Data processing method and device, computing equipment and storage medium

Abstract

Embodiments of the present specification provide a data processing method and apparatus, a computing device, and a storage medium, the data processing method comprising: determining an initial text corresponding to a target item, and determining a target item variable corresponding to the target item, the target item variable being a variable that changes with time increase; the initial text and the target item variable are processed through a data processing model, target data containing the target item variable are obtained, the target data are data generated based on the initial text, and the target data are any modal data in multi-modal data; target project parameters corresponding to the target project variables are determined, a data processing result corresponding to the target project is generated based on the target project parameters and the target data, and the target project parameters are project parameters corresponding to the target project variables at the current time.

Description

Technical Field

[0001] Embodiments of this specification relate to the field of artificial intelligence technology, and in particular, to a data processing method. One or more embodiments of this specification also relate to a data processing device, a computing device, a computer-readable storage medium, and a computer program product. Background Art

[0002] With the continuous development of computer technology and artificial intelligence technology, neural network models can also be applied to various data processing projects for data processing and obtain data processing results.

[0003] Currently, in the process of using neural network models to process data, due to the failure to consider the continuously changing variables in various data processing projects, it is impossible to accurately obtain data processing results that are adapted to the data processing projects. Therefore, how to accurately determine the data processing results has become a technical problem that urgently needs to be solved. Summary of the Invention

[0004] In view of this, embodiments of this specification provide a data processing method. One or more embodiments of this specification also relate to a data processing device, a computing device, a computer-readable storage medium, and a computer program product to solve the technical defects existing in the prior art.

[0005] According to the first aspect of the embodiments of this specification, a data processing method is provided, including: Determine the initial text corresponding to the target project, and determine the target project variables corresponding to the target project, where the target project variables are variables that change over time; Use a data processing model to process the initial text and the target project variables to obtain target data including the target project variables, where the target data is data generated based on the initial text, and the target data is any one of the multimodal data; Determine the target project parameters corresponding to the target project variables, and generate the data processing result corresponding to the target project based on the target project parameters and the target data, where the target project parameters are the project parameters corresponding to the target project variables at the current time.

[0006] According to the second aspect of the embodiments of this specification, a data processing device is provided, including: A variable determination module configured to determine the initial text corresponding to the target project, and determine the target project variables corresponding to the target project, where the target project variables are variables that change over time; A model processing module, configured to process the initial text and the target project variable by using a data processing model to obtain target data including the target project variable, where the target data is data generated based on the initial text, and the target data is any one of the multimodal data; A data processing module, configured to determine target project parameters corresponding to the target project variable, and generate a data processing result corresponding to the target project based on the target project parameters and the target data, where the target project parameters are project parameters corresponding to the target project variable at the current time.

[0007] According to a third aspect of the embodiments of the present specification, a computing device is provided, including: A memory and a processor; The memory is used to store computer programs / instructions, and the processor is used to execute the computer programs / instructions. When the computer programs / instructions are executed by the processor, the steps of the above data processing method are implemented.

[0008] According to a fourth aspect of the embodiments of the present specification, a computer-readable storage medium is provided, which stores computer programs / instructions. When the computer programs / instructions are executed by a processor, the steps of the above data processing method are implemented.

[0009] According to a fifth aspect of the embodiments of the present specification, a computer program product is provided, including computer programs / instructions. When the computer programs / instructions are executed by a processor, the steps of the above data processing method are implemented.

[0010] A data processing method provided by one or more embodiments of the present specification can, during the data processing process, determine the initial text that needs to be processed by a data processing model, and the target project variable corresponding to the target project and changing over time; use the data processing model to perform data generation processing on the initial text and the target project variable to obtain target data generated from the initial text and including the target project variable; finally, use the target project parameters corresponding to the target project variable to process the target data, so as to accurately generate a data processing result corresponding to the target project, thereby realizing that the continuously changing variables in the target project can be considered during the data processing process, and solving the problem that it is impossible to accurately obtain a data processing result adapted to the data processing project. Description of the Drawings

[0011] Figure 1 is an application schematic diagram of a data processing method provided by an embodiment of the present specification; Figure 2 is a flowchart of a data processing method provided by an embodiment of the present specification; Figure 3 It is a schematic diagram of a dynamic factor model and a factor service system in a data processing method provided by an embodiment of this specification; Figure 4 It is a schematic diagram of marketing promotion in a data processing method provided by an embodiment of this specification; Figure 5 It is a flowchart of the processing procedure of a data processing method provided by an embodiment of this specification; Figure 6 It is a structural block diagram of a computing device provided by an embodiment of this specification. Detailed implementation manners

[0012] In the following description, many specific details are set forth in order to provide a thorough understanding of this specification. However, this specification can be implemented in many other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the connotation of this specification. Therefore, this specification is not limited by the specific implementations disclosed below.

[0013] The terms used in one or more embodiments of this specification are only for the purpose of describing specific embodiments, and are not intended to limit one or more embodiments of this specification. The singular forms "a", "the", and "said" used in one or more embodiments of this specification and the appended claims are also intended to include the plural forms, unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used in one or more embodiments of this specification refers to and includes any or all possible combinations of one or more of the associated listed items.

[0014] It should be understood that although the terms first, second, etc. may be used in one or more embodiments of this specification to describe various information, such information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of one or more embodiments of this specification, the first may also be referred to as the second, and similarly, the second may also be referred to as the first. Depending on the context, the word "if" as used herein may be interpreted as "when" or "while" or "in response to determining".

[0015] In addition, it should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in one or more embodiments of this specification are all information and data that have been authorized by the user or fully authorized by all parties, and the collection, use, and processing of relevant data need to comply with the relevant laws, regulations, and standards of relevant countries and regions, and corresponding operation entrances are provided for users to select to authorize or reject.

[0016] In one or more embodiments of this specification, a large model refers to a deep learning model with a large number of model parameters, usually including hundreds of millions, tens of billions, hundreds of billions, trillions or even more than one quadrillion model parameters. A large model can also be called a Foundation Model. Through pre-training of the large model with a large amount of unlabeled corpus, a pre-trained model with more than one hundred million parameters is produced. This model can adapt to a wide range of downstream tasks and has good generalization ability. For example, large language models (LLMs), multi-modal pre-training models, etc.

[0017] When a large model is actually applied, it only needs to fine-tune the pre-trained model with a small number of samples to be applied to different tasks. Large models can be widely applied in the fields of natural language processing (NLP), computer vision, etc. Specifically, they can be applied to tasks in the field of computer vision such as visual question answering (VQA), image captioning (IC), image generation, etc., and tasks in the field of natural language processing such as text-based sentiment classification, text summary generation, machine translation, etc. The main application scenarios of large models include digital assistants, intelligent robots, search, online education, office software, e-commerce, intelligent design, etc.

[0018] First, the noun terms involved in one or more embodiments of this specification are explained.

[0019] AIGC (AI-Generated Content): refers to content generated by artificial intelligence; this technology uses artificial intelligence algorithms such as machine learning and natural language processing to create various types of content such as text, images, audio, and video.

[0020] Dynamic factor: refers to a factor that changes dynamically; in this specification, by defining a factor service, the input and output parameter formats of the dynamic factor are determined, and the calculation logic is customized inside the factor service, so as to achieve the effect of inputting dynamic input parameters and outputting dynamic output parameters.

[0021] Large Language Model (LLM): It refers to an artificial intelligence model trained based on deep learning technology, especially neural networks. Such models are designed to understand and generate natural language text. For example, large language models (or LLMs) include the GPT series, BERT, T5, etc. They have achieved remarkable results in the field of natural language processing (NLP) and can be applied to various fields such as customer service, education, creation, content recommendation, etc.

[0022] Material placement: It refers to the process of releasing or spreading pre-produced materials (such as pictures, videos, texts, etc.) through specific channels in advertising, marketing, or content promotion to reach the target audience and achieve marketing goals. Material placement is an important link in achieving advertising effects and brand communication.

[0023] Fund material placement: In the operation of fund projects, it is necessary to release the produced marketing materials on the internal and external channels of the fund to attract potential users or improve the efficiency of traffic conversion.

[0024] Fund material booth: It means that there are respective display contents on different floors or channels within the corresponding publicity positions of the fund. For these display contents, material management and material placement can be carried out in the form of booths for the display contents.

[0025] With the continuous development of computer technology and artificial intelligence technology, neural network models can also be applied to various data processing projects for data processing and obtain data processing results. Currently, in the process of using neural network models to process data, due to the lack of consideration of the continuously changing variables in various data processing projects, it is impossible to accurately obtain data processing results that are suitable for the data processing projects.

[0026] For example, when a user clicks or triggers a specific link to enter the publicity position of an Internet finance project (such as a fund), different materials will be seen on the display booths of the publicity position. In the past, the materials of all booths were manually created and configured by the operation staff. With the development of large language models, the application of AIGC (Automated Content Generation) in copywriting creation has become increasingly widespread, improving the creation efficiency. In the operation of Internet finance projects, the operation mode of publicity personnel has gradually changed from manually creating limited materials to using large models to generate a batch of materials for placement. In this context, publicity personnel sort out the description of product-related benefits for Internet finance projects and rely on large models to produce the material copywriting, pictures, videos, etc. to be placed. The key information of the product benefits of Internet finance projects can be abstracted into two types: static and dynamic. Therefore, it is necessary to design a set of solutions to make the material AIGC compatible with the recognition, creation, and rendering to customers of static and dynamic information at the same time.

[0027] In view of the above problems, the present method provides a solution which is to provide a fixed piece of text for the large model to extract relevant benefit points, so as to form a short fixed text output. The specific steps include: 1. Input of static text. For example, the static text can be "Self-health Fund Recall Strategy: For the funds redeemed by users, and at the same time, the product meets the condition that the net value has continuously increased recently". 2. AIGC by the large model. 3. Output of fixed text, for example, (The fund you redeemed has had consecutive increases recently).

[0028] However, the disadvantages of this solution are as follows: 1. Diversity limitation: If the products to be launched are multiple, the pure static text limits the diversity of text expression and cannot generate personalized texts according to different product characteristics. Personalized texts such as "It has had consecutive increases for X days recently and is worth repurchasing", where X can vary dynamically according to different redeemed products. 2. Need to update time-sensitive materials periodically: If the material launch is restricted to a single product, the text can reflect the characteristics of the product, and directly display the static text such as "It has had consecutive increases for 5 days recently". Since the numbers are time-sensitive, it is necessary to update the materials periodically, which reduces the life cycle of the materials and increases the creation workload.

[0029] Based on this, in this specification, a data processing method is provided. This specification also relates to a data processing device, a computing device, and a computer-readable storage medium, which will be described in detail one by one in the following embodiments.

[0030] Considering that the number of model parameters of the large model is huge and the computing resources of mobile terminals are limited, the data processing method provided by the embodiments of the present application can be applied to Figure 1 the application scenarios shown, but not limited thereto. In the application scenarios shown in Figure 1 the large model is deployed in the server 104. The server 104 can be connected to one or more clients 102 through a local area network connection, a wide area network connection, an Internet connection, or other types of data networks. Here, the clients 102 can include but are not limited to: smart phones, tablet computers, laptop computers, palm computers, personal computers, smart home devices, in-vehicle devices, etc. The client 102 can interact with the user through a graphical user interface to realize the invocation of the large model, and further realize the method provided by the embodiments of this specification.

[0031] In the embodiments of this specification, the system composed of the client 102 and the server 104 may perform the following steps: The server 104 executes using the dynamic factor and the static copywriting as the large model prompt raw materials and inputs them into the large model, uses the large model to generate marketing promotion information including the dynamic factor, and then uses the index parameters corresponding to the dynamic factor to complete the marketing promotion information to obtain the marketing promotion data for marketing promotion. The client 102 executes rendering the marketing promotion data. It should be noted that in the case where the operating resources of the client 102 can meet the deployment and operating conditions of the large model, the embodiments of this application can be carried out in the client device.

[0032] See Figure 2 , Figure 2 shows a flowchart of a data processing method provided according to an embodiment of this specification, which specifically includes the following steps.

[0033] Step 202: Determine the initial text corresponding to the target project, and determine the target project variable corresponding to the target project, where the target project variable is a variable that changes over time.

[0034] Among them, the target project can be understood as a project that needs to be marketed, for example, the target project can be understood as an application program, a computer service (such as a financial management service, a health service, an online shopping service, etc.), a hardware device, a product, etc.

[0035] The initial text can be understood as the text that can generate the target data. For example, the initial text can be understood as the text or prompt information of the target data such as the marketing promotion copywriting and the marketing promotion image; The target project variable can be understood as a variable that changes over time during the execution of the target project. The target project variable can be understood as a kind of dynamic factor; the target project variable can be understood as a marketing promotion variable, for example, the growth amount, the sales amount, the performance index, etc.

[0036] In one or more embodiments provided in this specification, the determining the target project variable corresponding to the target project includes: Determine the project variable identifier corresponding to the target project, and based on the project variable identifier, obtain the target project variable corresponding to the target project from the variable storage structure; Among them, the project variable identifier can be understood as an identifier that uniquely identifies a target project variable. For example, the project variable identifier can be information such as a number or a name; the variable storage structure can be understood as a data structure for storing the target project variable and the project variable parameters. For example, the variable storage structure can be a dynamic factor model.

[0037] Taking the application of the data processing method provided in this specification in the marketing and promotion scenario as an example, the data processing method will be described. It should be noted that the data processing method provided in this specification is an AIGC method based on dynamic factors, where the dynamic factor is the target project variable; AIGC is the data processing result. Based on this, this method provides a dynamic factor (such as the number of consecutive rising months describing a product), and allows the large model to generate copywriting (i.e., the data processing result) according to this factor. In the generated copywriting, placeholders are used to represent this dynamic factor, so that when finally presented to the user, it can be filled in real-time through relevant services.

[0038] In order to quickly and accurately obtain the dynamic factor (referred to as the factor for short), this method designs a dynamic factor model; this dynamic factor model is used to uniquely identify the factor and its execution logic. Specifically refer to Figure 3 , Figure 3 Figure 7 is a schematic diagram of the dynamic factor model and the factor service system in a data processing method provided by an embodiment of this specification; among them, the dynamic factor model includes a factor unique code (i.e., the project variable identifier), an input parameter model, an output parameter model, and a factor execution logic; among them, the factor unique code refers to the code (code) that uniquely identifies a dynamic factor; the input parameter model is an interface for writing the actual parameters corresponding to the factor (i.e., the number of consecutive rising months, the consecutive rising turnover) into the dynamic factor model; the output parameter model is an interface for obtaining the actual parameters corresponding to the factor from the dynamic factor model; the factor execution logic refers to the underlying logic (such as underlying code) required for the dynamic factor model to perform processes such as input parameters and output parameters on the dynamic factor. Based on this, this method can accurately obtain the dynamic factor corresponding to the target project (i.e., the target project variable) from the dynamic factor model (i.e., the variable storage structure) according to the factor unique code; facilitating subsequent generation of marketing and promotion data corresponding to the target project based on the dynamic factor (i.e., the data processing result).

[0039] In addition, this method also designs a factor service system for factor service registration and execution; this factor service system includes a dynamic factor model and a large model; what the factor service system can achieve includes: factor registration and factor execution, where factor registration refers to defining an instance of the factor model. Specifically, it can be understood that the publicity staff pre-define multiple dynamic factors according to experience and register the dynamic factors (such as the factor of the consecutive rising months of the fund, and the identifier of this factor is "fund_continue_growup_month") into the dynamic factor model.

[0040] Factor execution refers to returning the result output parameter according to the factor code and input parameters; specifically, it can be understood as the execution logic in the process of using the dynamic factor model and the large model to generate marketing and promotion materials.

[0041] In one or more embodiments provided in this specification, before determining the target project variable corresponding to the target project, the following steps are further included: Create the target project variable in the variable storage structure using the project variable identifier; Obtain the target project parameter corresponding to the target project variable and store the target project parameter in the variable storage structure.

[0042] Among them, the target project parameter can be understood as the specific parameter corresponding to the target project variable, and this target project parameter will change with the passage of time. For example, the target project parameter can be an index parameter, a project revenue parameter, etc.

[0043] Continuing with the above example, this method can pre-define multiple dynamic factors and register the dynamic factors (such as the fund consecutive rising month factor, and the identifier of this factor is "fund_continue_growup_month") into the dynamic factor model; Then, real-time detect the current index parameter of this dynamic factor (for example, the fund consecutive rising month factor is May), and write this current index parameter into the dynamic factor model. Thus, it is convenient to quickly and accurately obtain the target project variable and the target project parameter subsequently.

[0044] Step 204: Process the initial text and the target project variable using a data processing model to obtain target data containing the target project variable, where the target data is data generated based on the initial text, and the target data is any one of the multimodal data.

[0045] In one or more embodiments provided in this specification, the initial text is an initial promotional text, the target data is a target promotional text, and the data processing model is a multimodal large model; The process of using the data processing model to process the initial text and the target project variable to obtain target data containing the target project variable includes: Input the initial promotional text and the target project variable into the multimodal large model, and use the multimodal large model to generate a promotional text according to the initial promotional text and the target project variable to obtain a target promotional text containing the target project variable.

[0046] Continuing with the above example, this method uses static copy input + dynamic factors as the model input data. Among them, the static copy can be "Robust Fund Recall Strategy: For the robust funds redeemed by users, and at the same time, the net value of this product has continuously increased in recent months", and the corresponding dynamic factor can be "fund_continue_growup_month"; combine the two as the model input data to obtain "Robust Fund Recall Strategy: For the robust funds redeemed by users, and at the same time, the net value of this product has continuously increased in recent months, and the dynamic factor for the continuous growth of the net value is 'fund_continue_growup_month'".

[0047] Then, input the above model input data into the large model, and use the large model to perform language inference on the model input data to obtain the model output copy; the model output copy can be "The net value has continuously increased for {fund_continue_growup_month} months recently and is worthy of repurchase". Based on this, it can be seen that this method uses the dynamic factor as the key large model prompt raw material and emphasizes using the unique code of the dynamic factor as a placeholder output when generating and outputting the copy.

[0048] Based on the above embodiments, it can be seen that this method uses the large model to generate the target copy containing the dynamic factor based on the dynamic factor and the initial copy, thereby improving the personalized expression of the copy and enhancing the timeliness of the copy. This operation brings the following significant technical effects: First, using the advanced natural language processing large model, it can quickly generate high-quality and attractive promotional copies. Compared with the traditional manual writing method, it greatly shortens the time cycle of copy creation and improves work efficiency. Second, the automated copy generation process reduces the dependence on professional copywriters and lowers the labor cost. At the same time, since it reduces errors caused by human factors, such as spelling mistakes or inaccurate information, it ensures the quality and reliability of the copy output. Finally, this method is suitable for projects or products that need to generate a large number of different types of promotional copies. It can provide diverse content support for different marketing activities in a short time and improves the speed and flexibility of responding to market changes.

[0049] In one or more embodiments provided in this specification, the target data is a target image, and the data processing model is a multi-modal large model; The processing of the initial text and the target project variable by using the data processing model to obtain the target data containing the target project variable includes: Input the initial text and the target project variable into the multi-modal large model, and use the multi-modal large model to perform image generation according to the initial text and the target project variable to obtain the target image containing the target project variable.

[0050] Continuing with the above example, input the above model input data into the large model, and use the large model to perform inference on the model input data to obtain the model output image; the model output image can be an image containing "Continuously rising for {fund_continue_growup_month} months recently, worthy of repurchase". It should be noted that when promoting the production and output of the image, the unique code of the dynamic factor can be embedded in the model output image.

[0051] Based on the above embodiments, the method inputs the model input data into the large model and uses the large model to perform inference processing on the input data, thereby obtaining a high-quality model output image. The following remarkable technical effects are achieved: First, by adopting advanced artificial intelligence large model technology, it can quickly and accurately generate attractive and targeted promotional images. Compared with traditional manual design, it greatly shortens the image production cycle and improves work efficiency. Second, the method allows adjusting the model input data according to the dynamic factor, and then generating promotional images that meet the requirements of different application scenarios. Moreover, the process of automatically generating promotional images reduces the dependence on designers, lowers labor costs, and also avoids design errors caused by human factors, ensuring the stability and reliability of image output.

[0052] In one or more embodiments provided in this specification, the target data is a target video, and the data processing model is a multi-modal large model; The step of using the data processing model to process the initial text and the target project variable to obtain the target data containing the target project variable includes: Input the initial text and the target project variable into the multi-modal large model, and use the multi-modal large model to generate a video according to the initial text and the target project variable to obtain the target video containing the target project variable.

[0053] Continuing with the above example, input the above model input data into the large model, and use the large model to perform inference on the model input data to obtain the model output video; the model output video can be a video containing "Continuously rising for {fund_continue_growup_month} months recently, worthy of repurchase". It should be noted that when promoting the production and output of the video, the unique code of the dynamic factor can be embedded in all or part of the video frames of the model output video.

[0054] Based on the above embodiments, the method inputs the model input data into a large model and uses the large model to perform inference processing on the input data to generate a video suitable for publicity and promotion. This process brings the following significant technical effects: First, by using a large model that combines natural language processing and video generation technologies, the method provides a new video content creation method. It allows users to automatically generate corresponding high-quality publicity and promotion videos only by inputting text descriptions or concepts, greatly simplifying the video production process. Second, by adjusting the model input data according to specific market demands or brand positioning, the method can create highly customized video content. This can not only accurately convey the core value of products or services but also effectively attract the attention of the target audience and enhance market competitiveness. Second, the automated video generation process reduces the dependence on professional video production teams and lowers labor costs. At the same time, it also greatly shortens the time cycle from creative concept to finished product output, improves work efficiency, and makes it possible to quickly respond to market changes.

[0055] Step 206: Determine the target project parameters corresponding to the target project variable, and generate the data processing result corresponding to the target project based on the target project parameters and the target data, where the target project parameters are the project parameters corresponding to the target project variable at the current time.

[0056] In one or more embodiments provided in this specification, the determining the target project parameters corresponding to the target project variable includes: Obtain the target project parameters corresponding to the target project variable at the current time from the variable storage structure, where the target project parameters are the parameters corresponding to the target project variable at different time points.

[0057] Continuing with the above example, after obtaining the AIGC output by the large model, the index parameters corresponding to the dynamic factors can be obtained from the dynamic factor model, which is convenient for generating the marketing promotion data (i.e., the data processing result) of the target project based on the index parameters subsequently.

[0058] Based on the above embodiments, the method can accurately and quickly obtain the index parameters of the dynamic factors, thereby facilitating the subsequent generation of the marketing promotion data of the target project based on the index parameters, ensuring that the generated marketing promotion data highly matches the actual situation. This enables marketing promotion to quickly adapt to changing market conditions and adjust marketing strategies in a timely manner according to the latest market feedback to maintain a competitive advantage.

[0059] In one or more embodiments provided in this specification, the data processing result is the project marketing data corresponding to the target project; Generating the data processing result corresponding to the target project based on the target project parameters and the target data includes: Fusing the target project parameters and the target data to generate project marketing data corresponding to the target project.

[0060] Continuing with the previous example, after obtaining the current metric parameter, it can be embedded into the AIGC output by the large model to obtain the marketing promotion data for the target project.

[0061] Based on the above embodiments, this method generates marketing promotion data for the target project according to the metric parameters of the dynamic factor, thus ensuring that the generated marketing promotion data highly coincides with the actual situation. This enables the enterprise to make more accurate and effective marketing based on the actual situation, enhancing its market competitiveness. Moreover, this approach not only improves the relevance of marketing activities but also greatly shortens the time cycle from data collection to strategy execution, enabling the enterprise to quickly respond to market changes; since the marketing promotion data is generated based on real-time metric parameters, it can better meet the actual needs and preferences of target customers; personalized marketing information can not only attract the attention of more potential customers but also effectively improve the satisfaction of existing customers.

[0062] In one or more embodiments provided in this specification, the target data is target promotional text, target image, or target video, and the target project parameters are target project marketing metrics; The fusing the target project parameters and the target data to generate project marketing data corresponding to the target project includes: Using the target project marketing metric to replace the target project variables included in the target promotional text to obtain an updated promotional text, and determining the updated promotional text as the project marketing data corresponding to the target project; or Using the target project marketing metric to replace the target project variables included in the target image to obtain an updated image, and determining the updated image as the project marketing data corresponding to the target project; or Using the target project marketing metric to replace the target project variables included in the target video to obtain an updated video, and determining the updated video as the project marketing data corresponding to the target project.

[0063] Continuing with the above example, in the stage of material placement and rendering of this method, when it is determined that the text, image, or video output by the model contains dynamic factor placeholders, the service of the factor model is called to obtain the index parameters corresponding to the dynamic factors, and the text information, image information, or video is complemented using these index parameters, so as to be combined into the final marketing promotion materials (such as text, advertising images, marketing videos, etc.).

[0064] Based on the above embodiments, this method generates various types of marketing promotion data including promotional text, promotional images (such as posters), and promotional videos by using real-time index parameters. Thus, it significantly improves the efficiency and diversity of marketing promotion, and greatly shortens the content creation cycle. Compared with the traditional manual production method, this technology can quickly respond to market changes and timely launch marketing materials that meet market needs. Moreover, through real-time index parameters, the system can generate corresponding marketing content according to information such as current market trends and growth conditions, which ensures that each piece of text, image, or video produced has a high degree of relevance and pertinence, and can more effectively attract the attention of the target audience. This method supports the generation of various forms of marketing promotion materials, from text to vision to multimedia, greatly enriching the manifestation forms of marketing strategies. At the same time, due to the powerful algorithm support of the large model, an automated generation process is realized, reducing the dependence on professional designers and copywriters and lowering the labor cost.

[0065] In one or more embodiments provided in this specification, after generating the data processing result corresponding to the target project based on the target project parameters and the target data, it further includes: Real-time monitoring of the target project parameters corresponding to the target project variables, and obtaining the updated project parameters corresponding to the target project variables when it is determined that the target project parameters have been updated; Using the updated project parameters and the target data to generate the updated data processing result corresponding to the target project.

[0066] Continuing with the above example, this method can detect in real time whether the indicator parameters corresponding to the dynamic factors have been updated. In the case of determining that an update has occurred, the service of the factor model can be called to obtain the latest indicator parameters corresponding to the dynamic factors, and the historical indicator parameters in the copywriting information, image information, or video can be replaced with the latest parameters. Alternatively, using the latest parameters, the copywriting information, image information, or video output by the large model can be replenished again, so as to obtain marketing promotion materials (such as copywriting, advertising images, marketing videos, etc.) that conform to the latest project progress. Thus, by monitoring the changes in indicator parameters in real time, the system can adjust and update the marketing promotion content in a timely manner. This ensures that the generated marketing promotion data always reflects the latest market conditions, improving the accuracy and effectiveness of the information. The need for manual intervention is reduced through an automated update process, which not only reduces labor costs but also avoids information lag caused by human delays.

[0067] In one or more embodiments provided in this specification, after generating the data processing result corresponding to the target project based on the target project parameters and the target data, it further includes: Render the data processing result to the marketing promotion module so that the marketing promotion module can display the data processing result.

[0068] Among them, the marketing promotion module can be understood as a module for conducting marketing promotion. For example, the marketing promotion module can be a hardware device such as a smartphone or an advertising screen; or, the marketing promotion module can be a software module such as a pop-up window or an advertising space of an application.

[0069] Continuing with the above example, after obtaining the marketing promotion materials, the marketing promotion materials can be rendered to various marketing promotion modules (such as smartphones, advertising screens, clients, web pages, etc.) to achieve marketing promotion. For example, Figure 4 , Figure 4 A schematic diagram of material placement for the display content (i.e., marketing promotion materials) is provided. Figure 4 The two boxes in are, respectively, the focused card booth - material placement and the shelf booth - material placement; the materials displayed in the two booths are the materials generated using the large model. And, Figure 4 In the key material copywriting "Positive returns in X quarters since its establishment" in, and Figure 4 In the key material copywriting "Has risen continuously for X months recently, worthy of repurchase" in, X in the two copywritings is a dynamic factor, and this value will display different values based on different funds and will also change over time for the same product.

[0070] A data processing method provided by one or more embodiments of this specification can, during the data processing process, determine the initial text that needs to be processed by a data processing model, as well as the target project variables corresponding to the target project and that change over time; use the data processing model to perform data generation processing on the initial text and the target project variables to obtain target data generated from the initial text and containing the target project variables; and finally, use the target project parameters corresponding to the target project variables to process the target data, thereby accurately generating a data processing result corresponding to the target project, thus achieving the consideration of the continuously changing variables in the target project during the data processing process and solving the problem of being unable to accurately obtain a data processing result adapted to the data processing project.

[0071] The following combines the attached Figure 5 , taking the application of the data processing method provided by this specification in marketing promotion as an example, to further illustrate the data processing method. Among them, Figure 5 shows the processing process flowchart of a data processing method provided by an embodiment of this specification, specifically including the following steps.

[0072] It should be noted that the data processing method of this specification provides a factor service system for factor service registration and execution; this factor service system includes a dynamic factor model and a large model; what the factor service system can achieve includes: factor registration and factor execution. Among them, factor registration refers to defining an instance of a factor model, which can specifically be understood as that a publicity person pre-defines multiple dynamic factors according to experience and registers the dynamic factors (such as the fund consecutive rising months factor, and the identifier of this factor is "fund_continue_growup_month") into the dynamic factor model.

[0073] In addition, this method designs a dynamic factor model; this dynamic factor model is used to uniquely identify a factor and its execution logic. Specifically referring to Figure 3 , the dynamic factor model includes a factor unique code (i.e., the project variable identifier), an input parameter model, an output parameter model, and a factor execution logic; among them, the factor unique code refers to the code that uniquely identifies a dynamic factor; the input parameter model is an interface for writing the actual parameters corresponding to the factor (i.e., the consecutive rising months, consecutive rising turnover) into the dynamic factor model; the output parameter model is an interface for obtaining the actual parameters corresponding to the factor from the dynamic factor model; the factor execution logic refers to the underlying logic (such as underlying code) that needs to be executed during the process of the dynamic factor model performing input parameter, output parameter, etc. processing on the dynamic factor.

[0074] Specifically, the execution steps of this data processing method are as follows: Step 502: Preparation stage.

[0075] This preparation node points to the factor service system to register dynamic factors. For example, the dynamic factor can be: register the factor of consecutive rising months of the fund: fund_continue_growup_month.

[0076] Step 504: Production stage.

[0077] This production stage includes: static copy input + dynamic factor, large model AIGC, dynamic factor placeholder + static copy output.

[0078] Among them, static copy input + dynamic factor means: combining static copy input and dynamic factors into model input data.

[0079] The static copy can be "Robust fund recall strategy: For the robust funds redeemed by users, and at the same time this product meets the condition that the net value has continuously increased in recent months", and the corresponding dynamic factor can be "fund_continue_growup_month"; combining the two into model input data, we get "Robust fund recall strategy: For the robust funds redeemed by users, and at the same time this product meets the condition that the net value has continuously increased in recent months, and the dynamic factor for continuous net value growth is 'fund_continue_growup_month'".

[0080] Among them, large model AIGC means: inputting the above model input data into the large model, and using the large model to generate material content for the model input data.

[0081] Among them, dynamic factor placeholder + static copy output means: obtaining the model output copy; the initial copy can be in the form of dynamic factor placeholder + static copy, for example, "Recently, it has continuously risen for {fund_continue_growup_month} months, which is worth repurchasing". Based on this, it can be seen that this method uses dynamic factors as the key large model prompt raw materials, and emphasizes using the unique code of the dynamic factor as a placeholder output when generating and outputting the copy.

[0082] In addition, the AIGC output by the large model can also be images and videos. The specific implementation method is as follows: Input the above model input data into the large model, and use the large model to perform language inference on the model input data to obtain the model output image; the model output image can be an image containing "Recently, it has continuously risen for {fund_continue_growup_month} months, which is worth repurchasing". Among them, this method can embed the unique code of the dynamic factor into the model output image.

[0083] Input the above model input data into a large model, and use the large model to perform language inference on the model input data to obtain a model output video; the model output video can be a video containing "The price has been rising continuously for {fund_continue_growup_month} months recently, and it is worth repurchasing". Among them, this method can embed the unique code of the dynamic factor into all or part of the video frames of the model output video.

[0084] Step 506: Rendering stage.

[0085] Specifically, the rendering stage includes: requesting from the factor service system and assembling the final copy expression according to the return of the factor service system.

[0086] Among them, requesting from the factor service system means that in the material placement and rendering stage of this method, when it is determined that the copywriting, image, or video output by the model contains a dynamic factor placeholder, the index parameters corresponding to the dynamic factor are obtained by calling the service of the factor model.

[0087] Among them, assembling the final copy expression according to the return of the factor service system means using the index parameters obtained from the dynamic factor model to complete the copywriting information, image information, or video, so as to combine them into the final marketing promotion materials (such as copywriting, advertising images, marketing videos, etc.).

[0088] After obtaining the marketing promotion materials, the marketing promotion materials can be rendered to various marketing promotion modules (such as smartphones, advertising screens, clients, web pages, etc.) to achieve the marketing promotion of products.

[0089] In addition, it should also be noted that this method can detect in real time whether the index parameters corresponding to the dynamic factor have been updated. In the case of determining that an update has occurred, the latest index parameters corresponding to the dynamic factor can be obtained by calling the service of the factor model; Then, use the latest parameters to replace the historical index parameters in the copywriting information, image information, or video, or use the latest parameters to re-complete the copywriting information, image information, or video output by the large model, so as to obtain marketing promotion materials (such as copywriting, advertising images, marketing videos, etc.) that conform to the latest project progress, so as to ensure that the marketing promotion materials always keep up with the times.

[0090] Based on the above embodiments, the data processing method described in this specification provides an AIGC dynamic material solution based on dynamic factors, which generates personalized and timely copywriting by leveraging large language models. By introducing dynamic factors, such as the number of consecutive rising days of a fund product, the large model can identify and accurately use these dynamic factors, enabling the created materials to support the dynamic display of multiple products over different time periods, effectively coping with changes in different product characteristics and time, and improving the overall creation efficiency. Moreover, it realizes the real-time adjustment and intelligent generation of copywriting, thus effectively responding to changes in product characteristics. Compared with traditional static copywriting generation methods, this method enhances the flexibility and market adaptability of the copywriting, ensures that the copywriting always keeps up with the times, and improves the overall creation efficiency.

[0091] Compared with the above-mentioned fixed copywriting output solution, through the design of dynamic factors, different input parameters can dynamically return different output parameters. In the case of promoting multiple products, different products can achieve personalized expression; through the design of dynamic factors, when the same input parameters are rendered and consulted at different times, the underlying service model will return corresponding values based on the current time, thus realizing the dynamic expression of the copywriting over time and ensuring real-time correctness and effectiveness.

[0092] In summary, this method expands the available range and lifecycle of material copywriting by introducing the design of dynamic factors at different stages of AIGC production, delivery, and rendering, enhancing the personalized expression and timeliness of the copywriting. Compared with static copywriting, it can flexibly provide suitable material descriptions under different times and different product conditions, improving the overall copywriting creation efficiency and material drainage effect.

[0093] Corresponding to the above method embodiments, this specification also provides embodiments of a data processing device, which includes: A variable determination module, configured to determine the initial text corresponding to the target project and determine the target project variable corresponding to the target project, where the target project variable is a variable that changes over time; A model processing module, configured to process the initial text and the target project variable using a data processing model to obtain target data containing the target project variable, where the target data is data generated based on the initial text and the target data is any one of the multimodal data modalities; A data processing module, configured to determine the target project parameters corresponding to the target project variable and generate a data processing result corresponding to the target project based on the target project parameters and the target data, where the target project parameters are the project parameters corresponding to the target project variable at the current time.

[0094] Optionally, the variable determination module is further configured to: Determine the project variable identifier corresponding to the target project, and based on the project variable identifier, obtain the target project variable corresponding to the target project from the variable storage structure; The data processing module is further configured to: Obtain the target project parameters corresponding to the target project variable at the current time from the variable storage structure, where the target project parameters are the parameters corresponding to the target project variable at different time points.

[0095] The data processing device further includes a variable creation module, which is configured to: Create the target project variable in the variable storage structure by using the project variable identifier; Obtain the target project parameters corresponding to the target project variable, and store the target project parameters in the variable storage structure.

[0096] Optionally, the initial text is the initial promotional text, the target data is the target promotional text, and the data processing model is a multi-modal large model; The model processing module is further configured to: Input the initial promotional text and the target project variable into the multi-modal large model, and use the multi-modal large model to generate a promotional text according to the initial promotional text and the target project variable, and obtain the target promotional text containing the target project variable; Optionally, the target data is the target image, and the data processing model is a multi-modal large model; The model processing module is further configured to: Input the initial text and the target project variable into the multi-modal large model, and use the multi-modal large model to generate an image according to the initial text and the target project variable, and obtain the target image containing the target project variable; Optionally, the initial text is the initial text, the target data is the target video, and the data processing model is a multi-modal large model; The model processing module is further configured to: Input the initial text and the target project variable into the multi-modal large model, and use the multi-modal large model to generate a video according to the initial text and the target project variable, and obtain the target video containing the target project variable.

[0097] Optionally, the data processing result is the project marketing data corresponding to the target project; The data processing module is further configured to: Fuse the target project parameters and the target data to generate project marketing data corresponding to the target project.

[0098] Optionally, the target data is a target promotional text, a target image, or a target video, and the target project parameters are target project marketing metrics; The data processing module is further configured to: Use the target project marketing metrics to replace the target project variables included in the target promotional text to obtain an updated promotional text, and determine the updated promotional text as the project marketing data corresponding to the target project; or Use the target project marketing metrics to replace the target project variables included in the target image to obtain an updated image, and determine the updated image as the project marketing data corresponding to the target project; or Use the target project marketing metrics to replace the target project variables included in the target video to obtain an updated video, and determine the updated video as the project marketing data corresponding to the target project.

[0099] Optionally, the data processing device further includes a parameter update module, configured to: Monitor in real time the target project parameters corresponding to the target project variables, and obtain updated project parameters corresponding to the target project variables when it is determined that the target project parameters have been updated; Use the updated project parameters and the target data to generate an updated data processing result corresponding to the target project.

[0100] Optionally, the data processing device further includes a result rendering block, configured to: Render the data processing result to a marketing promotion module so that the marketing promotion module can display the data processing result.

[0101] A data processing device provided by one or more embodiments of this specification can, during the data processing process, determine the initial text that needs to be processed by a data processing model, and the target project variables corresponding to the target project and changing over time; use the data processing model to perform data generation processing on the initial text and the target project variables to obtain target data generated from the initial text and including the target project variables; and finally, use the target project parameters corresponding to the target project variables to process the target data, thereby accurately generating a data processing result corresponding to the target project, thus realizing taking into account the continuously changing variables in the target project during the data processing process, and solving the problem of being unable to accurately obtain a data processing result adapted to the data processing project.

[0102] The above is a schematic solution of a data processing device according to this embodiment. It should be noted that the technical solution of this data processing device and the technical solution of the above data processing method belong to the same concept. For the details not described in detail in the technical solution of the data processing device, reference can be made to the description of the technical solution of the above data processing method.

[0103] Figure 6 FIG. shows a block diagram of a computing device 600 according to an embodiment of the present specification. The components of the computing device 600 include, but are not limited to, a memory 610 and a processor 620. The processor 620 is connected to the memory 610 through a bus 630, and a database 650 is used to store data.

[0104] The computing device 600 further includes an access device 640, which enables the computing device 600 to communicate via one or more networks 660. Examples of these networks include the Public Switched Telephone Network (PSTN), Local Area Network (LAN), Wide Area Network (WAN), Personal Area Network (PAN), or a combination of communication networks such as the Internet. The access device 640 may include one or more of any type of wired or wireless network interfaces (e.g., a network interface card (NIC)), such as an IEEE 802.11 Wireless Local Area Network (WLAN) wireless interface, a Worldwide Interoperability for Microwave Access (Wi-MAX) interface, an Ethernet interface, a Universal Serial Bus (USB) interface, a cellular network interface, a Bluetooth interface, a Near Field Communication (NFC).

[0105] In an embodiment of the present specification, the above components of the computing device 600 and Figure 6 other components not shown in the figure may also be connected to each other, for example, through a bus. It should be understood that Figure 6 the block diagram of the computing device shown is only for illustrative purposes and is not a limitation on the scope of the present specification. Those skilled in the art can add or replace other components as needed.

[0106] The computing device 600 can be any type of stationary or mobile computing device, including mobile computers or mobile computing devices (e.g., tablet computers, personal digital assistants, laptop computers, notebook computers, netbooks, etc.), mobile phones (e.g., smartphones), wearable computing devices (e.g., smartwatches, smart glasses, etc.) or other types of mobile devices, or stationary computing devices such as desktop computers or personal computers (PCs). The computing device 600 can also be a mobile or stationary server.

[0107] The processor 620 is configured to execute computer-executable instructions, which, when executed by the processor, implement the steps of the above-described data processing method.

[0108] Each embodiment in this specification is described in a progressive manner. For the same or similar parts among the embodiments, reference can be made to each other. Each embodiment focuses on the differences from other embodiments. In particular, for the embodiment of the computing device, since it is basically similar to the embodiment of the data processing method, the description is relatively simple. For the relevant parts, reference can be made to the partial description of the embodiment of the data processing method.

[0109] An embodiment of this specification also provides a computer-readable storage medium storing computer programs / instructions, which, when executed by a processor, implement the steps of the above-described data processing method.

[0110] Each embodiment in this specification is described in a progressive manner. For the same or similar parts among the embodiments, reference can be made to each other. Each embodiment focuses on the differences from other embodiments. In particular, for the embodiment of the computer-readable storage medium, since it is basically similar to the embodiment of the data processing method, the description is relatively simple. For the relevant parts, reference can be made to the partial description of the embodiment of the data processing method.

[0111] An embodiment of this specification also provides a computer program product including computer programs / instructions, which, when executed by a processor, implement the steps of the above-described data processing method.

[0112] The above is a schematic solution of a computer program product according to this embodiment. It should be noted that the technical solution of this computer program product and the technical solution of the above-described data processing method belong to the same concept. For the details not described in the technical solution of the computer program product, reference can be made to the description of the technical solution of the above-described data processing method.

[0113] The above describes 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 in the embodiments and still achieve the desired results. Additionally, the processes depicted in the drawings do not necessarily require the particular order or sequential order shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0114] The computer instructions include computer program code, which may be in source code form, object code form, executable file, or some intermediate form, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, removable hard disk, magnetic disk, optical disc, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc. It should be noted that the content included in the computer-readable medium can be appropriately increased or decreased according to the requirements of patent practice. For example, in some regions, according to patent practice, the computer-readable medium does not include electrical carrier signals and telecommunication signals.

[0115] It should be noted that for the foregoing method embodiments, for the sake of simplicity of description, they are all expressed as a series of action combinations. However, those skilled in the art should know that the embodiments of this specification are not limited by the described order of actions, because according to the embodiments of this specification, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to the embodiments of this specification.

[0116] In the above embodiments, the descriptions of the various embodiments have their own focuses. For the parts not detailed in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0117] The preferred embodiments of this specification disclosed above are only used to help explain this specification. The alternative embodiments do not elaborate on all the details and do not limit the invention to only the specific embodiments described. Obviously, many modifications and variations can be made according to the content of the embodiments of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the embodiments of this specification, so that those skilled in the art can well understand and utilize this specification. This specification is only limited by the claims and their full scope and equivalents.

Claims

1. A data processing method, comprising: Determine an initial text corresponding to a target project, and determine a target project variable corresponding to the target project, wherein the target project variable is a variable that changes with time; Processing the initial text and the target project variables using a data processing model to obtain target data containing the target project variables, wherein the target data is data generated based on the initial text, and the target data is any one modality of multimodal data; Determine a target project parameter corresponding to the target project variable, and generate a data processing result corresponding to the target project based on the target project parameter and the target data, wherein the target project parameter is a project parameter corresponding to the target project variable at the current time.

2. The data processing method according to claim 1, wherein determining the target project variable corresponding to the target project comprises: Determine a project variable identifier corresponding to the target project, and based on the project variable identifier, obtain a target project variable corresponding to the target project from a variable storage structure; The determining of the target project parameter corresponding to the target project variable includes: The target project parameter corresponding to the target project variable at the current time is obtained from the variable storage structure, wherein the target project parameter is a parameter corresponding to the target project variable at different time points.

3. The data processing method according to any one of claims 1 to 2, before determining the target project variable corresponding to the target project, further comprising: Using the project variable identifier, creating the target project variable in the variable storage structure; A target project parameter corresponding to the target project variable is obtained, and the target project parameter is stored in the variable storage structure.

4. The data processing method according to claim 1, wherein the initial text is an initial promotional text, the target data is a target promotional text, and the data processing model is a multimodal large model; The using of the data processing model to process the initial text and the target project variable to obtain target data containing the target project variable includes: The initial promotional text and the target project variables are input into the multimodal large model, and the multimodal large model is used to generate a promotional text according to the initial promotional text and the target project variables to obtain a target promotional text containing the target project variables.

5. The data processing method according to claim 1, wherein the target data is a target image, and the data processing model is a multimodal large model; The using of the data processing model to process the initial text and the target project variable to obtain target data containing the target project variable includes: The initial text and the target project variables are input into the multimodal large model, and the multimodal large model is used to generate an image according to the initial text and the target project variables to obtain the target image containing the target project variables.

6. The data processing method according to claim 1, wherein the initial text is an initial text, the target data is a target video, and the data processing model is a multimodal large model; The using of the data processing model to process the initial text and the target project variable to obtain target data containing the target project variable includes: The initial text and the target project variables are input into the multimodal large model, and the multimodal large model is used to generate a video according to the initial text and the target project variables to obtain the target video containing the target project variables.

7. The data processing method according to claim 1, wherein the data processing result is project marketing data corresponding to the target project; The step of generating a data processing result corresponding to the target project based on the target project parameters and the target data includes: The target project parameters and the target data are fused to generate project marketing data corresponding to the target project.

8. The data processing method according to claim 7, wherein the target data is a target promotional text, a target image or a target video, and the target project parameter is a target project marketing indicator; The step of fusing the target project parameters and the target data to generate project marketing data corresponding to the target project includes: Using the target project marketing indicator, the target project variable contained in the target promotion text is replaced to obtain an updated promotion text, and the updated promotion text is determined as the project marketing data corresponding to the target project; or Using the target project marketing indicator, replacing the target project variable contained in the target image to obtain an updated image, and determining the updated image as the project marketing data corresponding to the target project; or The target project marketing indicator is used to replace the target project variables contained in the target video to obtain an updated video, and the updated video is determined as the project marketing data corresponding to the target project.

9. The data processing method according to claim 1, after generating the data processing result corresponding to the target project based on the target project parameters and the target data, further comprising: monitoring the target project parameters corresponding to the target project variables in real time, and obtaining updated project parameters corresponding to the target project variables when it is determined that the target project parameters are updated; The update data processing result corresponding to the target project is generated by using the update project parameters and the target data.

10. The data processing method according to claim 1, after generating the data processing result corresponding to the target project based on the target project parameters and the target data, further comprising: The data processing result is rendered to a marketing promotion module so that the marketing promotion module displays the data processing result.

11. A data processing device, comprising: A variable determination module is configured to determine an initial text corresponding to a target project, and determine a target project variable corresponding to the target project, wherein the target project variable is a variable that changes with time; A model processing module is configured to process the initial text and the target project variables using a data processing model to obtain target data containing the target project variables, wherein the target data is data generated based on the initial text, and the target data is any one modality of multimodal data; The data processing module is configured to determine the target project parameters corresponding to the target project variables, and generate data processing results corresponding to the target project based on the target project parameters and the target data, wherein the target project parameters are project parameters corresponding to the target project variables at the current time.

12. A computing device comprising: Memory and processor; The memory is used to store computer programs / instructions, and the processor is used to execute the computer programs / instructions. When the computer programs / instructions are executed by the processor, the steps of the method described in any one of claims 1 to 10 are implemented.

13. A computer-readable storage medium storing a computer program / instruction, wherein the computer program / instruction, when executed by a processor, implements the steps of the method according to any one of claims 1 to 10.

14. A computer program product, comprising a computer program / instruction, which, when executed by a processor, implements the steps of the method according to any one of claims 1 to 10.

Images