provided in an embodiment of this application; Figure 2 A flowchart illustrating a copywriting recommendation method provided in an embodiment of this application is shown; Figure 3 This is a schematic diagram of the structure of an electronic device 300 provided in an embodiment of this application. Detailed Implementation

[0013] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the accompanying drawings in this application are for illustrative and descriptive purposes only and are not intended to limit the scope of protection of this application. Furthermore, it should be understood that the schematic drawings are not drawn to scale. The flowcharts used in this application illustrate operations implemented according to some embodiments of this application. It should be understood that the operations in the flowcharts may not be implemented in sequence, and steps without logical contextual relationships may be reversed or implemented simultaneously. In addition, those skilled in the art, guided by the content of this application, may add one or more other operations to the flowcharts, or remove one or more operations from the flowcharts.

[0014] Furthermore, the described embodiments are merely some, not all, of the embodiments of this application. The components of the embodiments of this application described and illustrated herein can typically be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0015] It should be noted that the term "comprising" will be used in the embodiments of this application to indicate the presence of the features declared thereafter, but does not exclude the addition of other features.

[0016] In existing technologies, users typically search for relevant templates online or use AI-generated models to generate relevant opening remarks as reference text. However, these existing technologies suffer from several drawbacks. First, due to a lack of awareness of external information, the text is created based on existing prior knowledge, resulting in similar reference texts and a lack of diversity in text creation. Second, after generating reference texts through AI-generated models, the lack of refined distribution makes it difficult for AI-generated models to distribute different reference texts to different users in a precise manner, even if they generate a sufficient number of reference texts.

[0017] Based on this, the embodiments of this application provide a copywriting recommendation method, system, device and storage medium, which can generate a sufficient number of high-quality candidate copywritings and then distribute target copywritings with the best matching degree to users in a refined manner, thereby effectively overcoming the technical defects of the prior art, such as the lack of diversity in copywriting creation and the insufficient refinement of copywriting distribution.

[0018] To facilitate understanding of the embodiments of this application, a detailed description of a document recommendation method, system, device, and storage medium provided in the embodiments of this application is provided below.

[0019] Here, the text recommendation method provided in this application embodiment is applied to the text recommendation system corresponding to the target application; wherein, the target application can be a social application or other applications involving the sending and receiving of text information, and the text recommendation system is composed of multiple agent intelligent agents with different functional divisions. Each agent intelligent agent represents a system that can make decisions and take actions in combination with environmental changes. It can be an AI model, an application, or a physical device. This application embodiment does not limit it in any way.

[0020] like Figure 1 As shown, Figure 1The diagram illustrates the structure of a copywriting recommendation system provided in an embodiment of this application. The copywriting recommendation system includes: a Selector Agent, a Data Source Agent, a Production Agent, a Review Agent, an Online Agent, a Distribution Agent, and a Summary Agent.

[0021] Specifically, in the copywriting recommendation system, the selector agent, as the leader of the system, can receive copywriting generation requests sent by users through the client of the target application. It then recommends copywriting matching the request as inference targets. Based on the functions of multiple other agents (i.e., the aforementioned data source agent, production agent, review agent, online agent, distribution agent, and summary agent), it infers the tasks to be assigned to each agent, determines the tasks to be assigned to each agent, and sends the matching tasks to each agent. Thus, in the copywriting recommendation system, all agents can interact, collaborate, or compete with each other to jointly complete complex tasks that are difficult for a single agent to handle.

[0022] Specifically, the data source agent is used to perceive the constantly changing information data in the world (i.e., inside and outside the target application). Through the data source agent, the reference data required to complete the copy generation request can be obtained from multiple data sources corresponding to the target application inside and outside (which is also equivalent to the reference data required to generate copy that matches the copy generation request).

[0023] Specifically, the production agent is used to produce multiple candidate texts that match the copywriting generation request through diversified production methods (that is, to produce a sufficient amount of high-quality texts).

[0024] Specifically, the review agent is used to conduct quality reviews and optimizations on multiple candidate documents produced by the production agent, in order to determine the target document that best matches the user.

[0025] Specifically, the online intelligent agent is used to write the optimal copy (i.e., the target copy) determined by the review intelligent agent into the offline / online storage engine of the target application.

[0026] Specifically, the distribution agent is used to distribute the target copy that the review agent has determined to be most suitable for each user to different users in a refined manner.

[0027] Specifically, this study summarizes how intelligent agents listen to the latest messages from other intelligent agents and summarizes the task execution process of all intelligent agents in the copywriting recommendation system.

[0028] In summary, in a copywriting recommendation system, different agents can perform their respective functions. By having the selector agent autonomously formulate a complete task execution process based on the copywriting generation request currently input by the user, the copywriting recommendation system can recommend the most suitable copywriting to each user after receiving copywriting generation requests from different users. This effectively overcomes the technical shortcomings of existing technologies, such as the lack of diversity in copywriting creation and the insufficient precision of copywriting distribution. In addition, since the production agent can produce a sufficient quantity of high-quality copywriting, the copywriting recommendation system also helps to improve the efficiency of copywriting production and the quantity of high-quality copywriting produced.

[0029] The following is combined with Figure 1 The document recommendation system shown herein provides a detailed description of the document recommendation method provided in the embodiments of this application.

[0030] here, Figure 2 This illustration shows a flowchart of a copywriting recommendation method provided in an embodiment of this application, wherein, as shown... Figure 2 As shown, the copywriting recommendation method includes steps S201-S204, specifically: S201, in response to receiving a copy generation request input by the user, topic data related to the target topic is obtained from multiple data sources based on the target topic that matches the copy generation request.

[0031] Here, in the copywriting recommendation system, the Selector Agent, in response to receiving a copywriting generation request from a user, can determine the target topic that matches the copywriting generation request and assign a task to the data source agent to instruct it to obtain topic data related to the target topic.

[0032] Specifically, after receiving the aforementioned task to be executed, the data source agent can obtain topic data related to the target topic from multiple data sources according to the methods shown in steps a1-a3 below: Step a1: Obtain target internal data related to the target topic from the target application's internal data.

[0033] Here, the copywriting recommendation system is the internal system of the target application. Therefore, the data source agent can directly access the target application's internal data (i.e., internal data) and obtain target internal data related to the aforementioned target topic from the internal data.

[0034] Step a2: Obtain target external data related to the target topic from external data sources.

[0035] Here, external data source refers to other data sources located outside the target application; external data sources may include, but are not limited to, external applications, external data platforms, etc.

[0036] Specifically, in the copywriting recommendation system, an MCP (Model Context Protocol) server can be pre-deployed for the data source agent; the MCP server is used to achieve seamless integration of large language model applications (such as the multiple agents involved in this application) with external data sources and tools, so that the data source agent can access the aforementioned external data sources through its corresponding MCP server.

[0037] It should be noted that, in this embodiment of the application, in addition to the data source agent, an MCP server can be pre-deployed for each other agent. In terms of underlying technical implementation, the deployment of the MCP server can be achieved through FAST-MCP and FAST-API, so that each agent in the copywriting recommendation system can perform Function Calling through the MCP protocol (during program execution, by specifying the function name and passing in the necessary parameters, the operation of the code block defined in the function is executed, and the AI ​​model and the application can achieve point-to-point calling), thereby improving the richness and production efficiency of the copywriting.

[0038] Step a3: Use the data within the target site and the data outside the target site as the topic data.

[0039] Here, the data source agent acquires target site data and target site data as topic data required for generating copy, which enables the copy recommendation system to have more prior knowledge and brings more inspiring thinking to the creation of copy. In addition, the topic data acquired by the data source agent comes from more data sources, which also means more creativity and possibilities. This helps to overcome the technical deficiency of copy creation lacking diversity in copy creation that exists in manually created copy or copy generated by a single AI model.

[0040] S202, generate target prompt information according to the preset prompt information generation method to indicate the generation of text related to the target topic.

[0041] Here, in the copywriting recommendation system, for the same target topic (i.e. the same copywriting generation request), the production agent can generate target prompts with the same prompt function (i.e., all used to instruct the generation of copywriting related to the target topic) but different information content through different preset prompt information generation methods.

[0042] Specifically, as an optional embodiment, the production agent can directly generate a second prompt message as the target prompt message based on the target topic according to a pure topic production method (i.e., an optional preset prompt message generation method); wherein, the second prompt message is used to instruct the generated copy to be related to the target topic.

[0043] For example, taking entertainment news as the target topic, the production agent can directly generate a prompt message such as "Please generate a social media post about entertainment news" based on the target topic of entertainment news, using a pure topic production method.

[0044] Specifically, as another optional embodiment, the production agent can also generate the above-mentioned target prompt information by using a combination of high response rate samples and topics (i.e., another optional preset prompt information generation method) through the steps shown in b1-b2: Step b1: Extract historical copywriting data with a user response rate higher than a preset threshold as target historical copywriting.

[0045] Here, historical copywriting data can be copywriting data generated and recalled by the production agent at a historical moment. This historical copywriting data can be obtained from the offline / online storage engine by the online agent.

[0046] Specifically, the production agent can select historical copywriting with a high user response rate (i.e., above a preset threshold) from historical copywriting data as target historical copywriting. In the generated target prompt information, it can prompt the copywriting generation model to use the target historical copywriting as a reference sample for copywriting generation. This helps the copywriting generation model output candidate copywriting with a similar style to the target historical copywriting, thereby obtaining higher quality candidate copywriting (i.e., with a higher user response rate).

[0047] Step b2: Generate a first prompt message as the target prompt message based on the target historical text and the target topic.

[0048] Here, the first prompt is used to indicate that the generated copy style matches the target historical copy and the copy content is related to the target topic. That is, compared with the second prompt generated by the pure topic method, when the first prompt is generated by the combination of high response rate samples and topic, the candidate copy generated by the same copy generation model based on the input first prompt will have a higher user response rate than the candidate copy generated based on the input second prompt. This is beneficial for the copy recommendation system to obtain higher quality (i.e., higher user response rate) candidate copy.

[0049] S203, the target prompt information and the topic data are respectively input into multiple copywriting generation models to obtain the copywriting output by the multiple copywriting generation models as candidate copywriting corresponding to the target topic.

[0050] Here, through the production agent in the copywriting recommendation system, target prompt information and topic data can be input into multiple copywriting generation models located outside the target application, and the copywriting output by each copywriting generation model can be recalled as the candidate copywriting.

[0051] Specifically, in the copywriting recommendation system, an MCP server is pre-deployed for the production agent. The production agent accesses the copywriting generation model located outside the target application through the MCP server. For the specific deployment method of the MCP server, please refer to the relevant description of the MCP server in the aforementioned step S201. The repeated parts will not be repeated here.

[0052] It should be noted that the above-mentioned copywriting generation models include, but are not limited to, AI generation models such as the LLM large language model and the Qwen model; however, this application embodiment does not impose any limitations on the specific model structure and number of the above-mentioned copywriting generation models.

[0053] S204, from the multiple candidate texts, determine the candidate text with the highest matching degree between the user and the target topic as the target text, and recommend the target text to the user.

[0054] Here, after the production agent receives multiple candidate texts from multiple recalls (i.e., candidate texts output by multiple text generation models respectively), it can hand them over to the review agent to optimize the multiple candidate texts and determine the optimal candidate text as the target text. After the target text is determined, it can be handed over to the online agent to write the target text into the offline / online storage engine for storage, and then handed over to the distribution agent to distribute the target text in a refined manner to the user who sent the above text generation request (i.e., recommend the target text to the user).

[0055] Specifically, the review agent can determine the candidate text with the highest degree of matching between the user and the target topic from multiple candidate texts according to the method shown in steps c1-c4 below: Step c1: Generate multi-dimensional user features that match the user information in the target application.

[0056] Here, user information refers to the user's multimodal information in the target application, including but not limited to: the user's basic information, historical session information, and image content (such as the user's avatar) in the target application.

[0057] Specifically, the aforementioned basic user information may include, but is not limited to: age, whether a user is new, occupation, height, weight, educational background, income, verification status, registration time, offline time, whether verified by a real person, city level, current province, current city, number of album photos, number of followers, number of friends, charm value, wealth value, charm level, wealth level, family status, balance of paid points, balance of free points, balance of free coins, whether a user is a VIP, whether a user with privileges, appearance score, level, distance, and other basic information from multiple dimensions. Based on the aforementioned basic user information, the review intelligence agent can construct multi-dimensional basic user features corresponding to the aforementioned basic user information (the number of feature values ​​in the multi-dimensional basic user features is the same as the number of information in the aforementioned basic user information).

[0058] Specifically, based on the aforementioned historical conversation information, the review agent can construct historical conversation sequence features corresponding to the aforementioned historical conversation information.

[0059] Specifically, for user avatar information in the target application, the review agent can extract image features from the user avatar information through the VIT (Vision Transformer) model, generate multi-dimensional (such as 768-dimensional) initial avatar embedding features, and then compress the initial avatar embedding features into 32-dimensional final avatar embedding features through the AutoEncoder.

[0060] Step c2: Calculate the matching degree between the multiple candidate texts and the multidimensional user features to obtain the user matching degree corresponding to the multiple candidate texts.

[0061] Here, by extracting multi-dimensional user features corresponding to users in a fine-grained manner, the review agent can input the multi-dimensional user features and each candidate text into the pre-trained fine-ranking model, and obtain the matching score between the multi-dimensional user features output by the fine-ranking model and each candidate text as the user matching degree corresponding to each candidate text.

[0062] Step c3: Calculate the matching degree between the multiple candidate texts and the target topic, and obtain the topic matching degree corresponding to the multiple candidate texts.

[0063] Here, the textual similarity or semantic similarity between each candidate text and the target topic can be calculated separately as the topic matching degree for each candidate text. Step c4: Based on the user matching degree and the topic matching degree, determine the target copy from the multiple candidate copy texts.

[0064] Here, in the reviewing intelligent agent, the candidate copy with the highest combined ranking of user matching degree and topic matching degree is taken as the optimal target copy for the user.

[0065] Based on the copywriting recommendation method provided in this application embodiment, in response to receiving a copywriting generation request input by a user, topic data related to the target topic is obtained from multiple data sources according to the target topic matching the copywriting generation request; target prompt information is generated to indicate the generation of copywriting related to the target topic according to a preset prompt information generation method; the target prompt information and topic data are respectively input into multiple copywriting generation models to obtain the copywriting output by the multiple copywriting generation models as candidate copywriting corresponding to the target topic; from the multiple candidate copywritings, the candidate copywriting with the highest matching degree between the user and the target topic is determined as the target copywriting, and the target copywriting is recommended to the user. In this way, this application can, on the basis of generating a sufficient number of high-quality candidate copywritings, finely distribute the target copywriting with the best matching degree to the user, thereby effectively overcoming the technical defects of the prior art in the lack of diversity in copywriting creation and the insufficient fineness of copywriting distribution.

[0066] Based on the same inventive concept, this application also provides a copywriting recommendation system corresponding to the above-mentioned copywriting recommendation method. Since the copywriting recommendation system in the embodiments of this application solves the problem in a similar way to the copywriting recommendation method in the embodiments of this application, the implementation of the copywriting recommendation system can refer to the implementation of the above-mentioned copywriting recommendation method, and the repeated parts will not be described again.

[0067] like Figure 1 As shown, the copywriting recommendation system corresponds to the target application, wherein the copywriting generation system includes: a selector agent, a data source agent, a production agent, a review agent, a deployment agent, a distribution agent, and a summary agent; wherein: A data source agent is used to respond to a text generation request input by a user, and to obtain topic data related to the target topic from multiple data sources based on the target topic that matches the text generation request. A production intelligent agent is used to generate target prompt information according to a preset prompt information generation method, which is used to instruct the generation of text related to the target topic; The target prompt information and the topic data are respectively input into multiple copywriting generation models, and the copywriting output by the multiple copywriting generation models is used as the candidate copywriting corresponding to the target topic. A distribution agent is used to determine the candidate text with the highest matching degree between the candidate text and the user and the target topic from a plurality of candidate texts, and recommend the target text to the user.

[0068] In one optional implementation, when acquiring topic data related to the target topic from multiple data sources, the data source agent is configured to: Obtain target internal data related to the target topic from the target application's internal data; Obtain target external data related to the target topic from external data sources; wherein, the external data source refers to other data sources located outside the target application; The target site's internal data and the target site's external data are used as the topic data.

[0069] In one optional implementation, when generating target prompt information to instruct the generation of text related to the target topic according to a preset prompt information generation method, the production agent is used to: From historical copywriting data, extract historical copywriting with a user response rate higher than a preset threshold as target historical copywriting; Based on the target historical text and the target topic, a first prompt message is generated as the target prompt message; wherein, the first prompt message is used to instruct the generation of text with a style that matches the target historical text and a content that is related to the target topic.

[0070] In an optional implementation, when generating target prompt information to instruct the generation of text related to the target topic according to a preset prompt information generation method, the production agent further performs the following: Based on the target topic, a second prompt message is generated as the target prompt message; wherein, the second prompt message is used to instruct the generation of copy content that is related to the target topic.

[0071] In an optional implementation, when the target prompt information and the topic data are respectively input into multiple copywriting generation models, and the copywriting output by the multiple copywriting generation models is obtained as candidate copywriting corresponding to the target topic, the production agent is used to: The target prompt information and the topic data are respectively input into multiple copywriting generation models located outside the target application, and the copywriting output by each copywriting generation model is recalled as the candidate copywriting.

[0072] In one alternative implementation, in the copy recommendation system, an MCP server is pre-deployed for the production agent, wherein the production agent accesses a copy generation model located outside the target application through the MCP server.

[0073] In one alternative implementation, when determining the candidate text with the highest matching degree between the user and the target topic from the plurality of candidate texts as the target text, the agent is used to: Based on the user information of the user in the target application, generate multi-dimensional user features that match the user information; Calculate the degree of matching between each of the candidate texts and the multidimensional user features to obtain the user matching degree corresponding to each of the candidate texts; Calculate the degree of matching between each of the candidate texts and the target topic to obtain the degree of topic matching corresponding to each of the candidate texts; Based on the user matching degree and the topic matching degree, the target copy is determined from multiple candidate copy texts.

[0074] Based on the copywriting recommendation system provided in this application embodiment, in response to receiving a copywriting generation request input by a user, the system obtains topic data related to the target topic from multiple data sources according to the target topic matching the copywriting generation request; generates target prompt information to indicate the generation of copywriting related to the target topic according to a preset prompt information generation method; inputs the target prompt information and topic data into multiple copywriting generation models respectively, and obtains the copywriting output by the multiple copywriting generation models as candidate copywriting corresponding to the target topic; from the multiple candidate copywritings, the candidate copywriting with the highest matching degree between the user and the target topic is determined as the target copywriting, and recommended to the user. In this way, this application can, on the basis of generating a sufficient number of high-quality candidate copywritings, finely distribute the target copywriting with the best matching degree to the user, thereby effectively overcoming the technical defects of the prior art, such as the lack of diversity in copywriting creation and the insufficient fineness of copywriting distribution.

[0075] Based on the same inventive concept, this application also provides an electronic device corresponding to the above-mentioned recommended method. Since the principle of solving the problem by the electronic device in the embodiments of this application is similar to that of the above-mentioned recommended method in the embodiments of this application, the implementation of the electronic device can refer to the implementation of the above-mentioned recommended method, and the repeated parts will not be described again.

[0076] Figure 3 A schematic diagram of the structure of an electronic device 300 provided in this application embodiment includes: a processor 301, a memory 302, and a bus 303. The memory 302 stores machine-readable instructions executable by the processor 301. When the electronic device runs a document recommendation method as described in the embodiment, the processor 301 communicates with the memory 302 via the bus 303. The processor 301 executes the machine-readable instructions, wherein the processor 301 executes the machine-readable instructions to perform the following steps: In response to receiving a text generation request from a user, topic data related to the target topic is obtained from multiple data sources based on the target topic that matches the text generation request. According to a preset prompt information generation method, generate target prompt information to instruct the generation of text related to the target topic; The target prompt information and the topic data are respectively input into multiple copywriting generation models, and the copywriting output by the multiple copywriting generation models is used as the candidate copywriting corresponding to the target topic. From the multiple candidate texts, the candidate text with the highest degree of matching between the user and the target topic is determined as the target text, and the target text is recommended to the user.

[0077] In an optional implementation, when acquiring topic data related to the target topic from multiple data sources, the processor 301 is configured to: Obtain target internal data related to the target topic from the target application's internal data; Obtain target external data related to the target topic from external data sources; wherein, the external data source refers to other data sources located outside the target application; The target site's internal data and the target site's external data are used as the topic data.

[0078] In an optional implementation, when generating target prompt information for instructing the generation of text related to the target topic according to a preset prompt information generation method, the processor 301 is configured to: From historical copywriting data, extract historical copywriting with a user response rate higher than a preset threshold as target historical copywriting; Based on the target historical text and the target topic, a first prompt message is generated as the target prompt message; wherein, the first prompt message is used to instruct the generation of text with a style that matches the target historical text and a content that is related to the target topic.

[0079] In an optional implementation, when generating target prompt information for instructing the generation of text related to the target topic according to a preset prompt information generation method, the processor 301 is further configured to: Based on the target topic, a second prompt message is generated as the target prompt message; wherein, the second prompt message is used to instruct the generation of copy content that is related to the target topic.

[0080] In an optional implementation, when the target prompt information and the topic data are respectively input into multiple copywriting generation models to obtain the copywriting output by the multiple copywriting generation models as candidate copywriting corresponding to the target topic, the processor 301 is used to: The target prompt information and the topic data are respectively input into multiple copywriting generation models located outside the target application, and the copywriting output by each copywriting generation model is recalled as the candidate copywriting.

[0081] In one alternative implementation, in the copy recommendation system, an MCP server is pre-deployed for the production agent, wherein the production agent accesses a copy generation model located outside the target application through the MCP server.

[0082] In an optional implementation, when determining the candidate text with the highest matching degree between the user and the target topic from the plurality of candidate texts as the target text, the processor 301 is configured to: Based on the user information of the user in the target application, generate multi-dimensional user features that match the user information; Calculate the degree of matching between each of the candidate texts and the multidimensional user features to obtain the user matching degree corresponding to each of the candidate texts; Calculate the degree of matching between each of the candidate texts and the target topic to obtain the degree of topic matching corresponding to each of the candidate texts; Based on the user matching degree and the topic matching degree, the target copy is determined from multiple candidate copy texts.

[0083] The electronic device provided in this application, in response to receiving a text generation request input by a user, retrieves topic data related to the target topic from multiple data sources based on the target topic matching the text generation request; generates target prompt information to instruct the generation of text related to the target topic according to a preset prompt information generation method; inputs the target prompt information and topic data into multiple text generation models respectively, and obtains the text output by the multiple text generation models as candidate texts corresponding to the target topic; from the multiple candidate texts, determines the candidate text with the highest matching degree between the user and the target topic as the target text, and recommends the target text to the user. Thus, this application can, based on generating a sufficient number of high-quality candidate texts, finely distribute the target text with the best matching degree to the user, thereby effectively overcoming the technical defects of the prior art, namely, the lack of diversity in text creation and the insufficient fineness of text distribution.

[0084] Based on the same inventive concept, embodiments of this application also provide a computer-readable storage medium storing a computer program, which is executed by a processor, wherein the processor performs the following steps: In response to receiving a text generation request from a user, topic data related to the target topic is obtained from multiple data sources based on the target topic that matches the text generation request. According to a preset prompt information generation method, generate target prompt information to instruct the generation of text related to the target topic; The target prompt information and the topic data are respectively input into multiple copywriting generation models, and the copywriting output by the multiple copywriting generation models is used as the candidate copywriting corresponding to the target topic. From the multiple candidate texts, the candidate text with the highest degree of matching between the user and the target topic is determined as the target text, and the target text is recommended to the user.

[0085] In one optional implementation, when acquiring topic data related to the target topic from multiple data sources, the processor is configured to: Obtain target internal data related to the target topic from the target application's internal data; Obtain target external data related to the target topic from external data sources; wherein, the external data source refers to other data sources located outside the target application; The target site's internal data and the target site's external data are used as the topic data.

[0086] In one optional implementation, when generating target prompt information for instructing the generation of text related to the target topic according to a preset prompt information generation method, the processor is configured to: From historical copywriting data, extract historical copywriting with a user response rate higher than a preset threshold as target historical copywriting; Based on the target historical text and the target topic, a first prompt message is generated as the target prompt message; wherein, the first prompt message is used to instruct the generation of text with a style that matches the target historical text and a content that is related to the target topic.

[0087] In an optional implementation, when generating target prompt information for instructing the generation of text related to the target topic according to a preset prompt information generation method, the processor is further configured to: Based on the target topic, a second prompt message is generated as the target prompt message; wherein, the second prompt message is used to instruct the generation of copy content that is related to the target topic.

[0088] In an optional implementation, when the target prompt information and the topic data are respectively input into multiple copywriting generation models to obtain the copywriting output by the multiple copywriting generation models as candidate copywriting corresponding to the target topic, the processor is configured to: The target prompt information and the topic data are respectively input into multiple copywriting generation models located outside the target application, and the copywriting output by each copywriting generation model is recalled as the candidate copywriting.

[0089] In one alternative implementation, in the copy recommendation system, an MCP server is pre-deployed for the production agent, wherein the production agent accesses a copy generation model located outside the target application through the MCP server.

[0090] In an optional implementation, when determining the candidate text with the highest matching degree between the user and the target topic from the plurality of candidate texts as the target text, the processor is configured to: Based on the user information of the user in the target application, generate multi-dimensional user features that match the user information; Calculate the degree of matching between each of the candidate texts and the multidimensional user features to obtain the user matching degree corresponding to each of the candidate texts; Calculate the degree of matching between each of the candidate texts and the target topic to obtain the degree of topic matching corresponding to each of the candidate texts; Based on the user matching degree and the topic matching degree, the target copy is determined from multiple candidate copy texts.

[0091] The computer-readable storage medium provided in this application embodiment, in response to receiving a text generation request input by a user, acquires topic data related to the target topic from multiple data sources based on the target topic matching the text generation request; generates target prompt information to instruct the generation of text related to the target topic according to a preset prompt information generation method; inputs the target prompt information and topic data into multiple text generation models respectively, and obtains the text output by the multiple text generation models as candidate texts corresponding to the target topic; from the multiple candidate texts, determines the candidate text with the highest matching degree between the user and the target topic as the target text, and recommends the target text to the user. Thus, this application can, based on generating a sufficient number of high-quality candidate texts, finely distribute the target text with the best matching degree to the user, thereby effectively overcoming the technical defects of the prior art, namely, the lack of diversity in text creation and the insufficient fineness of text distribution.

[0092] In this embodiment, the computer-readable storage medium can also execute other machine-readable instructions when the processor runs, to perform the copywriting recommendation method as described in other embodiments. For details on the specific steps and principles of the copywriting recommendation method, please refer to the description of the method-side embodiment, which will not be repeated here.

[0093] In the embodiments provided in this application, it should be understood that the disclosed systems and methods can be implemented in other ways. The system embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and there may be other division methods in actual implementation. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the coupling or direct coupling or communication connection shown or discussed may be through some communication interface; the indirect coupling or communication connection between systems or units may be electrical, mechanical, or other forms.

[0094] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0095] In addition, the functional units in the embodiments provided in this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0096] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0097] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. In addition, the terms "first", "second", "third", etc. are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0098] Finally, it should be noted that the above-described embodiments are merely specific implementations of this application, used to illustrate the technical solutions of this application, and not to limit them. The protection scope of this application is not limited thereto. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments, or make equivalent substitutions for some of the technical features, within the scope of the technology disclosed in this application; and these modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application. All should be covered within the protection scope of this application. Therefore, the protection scope of this application should be determined by the protection scope of the claims.

Claims

1. A copywriting recommendation method, characterized in that, The copywriting recommendation method is applied to the copywriting recommendation system corresponding to the target application, and the copywriting recommendation method includes: In response to receiving a text generation request from a user, topic data related to the target topic is obtained from multiple data sources based on the target topic that matches the text generation request. According to a preset prompt information generation method, generate target prompt information to instruct the generation of text related to the target topic; The target prompt information and the topic data are respectively input into multiple copywriting generation models, and the copywriting output by the multiple copywriting generation models is used as the candidate copywriting corresponding to the target topic. From the multiple candidate texts, the candidate text with the highest degree of matching between the user and the target topic is determined as the target text, and the target text is recommended to the user.

2. The copywriting recommendation method according to claim 1, characterized in that, The step of obtaining topic data related to the target topic from multiple data sources includes: Obtain target internal data related to the target topic from the target application's internal data; Obtain target external data related to the target topic from external data sources; wherein, the external data source refers to other data sources located outside the target application; The target site's internal data and the target site's external data are used as the topic data.

3. The copywriting recommendation method according to claim 1, characterized in that, The step of generating target prompt information according to a preset prompt information generation method to instruct the generation of text related to the target topic includes: From historical copywriting data, extract historical copywriting with a user response rate higher than a preset threshold as target historical copywriting; Based on the target historical text and the target topic, a first prompt message is generated as the target prompt message; wherein, the first prompt message is used to instruct the generation of text with a style that matches the target historical text and a content that is related to the target topic.

4. The copywriting recommendation method according to claim 1, characterized in that, The step of generating target prompt information to instruct the generation of text related to the target topic according to a preset prompt information generation method further includes: Based on the target topic, a second prompt message is generated as the target prompt message; wherein, the second prompt message is used to instruct the generation of copy content that is related to the target topic.

5. The copywriting recommendation method according to claim 1, characterized in that, The step of inputting the target prompt information and the topic data into multiple copywriting generation models respectively, and obtaining the copywriting output by the multiple copywriting generation models as candidate copywriting corresponding to the target topic, includes: The production agent in the copywriting recommendation system inputs the target prompt information and the topic data into multiple copywriting generation models located outside the target application, and recalls the copywriting output by each copywriting generation model as the candidate copywriting.

6. The copywriting recommendation method according to claim 5, characterized in that, In the copywriting recommendation system, an MCP server is pre-deployed for the production agent, wherein the production agent accesses a copywriting generation model located outside the target application through the MCP server.

7. The copywriting recommendation method according to claim 1, characterized in that, The step of determining the candidate text with the highest degree of matching between the candidate text and the user and the target topic from a plurality of candidate texts as the target text includes: Based on the user information of the user in the target application, generate multi-dimensional user features that match the user information; Calculate the degree of matching between each of the candidate texts and the multidimensional user features to obtain the user matching degree corresponding to each of the candidate texts; Calculate the degree of matching between each of the candidate texts and the target topic to obtain the degree of topic matching corresponding to each of the candidate texts; Based on the user matching degree and the topic matching degree, the target copy is determined from multiple candidate copy.

8. A copywriting recommendation system, characterized in that, The copywriting recommendation system corresponds to the target application, wherein the copywriting recommendation system includes: A data source agent is used to respond to a text generation request input by a user, and to obtain topic data related to the target topic from multiple data sources based on the target topic that matches the text generation request. A production intelligent agent is used to generate target prompt information according to a preset prompt information generation method, which is used to instruct the generation of text related to the target topic; The target prompt information and the topic data are respectively input into multiple copywriting generation models, and the copywriting output by the multiple copywriting generation models is used as the candidate copywriting corresponding to the target topic. A distribution agent is used to determine the candidate text with the highest matching degree between the candidate text and the user and the target topic from a plurality of candidate texts, and recommend the target text to the user.

9. An electronic device, characterized in that, include: The device includes a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor communicates with the memory via the bus. When the machine-readable instructions are executed by the processor, they perform the steps of the copywriting recommendation method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, performs the steps of the copywriting recommendation method as described in any one of claims 1 to 7.

Images