Content publishing method and device

Abstract

The embodiment of the invention provides a content publishing method and device, and the method comprises the steps: obtaining the project marketing information and the project reference content of a target project; project marketing information and the project reference content are input into a content generation model, to-be-published content is obtained, and the project marketing information is used for constraining the content generation direction of the target project; and publishing the to-be-published content to the content association area of the project reference content. The method can effectively ensure that the generated content to be published not only accords with a core marketing strategy of a target project, but also has an expression form and a context which are highly matched with the reference content of the project, so that the correlation, the consistency and the conversion effect of the content to be published are remarkably improved; and meanwhile, scene matching and efficient touch of content delivery are realized, the overall collaboration and intelligent level of project marketing propagation are enhanced, and the user participation degree and the project promotion efficiency are further improved.

Description

TECHNICAL FIELD

[0001] Embodiments of the present specification relate to the technical field of content publishing, and in particular to a content publishing method and device. BACKGROUND

[0002] With the rapid development of digital marketing, enterprises increasingly rely on publishing high-quality and scenario-based content in social media, official websites, content platforms, and other channels to attract potential users when promoting various projects (such as real estate, insurance products, and travel activities).

[0003] Currently, most projects still use manual methods for content writing and publishing: marketing personnel manually write scripts based on project materials and then manually upload them to designated locations on various content platforms. However, manual writing and publishing are inefficient, and the quality of the content is highly dependent on individual experience, which restricts the improvement of overall marketing effectiveness. SUMMARY

[0004] Therefore, embodiments of the present specification provide a content publishing method. One or more embodiments of the present specification also relate to a content publishing device, a computing device, a computer-readable storage medium, and a computer program product to solve the technical defects in the prior art.

[0005] According to a first aspect of embodiments of the present specification, a content publishing method is provided, comprising: obtaining project marketing information and project reference content of a target project; inputting the project marketing information and the project reference content into a content generation model to obtain to-be-published content, wherein the project marketing information is used to constrain the content generation direction of the target project; publishing the to-be-published content to a content association area of the project reference content.

[0006] According to a second aspect of embodiments of the present specification, a content publishing device is provided, comprising: an obtaining module configured to obtain project marketing information and project reference content of a target project; a generating module configured to input the project marketing information and the project reference content into a content generation model to obtain to-be-published content, wherein the project marketing information is used to constrain the content generation direction of the target project; a publishing module configured to publish the to-be-published content to a content association area of the project reference content.

[0007] According to a third aspect of embodiments of the present specification, a computing device is provided, comprising: 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, which, when executed by the processor, implement the steps of the method provided in the first aspect above.

[0008] According to a fourth aspect of the embodiments of this specification, a computer-readable storage medium is provided that stores a computer program / instructions that, when executed by a processor, implement the steps of the method provided in the first aspect described above.

[0009] According to a fifth aspect of the embodiments of this specification, a computer program product is provided, including a computer program / instructions that, when executed by a processor, implement the steps of the method provided in the first aspect described above.

[0010] This specification provides a content publishing method according to one embodiment, which involves obtaining project marketing information and project reference content for a target project; inputting the project marketing information and project reference content into a content generation model to obtain content to be published, wherein the project marketing information is used to constrain the content generation direction of the target project; and publishing the content to be published to the content association area of ​​the project reference content. By inputting the target project's marketing information and project reference content together into the content generation model, it is effectively ensured that the generated content to be published not only conforms to the core marketing strategy of the target project but also has an expression form and context that highly matches the project reference content, thereby significantly improving the relevance, consistency, and conversion effect of the content to be published. At the same time, accurately publishing the content to be published to the content association area achieves scenario-based matching and efficient reach of content delivery, enhances the overall synergy and intelligence level of project marketing communication, and further improves user engagement and project promotion efficiency. Attached Figure Description

[0011] Figure 1 This is a flowchart illustrating a content publishing method provided in one embodiment of this specification; Figure 2 This is a flowchart illustrating the processing procedure of a note-commenting method provided in one embodiment of this specification; Figure 3 This is an architecture diagram of a content publishing system provided in one embodiment of this specification; Figure 4 This is a schematic diagram of the structure of a content publishing device provided in one embodiment of this specification; Figure 5 This is a structural block diagram of a computing device provided in one embodiment of this specification. Detailed Implementation

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

[0013] The terminology used in one or more embodiments of this specification is for the purpose of describing particular embodiments only and is not intended to limit the scope of one or more embodiments of this specification. The singular forms “a,” “described,” and “the” as 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” as used in one or more embodiments of this specification refers to and includes any or all possible combinations of one or more associated listed items. The term “at least one” as used in one or more embodiments of this specification means “one or more,” and “a plurality of” means “two or more.” The term “comprising” is an open-ended description and should be understood as “including but not limiting,” and may include other content in addition to what has been described.

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

[0015] Furthermore, 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 used for analysis, stored data, displayed data, etc.) involved in one or more embodiments of this specification are all information and data authorized by the user or fully authorized by all parties. Moreover, the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and corresponding operation entry points are provided for users to choose to authorize or refuse.

[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, typically containing hundreds of millions, tens of billions, hundreds of billions, trillions, or even tens of trillions of model parameters. A large model can also be called a foundation model. It is pre-trained using large-scale unlabeled corpora to produce a pre-trained model with hundreds of millions of parameters. Such models can adapt to a wide range of downstream tasks and have good generalization ability. Examples include Large Language Models (LLMs) and multi-modal pre-training models.

[0017] In practical applications, large models only require a small number of samples to fine-tune the pre-trained model before they can be applied to different tasks. Large models can be widely used in fields such as Natural Language Processing (NLP) and Computer Vision. Specifically, they can be applied to computer vision tasks such as Visual Question Answering (VQA), Image Captioning (IC), and Image Generation, as well as natural language processing tasks such as text-based sentiment classification, text summarization, and machine translation. The main application scenarios of large models include digital assistants, intelligent robots, search, online education, office software, e-commerce, and intelligent design.

[0018] First, the terms and concepts used in one or more embodiments of this specification will be explained.

[0019] Hypertext Transfer Protocol (HTTP): A very fundamental and crucial rule of the Internet, responsible for transmitting information on the World Wide Web. When you enter a URL in the browser's address bar and press Enter, the browser uses the HTTP protocol to "request" that webpage from the server; after the server understands the request, it then uses the HTTP protocol to send the "content" of the webpage back to the browser.

[0020] Web crawler: refers to writing programs to simulate a browser sending HTTP requests to a platform's topic search page, and then parsing and extracting the required content information from the returned HTML webpage source code.

[0021] Robotic Process Automation (RPA) is an application software technology that uses configured "software robots" to simulate human-computer interaction, automatically executing numerous, repetitive, rule-based business processes. For example, opening a browser, entering a website address, clicking the "login" button, entering a phone number in the username box, entering a password in the password box, and clicking "login" again to submit; if a verification code or QR code scanning page pops up, it can even simulate waiting for the scan (or use other tools to handle this). All of this is done automatically by the program, without requiring a human to sit in front of the computer and manually click. The RPA tool controls the mouse and keyboard, allowing the user to operate the interface just like a human.

[0022] Login Cookie: After you successfully log in to the platform, the server will say, "Okay, you are a legitimate user," and then it will send an encrypted identity token to your browser. This token is stored in a cookie. The purpose of the cookie is that every time the browser visits the platform, it automatically sends this cookie. The platform server sees it, "Oh, I recognize this cookie, it belongs to a logged-in user!" and allows access directly without needing to enter the username and password again.

[0023] Third-party tools: Software or online services developed by third parties that can operate independently to complete a specific task or solve a problem. They may collaborate with other software but do not depend on their internal operation.

[0024] Using content publishing platforms to generate buzz for online insurance products is a crucial method for expanding customer acquisition and boosting premiums, both now and in the future. However, currently, this requires manual effort in the comment sections of content publishing platforms, involving significant manpower to screen and comment on notes. This process is tedious, lacks consistency, and is inefficient. Therefore, it is necessary to automate the process by searching for notes, generating comments, reviewing and revising those comments, and automatically generating buzz in the comment sections of content publishing platforms to attract customers and increase premium income.

[0025] In the relevant technical solutions, people manually search for corresponding note content based on specific search keywords and post comments to recommend products or services.

[0026] Humans need to search for relevant notes based on specific search keywords and then comment on them. This is limited by factors such as individual professional knowledge and working hours. Furthermore, when the latest notes are published, it is not possible to monitor them in time and comment on them, which takes a lot of time. In the future, when the number of notes to be commented on is in the thousands every day, the cost of human intervention will be huge, the efficiency will be extremely low, and the time consumption will be very long.

[0027] To address the shortcomings of existing pet note comments, this embodiment of the manual uses RPA to automatically log into a platform account, search for the latest pet notes based on specific search keywords, categorize and tag them, find pet notes suitable for recommendation, then use a large model combined with the selling points of pet insurance products to generate high-quality, anthropomorphic comments on the note content, and automatically review and revise the generated comments using a large model combined with expert experience-based review rules, and finally automatically publish the comments.

[0028] Based on this, this specification provides a content publishing method, and also relates to a content publishing device, a computing device, a computer-readable storage medium, and a computer program product, which will be described in detail in the following embodiments.

[0029] See Figure 1 , Figure 1 This specification illustrates a flowchart of a content publishing method according to an embodiment, which specifically includes the following steps: Step 102: Obtain the target project's marketing information and reference content.

[0030] It should be noted that the target project refers to the specific object that needs to be marketed and promoted. Target projects can be promotional projects in different industries, such as pet insurance in the insurance industry, or wealth management products in the financial industry.

[0031] Project marketing messages refer to a set of consistent and persuasive core communication messages carefully designed and delivered for a target project. Their fundamental purpose is to clearly articulate the project's value proposition to the target audience, address their pain points, and inspire them to take the expected action. For example, if the target project is pet insurance, the project marketing message could be the selling points of pet insurance.

[0032] Project reference content refers to content related to the target project that assists the content generation model in generating content to be published. The content generation model can use project reference content as a basis for creation, outputting published content related to that content. For example, project reference content can be a summary text, which the content generation model can expand to obtain published content; project reference content can also be a content requirement, which the model can use to generate published content; project reference content can also be notes / videos on a platform related to the project topic, which the model can comment on to obtain published content.

[0033] Taking note-taking as an example, notes refer to text, images, or other forms of content created by individuals or teams in their studies, work, or daily lives for purposes such as recording information, organizing thoughts, and providing reminders. Notes can be as simple as a few lines of text or as complex electronic documents containing various elements such as charts, links, and multimedia files. Notes are widely used in various life and work scenarios, such as knowledge point notes and lesson plan notes in learning and education, and dietary records, exercise records, and shopping sharing notes in health and lifestyle scenarios.

[0034] In practical applications, there are various ways to obtain project marketing information and project reference content for a target project. The specific method chosen depends on the actual situation, and this specification does not impose any limitations on these methods. In one possible implementation, the target project's marketing information and project reference content can be received from a user via a client. In another possible implementation, the target project's marketing information and project reference content can be read from other databases or data acquisition devices.

[0035] In another possible implementation of this specification, obtaining the target project's marketing information and reference content may include the following steps: Obtain project marketing information and project-related topics for the target project; Based on the project topic, retrieve reference content for the project from the content publishing platform.

[0036] It's important to note that "project topic" refers to search keywords or query commands used for content retrieval, precisely filtering and locating relevant project reference content from content publishing platforms. Project topics are related to the target project. For example, if the target project is pet insurance, project topics could include: how to buy pet insurance, whether pet medical insurance is necessary, what a first-time cat / dog owner needs to prepare, how to raise a cat / dog, is it advisable for college students to own a cat, and adopting a kitten / puppy.

[0037] A content publishing platform is a technological system or online service that allows users to create, edit, manage, and publicly or selectively publish digital content (such as text, images, videos, and audio) on the internet for others to view, interact with, and share. Simply put, it's a bridge and stage connecting content creators and content consumers. For example, a content publishing platform could be a short video platform or a note-taking platform.

[0038] In practical applications, project topics can be obtained either by receiving project topics from users via a client or by reading project topics from other databases or data acquisition devices.

[0039] Depending on the project topic, there are multiple ways to retrieve project reference content from the content publishing platform. The specific method chosen depends on the actual situation, and this specification does not impose any limitations on this approach. One possible implementation involves obtaining the account and password for the content publishing platform, logging into the platform using RPA technology based on the account and password, and retrieving the project reference content from the platform.

[0040] In another possible implementation of this specification, web crawling technology is used to retrieve project reference content from the content publishing platform based on the project topic.

[0041] In another possible implementation of this specification, retrieving project reference content from the content publishing platform based on the project topic may include the following steps: Based on the login credentials of the content publishing platform, retrieve the content index corresponding to the project topic from the content publishing platform; Based on content indexing, retrieve reference content for the project.

[0042] It's important to note that login credentials are small data fragments sent by the content publishing platform to the web browser after a user successfully logs in with their username and password. Their core function is to inform the content publishing platform, "You are already a verified user and do not need to log in again." For example, login credentials might be a login cookie. Without cookies, the content publishing platform cannot remember the user's login status. Having to re-enter the password every time a user clicks a new link or opens a page would result in an extremely poor user experience.

[0043] A content index is an address credential or digital pointer used to uniquely identify and locate referenced content within a project. It does not contain any actual content information (such as title, text, images, etc.), but rather acts like a precise "address strip" or a library's "call number," enabling the system to accurately and efficiently retrieve the corresponding complete content from a vast content repository. For example, a content index can be a link to or an identity document (ID) for referenced content within a project.

[0044] In practical applications, there are multiple ways to retrieve the content index corresponding to a project topic from a content publishing platform based on the platform's login credentials. The specific method chosen depends on the actual situation, and this specification does not impose any limitations on this approach. One possible implementation of this specification involves retrieving the content index corresponding to a project topic from the content publishing platform using a web crawler based on the platform's login credentials. Another possible implementation of this specification involves retrieving the content index corresponding to a project topic from the content publishing platform using RPA technology based on the platform's login credentials.

[0045] There are multiple ways to obtain project reference content based on content indexing, and the specific method chosen depends on the actual situation. This specification does not impose any limitations on these methods in the embodiments. One possible implementation in this specification involves calling a third-party tool to obtain project reference content based on content indexing. Another possible implementation in this specification involves obtaining project reference content based on content indexing using RPA technology.

[0046] For example, Table 1 shows the note links and note IDs corresponding to the search keywords obtained through RPA technology. As shown in Table 1, searching for the keyword "how to buy pet insurance" on the content publishing platform using RPA technology yields multiple note links (such as URL 1, URL 2, URL 3, ...) and multiple note IDs (such as 67f4c7d5000, 67ee9242000, 67f3a233000, ...).

[0047] Table 1. A diagram of note links and note IDs.

[0048] The solution implemented in the embodiments of this specification enables efficient and automated retrieval of content indexes based on login credentials, thereby improving retrieval efficiency. By first retrieving the content index and then obtaining the project reference content, network transmission load can be reduced, the pressure on the target platform can be lowered, and the processing can be made more flexible.

[0049] In another possible implementation of this specification, retrieving project reference content from the content publishing platform based on the project topic may include the following steps: Based on the project topic, candidate reference content is retrieved from the content publishing platform. The candidate reference content is content related to the project topic on the content publishing platform. Identify the categories of candidate reference content and obtain the candidate content categories, where the candidate content categories are used to reflect the publishing scenario corresponding to the candidate reference content; Based on the candidate content category, project reference content is selected from the candidate reference content, where the candidate content category corresponding to the project reference content matches the release scenario of the target project.

[0050] It should be noted that candidate reference content refers to a collection of original content related to or potentially related to the project topic, obtained through preliminary retrieval from content publishing platforms (such as social media, news websites, blogs, forums, video platforms, etc.). For example, candidate reference content might consist of multiple notes retrieved based on the topic "how to buy pet insurance".

[0051] Candidate content categories refer to tags or classifications that categorize candidate reference content according to its inherent attributes, format, or publishing context. Different candidate reference content can correspond to different candidate content categories. For example, candidate content categories could be beginner pet ownership, purebred / domestic cats and dogs, student groups, free adoption posts, and pet insurance related content.

[0052] A publishing context refers to the specific environment, situation, and purpose within which the target content is planned to be published and consumed. It defines the content's "living environment" and determines what kind of content is effective and appropriate within it. Each candidate reference content corresponds to a publishing context, which can be reflected through candidate content categories.

[0053] In practical applications, there are various ways to retrieve candidate reference content from a content publishing platform based on the project topic. The specific method chosen depends on the actual situation, and this specification does not impose any limitations on this approach. One possible implementation in this specification involves using RPA technology to search for the project topic on the content publishing platform's search page to obtain candidate reference content.

[0054] In another possible implementation of this specification, web crawling technology is used to retrieve candidate reference content from the content publishing platform based on the project topic.

[0055] There are multiple ways to identify the category of candidate reference content and obtain the candidate content category. The specific method should be selected according to the actual situation, and the embodiments in this specification do not limit this. In one possible implementation of this specification, the candidate reference content is input into the category recognition model to obtain the candidate content category. The category recognition model is trained on an initial category recognition model based on the sample reference content and the sample content category.

[0056] In another possible implementation of this specification, the candidate reference content is parsed to obtain the title and interactive text of the candidate reference content. Based on the title and interactive text, the category of the candidate reference content is identified to obtain the candidate content category.

[0057] In another possible implementation of this specification, identifying the category of candidate reference content and obtaining the candidate content category may include the following steps: Input the candidate reference content into the content parsing model to obtain the title and interactive text of the candidate reference content; Based on the title and interactive text, the categories of candidate reference content are identified to obtain the candidate content categories.

[0058] It's important to note that a content parsing model refers to a trained machine learning or deep learning model used to extract titles and interactive text from candidate reference content. Content parsing models are obtained through supervised fine-tuning based on sample reference content and its corresponding sample titles and interactive text, exhibiting good generalization ability. For example, a content parsing model can be a large-scale multimodal model, taking candidate reference content and prompts (used to guide the model in parsing the candidate reference content and outputting titles and interactive text) as input, and outputting the titles and interactive text of the candidate reference content. Alternatively, a content parsing model can be a traditional recognition model, trained based on sample reference content and its corresponding sample titles and interactive text.

[0059] A title is a concise and summarized text in the candidate reference content, designed to attract user attention and highlight the core theme.

[0060] Interactive text refers to the core content of candidate reference content, excluding the title. It is the main text created by the publisher to convey core information, explain viewpoints, tell stories, or showcase content.

[0061] In practical applications, there are various ways to identify the category of candidate reference content based on the title and interactive text. The specific method chosen depends on the actual situation, and this specification does not limit this approach. In one possible implementation, each candidate content category corresponds to a keyword library. The title and interactive text are matched against multiple keyword libraries, and the candidate content category is determined according to a judgment rule. For example, if more than three keywords related to the category "pet insurance" appear simultaneously in the title and interactive text, then the category of the candidate reference content is "pet insurance related".

[0062] In another possible implementation of this specification, the title, interactive text, and classification prompts are input into the category recognition model to obtain candidate content categories. The category recognition model is obtained through supervised fine-tuning based on the sample title, sample interactive content, and sample content category.

[0063] For example, the category suggestion information could be: "You are a semantic understanding expert. Based on the pet notes I provided, you need to deeply understand the content of the notes and help me categorize the notes:" 1. Beginner pet owner: This refers to someone who is just starting to own a pet, and must already have a pet or be planning to start owning a cat or dog. If you do not own a pet, you cannot be considered a beginner pet owner.

[0064] 2. Purebred cats / dogs: The content of the notes must be about a specific breed of cat or dog, but there must be no content about paid adoption or free adoption.

[0065] 3. College Students' Pet Ownership: The notes are about college students keeping cats or dogs.

[0066] 4. Free adoption: This means that the adopter can take the cat or dog home and raise it without paying any money or valuables.

[0067] 5. Pet insurance related: This refers to purchasing relevant insurance for cats or dogs.

[0068] 6. Claims related: This refers to the claims process when a cat or dog becomes ill after being insured.

[0069] 7. Other animals: refers to animals other than cats and dogs, such as birds.

[0070] 8. Popular Science Notes: Popular science notes include information about pets or adoption.

[0071] 9. Advertising notes: The notes repeatedly mention the specific name of a pet insurance policy, the insurance company, or product keywords (such as "pet insurance" or "pet protection") or a certain brand of product, such as a certain brand of cat food or dog food, and even directly use explicit guiding words such as "recommended", "essential", and "personally tested and effective".

[0072] The article devotes a large portion to describing the insurance coverage, claims processing speed, and price discounts (such as "only 200 yuan per month" and "limited-time discount"), rather than sharing actual daily experiences of using pets.

[0073] Compared to other similar products, highlight the product's "exclusive advantages" or imply the shortcomings of other insurance policies.

[0074] The notes contain direct links, QR codes, discount codes, or prompts such as "DM for guide" or "Comments section has benefits" to attract users. They use sales copy-like wording, such as "worry-free and cost-effective," "one-click insurance," "lowest price online," "must-read for pet owners," or excessively exaggerate the risks.

[0075] If it's a comparison and recommendation of products from multiple insurance companies, then it's not considered an advertisement. The note contains promotional content.

[0076] 10. Paid adoption: This refers to an adoption where the adopter needs to pay any money or valuables to take the cat or dog home and raise it.

[0077] 11. Showing off cats or dogs: This refers to posting pictures of your own cats or dogs that are not of a purebred breed, i.e., ordinary cats or dogs.

[0078] Requirements: No other textual explanation is needed; just output the category, such as "free adoption" or "pet ownership for beginners".

[0079] The output category must be in the following list: ["Beginner Pet Ownership", "Purebred Cats / Dogs", "College Student Pet Ownership", "Show Off Your Cat or Dog", "Free Adoption", "Pet Insurance Related", "Claims Related", "Other Animals", "Science Notes", "Advertising Notes", "Paid Adoption"].

[0080] If none of the above categories apply, you need to create a category that you deem appropriate based on your own understanding of the notes' content.

[0081] For example, Table 2 shows the extracted note titles and note content. As shown in Table 2, the search keyword "how to buy pet insurance" yields multiple note links (e.g., URL 1, URL 2, URL 3, ...), multiple note IDs (e.g., 67f4c7d5000, 67ee9242000, 67f3a233000, ...), multiple note titles (e.g., Title 1, Title 2, Title 3, ...), and multiple note contents (e.g., Content 1, Content 2, Content 3, ...).

[0082] Table 2. A diagram of note titles and note content.

[0083] Table 3 shows the categorization results of the searched notes. As shown in Table 3, the search keyword was "how to buy pet insurance," and the following results were extracted: multiple note links (e.g., URL 1, URL 2, URL 3, ...), multiple note IDs (e.g., 67f4c7d5000, 67ee9242000, 67f3a233000, ...), multiple note titles (e.g., Title 1, Title 2, Title 3, ...), multiple note contents (e.g., Content 1, Content 2, Content 3, ...), multiple images and videos (e.g., Image 1, Image 2, Image 3, ...), and multiple... The following data can be used to select authors (e.g., author 1, author 2, author 3, ...), multiple posting times (e.g., posting time 1, posting time 2, posting time 3, ...), multiple likes (e.g., 2, 0, 6, ...), multiple favorites (e.g., 2, 0, 0, ...), multiple comments (e.g., 0, 1, 0, ...), multiple comment contents (e.g., empty, comment 2, empty, ...), multiple new comment times (e.g., empty, comment time 2, empty, ...), and multiple category tags (e.g., advertisement, pet insurance related, pet insurance related).

[0084] Table 3. Illustration of Note Classification Results

[0085] By applying the solutions in the embodiments of this specification, the title and interactive text are first parsed to determine the candidate content categories, which can focus on core information, improve classification accuracy and robustness, achieve data dimensionality reduction, optimize processing efficiency and system performance, and enhance flexibility and scalability.

[0086] In practical applications, there are various ways to select project reference content from candidate reference content based on candidate content categories. The specific method chosen depends on the actual situation, and the embodiments in this specification do not impose any limitations on this. In one possible implementation of this specification, based on the publishing scenario of the target project, the target content category is determined from the candidate content categories, and the project reference content corresponding to the target content category is selected from the candidate reference content based on the target content category.

[0087] In another possible implementation of this specification, the candidate content categories include advertising categories and non-advertising categories; based on the candidate content categories, selecting project reference content from the candidate reference content may include the following steps: Candidate reference content that is not in the advertising category will be identified as project reference content; and / or, From the candidate reference content, remove the content to be deleted to obtain the project reference content, where the content to be deleted belongs to the advertising category.

[0088] It should be noted that the advertising category refers to the core purpose of the corresponding content to promote products, services, brands, or ideas, and to encourage users to take specific actions such as purchasing, registering, following, or agreeing with the content.

[0089] Non-advertising categories refer to content whose core purpose is to provide information, share knowledge, express opinions, tell stories, or provide entertainment. Their primary goal is not sales, but rather communication, education, inspiration, or entertainment. For example, for pet notes, non-advertising categories could include beginner pet owners, purebred / domestic cats and dogs, student groups, free adoption posts, and pet insurance related content.

[0090] Content to be deleted refers to specific content identified from the initial search of candidate reference content and required to be filtered out according to preset filtering rules (here, the rule is "belongs to the advertising category"). For example, for pet notes, pet notes that belong to advertising posts would be considered content to be deleted.

[0091] By applying the solutions in the embodiments of this specification, the relevance and credibility of content can be improved by identifying and filtering candidate reference content for advertising categories; users do not need to perform manual filtering, thus improving the efficiency of information processing and user experience.

[0092] The solution implemented in this specification retrieves candidate reference content based on project topics, ensuring the breadth of information; through classification, identification, and filtering, it achieves in-depth information extraction, systematically eliminates irrelevant content, and ensures that the final output content is highly relevant.

[0093] The solution implemented in this specification allows for the retrieval of project reference content through project topic searches, eliminating the need for complex operational procedures, simplifying system operations, improving search efficiency, and supporting batch processing.

[0094] Step 104: Input the project marketing information and project reference content into the content generation model to obtain the content to be published. The project marketing information is used to constrain the content generation direction of the target project.

[0095] It's important to note that a content generation model refers to a trained machine learning or deep learning model used to generate content for publication. Content generation models are obtained through supervised fine-tuning based on sample marketing information, sample reference content, and their corresponding published content, exhibiting good generalization ability. For example, a content generation model can be a large-scale multimodal model, taking project marketing information, project reference content, and prompts (used to guide the model in generating published content based on the project marketing information and project reference content), and outputting the published content. Content generation models can also be traditional models, trained based on sample marketing information, sample reference content, and their corresponding published content.

[0096] Content to be published refers to content customized by the content generation model based on project marketing information and project reference content, intended for publication in a specific region. The project marketing information constrains the content generation direction of the content to be published, while the project reference content provides the creative basis for the content.

[0097] Content generation direction refers to the core objectives, boundaries, and stylistic framework set for the content creation process when utilizing artificial intelligence or human expertise. It is a set of guiding principles that ensure the final content aligns with pre-defined marketing goals, brand tone, and audience needs. Its primary function is to constrain and guide, preventing the content generation process from deviating from the theme or generating irrelevant or invalid information. It answers the fundamental question of "what kind of content to generate," and is crucial for ensuring content quality and relevance. When generating content to be published based on project reference content, the content generation model needs to generate the content according to the content generation direction corresponding to the project's marketing information.

[0098] For example, the project's marketing information focuses on the selling points of a pet insurance product, while the project's reference content is pet notes. The content to be published can be comments on these pet notes. These comments should consider not only the content of the pet notes but also the selling points of the pet insurance product, integrating these selling points into the comments. When users search for and view pet notes, they can glean the selling points of the pet insurance product from the comments, increasing the awareness of pet insurance and expanding traffic.

[0099] Table 4 shows the comments to be published for the search notes. As shown in Table 4, the search keyword is "how to buy pet insurance". The results retrieved multiple note links (e.g., URL 1, URL 2, URL 3, ...), multiple note IDs (e.g., 67f4c7d5000, 67ee9242000, 67f3a233000, ...), multiple note titles (e.g., Title 1, Title 2, Title 3, ...), multiple note contents (e.g., Content 1, Content 2, Content 3, ...), multiple images and videos (e.g., Image 1, Image 2, Image 3, ...), and multiple authors (e.g., Author 1, Author 2, Author 3, ...). The following are considered as separate categories: multiple post times (e.g., post time 1, post time 2, post time 3, ...), multiple likes (e.g., 2, 0, 6, ...), multiple favorites (e.g., 2, 0, 0, ...), multiple comments (e.g., 0, 1, 0, ...), multiple comment contents (e.g., empty, comment 2, empty, ...), multiple new comment times (e.g., empty, comment time 2, empty, ...), multiple category tags (e.g., advertisement, pet insurance related, pet insurance related), and multiple pending comments (e.g., empty, pending comment 2, pending comment 3).

[0100] Table 4. A list of comments to be published

[0101] In practical applications, there are multiple ways to input project marketing information and project reference content into the content generation model to obtain the content to be published. The specific method chosen depends on the actual situation, and this specification does not impose any limitations on this approach. In one possible implementation, the reference content category of the project reference content is obtained, and the reference content category, project marketing information, and project reference content are input into the content generation model to obtain the content to be published. The content generation model is trained on an initial generation model based on sample content categories, sample marketing information, sample reference content, and the corresponding sample published content.

[0102] In another possible implementation of this specification, inputting project marketing information and project reference content into the content generation model to obtain the content to be published may include the following steps: The content generation model is input with the generated prompt information, project marketing information, and project reference content to obtain the content to be published. The generated prompt information is used to guide the content generation model to generate the content to be published based on the project marketing information and project reference content.

[0103] It's important to note that the generation prompt is a pre-designed, guiding text instruction. It doesn't directly provide content materials, but rather tells the content generation model: what its role is (who it plays), what its specific task is (what it needs to do), how it should utilize the "project marketing information" and "project reference content" (how to do it), and what format and style requirements the final output "content to be published" must meet. Through precise instructions, the content generation model's process is ensured to be controllable, targeted, and efficient.

[0104] The generated prompt message can be a fixed piece of content, or it can be adjustable content related to the reference content category of the project's reference content.

[0105] The solution implemented in the embodiments of this specification generates prompts that clearly define the specific objectives of the task. These prompts act as anchors, ensuring that the output content serves the project's marketing goals and preventing invalid or off-target content.

[0106] In one optional embodiment of this specification, before obtaining the content to be published by generating input content such as prompt information, project marketing information, and project reference content into the model, the following steps may also be included: Obtain the reference content categories for the project; Based on the reference content category, generate prompt information.

[0107] It should be noted that reference content categories refer to tags or classifications that categorize project reference content according to its inherent attributes, format, or publishing context. Different projects may correspond to different reference content categories. For example, reference content categories could be beginner pet ownership, purebred / domestic cats and dogs, student groups, free adoption posts, and pet insurance related content.

[0108] In practical applications, there are various ways to obtain the reference content category of project reference content, and the specific method should be selected according to the actual situation. This specification does not impose any limitations on this method in its embodiments. In one possible implementation of this specification, the reference content category of the project reference content sent by the user through a client can be received. In another possible implementation of this specification, after filtering the project reference content from candidate reference content based on candidate content categories, the candidate content category corresponding to the project reference content can be used as the reference content category.

[0109] There are multiple ways to construct and generate prompt information based on the reference content category, and the specific method chosen depends on the actual situation. This specification does not impose any limitations on this approach in the embodiments. In one possible implementation, a mapping table between reference content categories and generated prompt information is pre-built, and the corresponding generated prompt information is retrieved from the mapping table based on the reference content category. In another possible implementation, construction rules for generating prompt information are obtained, and the generated prompt information is constructed based on the reference content category and the construction rules. The construction rules specify how to construct the generated prompt information based on the reference content category.

[0110] For example, the generated prompt message could be: "Character: You are an expert in writing notes and comments, you enjoy browsing platforms, and you like to share and discuss pet-related issues on a daily basis."

[0111] Objective: Based on the content described in the notes, integrate the pet health insurance product. Write a comment that naturally conveys the importance of pet health insurance to users, based on the product's selling points and the context of the notes.

[0112] Writing style: [Specific writing style details].

[0113] Skills: Skill 1, analyze notes; Skill 2, write comments, make accurate judgments on the content of the notes based on rich pet-raising experience, and infer the possible symptoms of the pet; Skill 3, avoid using related words and never use absolute terms such as best, best, most profitable, optimal, most outstanding, best, largest, maximum degree, highest, lowest, cheapest, newest, most advanced, superb, awesome.

[0114] Product selling points: [Specific product selling points].

[0115] Specific requirements: [Specific requirements details].

[0116] For example, the generated prompt message can also be: "Character: You are an expert in writing notes and comments, you enjoy browsing platforms, and you like to share and discuss pet-related issues on a daily basis."

[0117] Objective: [Specific objective details].

[0118] Writing style: [Specific writing style details].

[0119] Scene: Scenario 1: The notes are about finding adopters (the author posts pet adoption information). The central point of these notes is to help the pet find a responsible new owner.

[0120] Writing direction: From the perspective of reviewing adoptive parents.

[0121] Scenario 2: The notes are about wanting to adopt a pet. These notes often express the author's love for pets and their desire to adopt one.

[0122] Writing direction: Praise the author and acknowledge their love and adoption of pets.

[0123] Skills: Skill 1, analyze the notes; Skill 2, write comments; Skill 3, avoid using related words and absolute descriptive adjectives.

[0124] Product selling points: [Specific product selling points].

[0125] Specific requirements: Comments must conform to the prescribed format.

[0126] By applying the solutions in the embodiments of this specification, based on the reference content category, the generated prompt information is constructed, which achieves a perfect integration of the generated prompt information with the context of the reference content, thereby enhancing the relevance and contextualization of the generated prompt information.

[0127] Step 106: Publish the content to be published to the content association area of ​​the project reference content.

[0128] It should be noted that a content-related area refers to a specific location or section that has a strong contextual connection to the project's reference content and allows users to interact or supplement information. For example, a content-related area could be the answer area for the content requirements, or it could be the comment area for the notes.

[0129] In practical applications, there are various ways to publish content to be published to the content association area of ​​the project reference content. The specific method chosen depends on the actual situation, and this specification does not limit this approach. In one possible implementation, RPA technology is used to locate the content association area of ​​the project reference content and publish the content to be published to that area.

[0130] In one optional embodiment of this specification, before publishing the content to be published to the content association area of ​​the project reference content, the following steps may also be included: The content to be published is evaluated using quality assessment standards and / or quality assessment models to obtain assessment indicators, which are used to reflect the quality of the content to be published. Based on the evaluation metrics, the target content to be published is determined; Publishing content to be published to the content association area of ​​the project reference content may include the following steps: Publish the target content to the content association area of ​​the project reference content.

[0131] It should be noted that quality assessment standards refer to a set of predefined, measurable criteria and parameters used to objectively evaluate whether content meets business needs, user experience, and platform specifications before or after publication. For example, quality assessment standards may include commonly used phrases in the content to be published, as well as prohibited words.

[0132] A quality assessment model is a specially trained artificial intelligence model whose core task is to automatically and quantitatively score and evaluate a piece of text (such as content to be published) to determine whether it meets preset quality assessment standards. The quality assessment model is obtained through supervised fine-tuning based on sample published content, optional sample assessment standards, and sample assessment metrics, and possesses good generalization ability. For example, a quality assessment model can be a large-scale multimodal model, with the input being the content to be published, optional quality assessment standards, and prompts (used to guide the quality assessment model in evaluating the content to be published and outputting assessment metrics), and the output being the assessment metrics. A quality assessment model can also be a traditional assessment model, trained based on sample published content, optional sample assessment standards, and sample assessment metrics.

[0133] Evaluation metrics are a series of quantifiable data points reflecting various aspects of the quality of content after it has been evaluated using quality assessment standards and / or models. Simply put, it transforms the abstract concept of "good" or "bad" content into concrete, measurable, and comparable "data scores" or "grading labels." For example, an evaluation metric could be an evaluation score; a higher score indicates better quality content.

[0134] Target release content refers to content that meets the requirements of quality assessment standards or models and can be released in the content association area of ​​the project reference content. Target release content can be unmodified content that meets the quality assessment standards or models, or it can be modified content that meets the quality assessment standards or models.

[0135] In practical applications, there are multiple ways to determine the target content to be published based on evaluation metrics. The specific method to be selected depends on the actual situation. In one possible implementation of this specification, the content to be published includes multiple candidate content, each candidate content corresponds to an evaluation metric, the evaluation metric is compared with an evaluation threshold, and the candidate content corresponding to the evaluation metric that is higher than the evaluation threshold is taken as the target content to be published.

[0136] In another possible implementation of this specification, the content to be published includes multiple candidate content to be published, each candidate content to be published corresponds to an evaluation index, and the candidate content with the highest evaluation index is selected as the target content to be published.

[0137] In another possible implementation of this specification, determining the target content to be published based on evaluation metrics may include the following steps: If the evaluation criteria determine that the content to be published has failed, obtain suggestions for content modification. Based on content modification suggestions, generate the target content for publication.

[0138] It should be noted that content modification suggestions refer to specific and actionable optimization guidelines generated based on evaluation indicators following a quality assessment. These guidelines are provided for content awaiting publication that does not meet the publication standards. For example, a content modification suggestion might be to change word 1 to word 2 in the content to be published.

[0139] Evaluation metrics can directly indicate whether the evaluation of content to be published is successful or unsuccessful. When metric thresholds exist, the evaluation metrics can also be compared to these thresholds to determine success or failure. For example, if the evaluation metric is an evaluation score, the score can be compared to a score threshold; if the score is lower than the threshold, the content is deemed to have failed the evaluation.

[0140] In practical applications, there are multiple ways to obtain content modification suggestions, and the specific method to be selected depends on the actual situation. In one possible implementation method of this specification, the content to be published and the quality assessment criteria are input into the quality assessment model to obtain the assessment indicators and content modification suggestions. The quality assessment model is trained based on sample published content, sample assessment criteria, sample assessment indicators, and sample modification suggestions.

[0141] In another possible implementation of this manual, the system pre-establishes a "problem-suggestion" rule base. The evaluation metrics indicate the reasons for the failure of the content to be published. Based on the specific reasons for failure, the system searches the rule base for content modification suggestions that match the reasons for failure. For example, if the evaluation metric is that key selling point A is missing, the rule base suggestion would be "Please add a description of selling point A in the second paragraph or at the end, emphasizing the value it can bring to users."

[0142] There are multiple ways to generate target published content based on content modification suggestions. The specific method to be selected depends on the actual situation. In one possible implementation method described in this specification, the items to be modified in the content to be published are determined based on the content modification suggestions, and the items to be modified are modified to obtain the target published content.

[0143] In another possible implementation of this specification, the content modification suggestions and the content to be published are input into the content modification model to obtain the target published content. The content modification model is trained based on the sample modification suggestions, the sample content to be published, and the sample target published content.

[0144] In another possible implementation of this specification, if the content modification suggestion indicates that the content to be published needs to be regenerated, the content modification suggestion, project marketing information, and project reference content are input into the content generation model to obtain the updated content to be published, and the updated content to be published is used as the target content to be published.

[0145] In one optional embodiment of this specification, the content to be published, quality assessment criteria, and modification prompts are input into a content modification model to obtain the target published content. The content modification model is a multimodal large model, obtained through supervised fine-tuning based on sample content to be published, sample assessment criteria, and sample target published content.

[0146] For example, the modification prompt message could be: "Role: You are a reviewer of comments on notes. Based on the provided comments and relevant rules, you help me make modifications."

[0147] Objective: [Specific objective details].

[0148] Product selling points: [Specific product selling points].

[0149] Skills: Skill 1, analyze comment content; Skill 2, deeply understand the rules, ensuring that "certain content" is not allowed and "some content" must be included; Skill 3, modify and output, based on the rules, the given comments, and the product's selling points, help me modify and correct.

[0150] Specific requirements: [Specific requirements details].

[0151] The solution implemented in the embodiments of this specification determines the target content to be published based on content modification suggestions, thereby automating and refining the modification process, improving optimization efficiency, and ensuring the correctness and consistency of the modification direction.

[0152] By applying the solutions in the embodiments of this specification, the content to be published is evaluated, and the target content to be published is determined based on the evaluation indicators, thereby effectively intercepting low-quality, non-compliant, or risky content and improving the content quality of the target content to be published.

[0153] The scheme described in this specification involves obtaining the target project's marketing information and reference content; inputting the marketing information and reference content into a content generation model to obtain content to be published, wherein the marketing information constrains the content generation direction of the target project; and publishing the content to be published in the content association area of ​​the reference content. By inputting the target project's marketing information and reference content together into the content generation model, it is effectively ensured that the generated content to be published not only conforms to the target project's core marketing strategy but also possesses an expression form and context highly consistent with the reference content, thereby significantly improving the relevance, consistency, and conversion effect of the content to be published. Simultaneously, accurately publishing the content to be published in the content association area achieves scenario-based matching and efficient reach of content delivery, enhancing the overall synergy and intelligence level of project marketing communication, and further improving user engagement and project promotion efficiency.

[0154] The following is in conjunction with the appendix Figure 2 Taking the content publishing method provided in this manual as an example in the field of note-taking and commenting, this paper further explains the content publishing method. Figure 2 A flowchart illustrating the processing procedure of a note-commenting method provided in one embodiment of this specification is shown.

[0155] like Figure 2 As shown, the process involves obtaining a login cookie and automatically logging into the platform account via RPA; on the platform search page, searching for corresponding pet notes based on specific search keywords (how to buy pet insurance, is pet medical insurance necessary, what do first-time cat / dog owners need to prepare, how to raise a cat / dog, is it recommended for college students to raise cats, adopting kittens / puppies), and obtaining platform note links; extracting text from the note links using a large model to obtain platform note text; filtering and tagging the searched note text using the large model, selecting platform note text that matches the scenario and removing advertisements; obtaining pet insurance product selling points from the pet insurance product database, and using the large model to process the platform note text that matches the scenario by writing prompts and combining them with pet insurance product selling points, generating note comments; obtaining expert review experience from the expert experience knowledge base, and automatically reviewing and revising the generated note comments using the large model and expert review experience, obtaining note comments to be published; and automatically publishing the note comments to be published under the corresponding platform notes.

[0156] Taking the search keyword "how to buy pet insurance" as an example, this describes the principle and implementation method of automatically generating comments: RPA logs in and obtains the latest platform note links; extracts the platform note titles and content; categorizes and tags the obtained pet notes, identifying those that match specific scenarios; writes prompts and uses a large model to categorize the given note content into: beginner pet owners, purebred / domestic cats and dogs, student groups, free adoption posts, and pet insurance related; generates comments for notes that match specific scenarios, reviews and revises them, and obtains the final comments. According to the above embodiment, comments are efficiently published in batches under specific notes through automation to promote pet insurance products, increase brand awareness, and expand premium income.

[0157] In this embodiment of the specification, note links are obtained and note text is extracted based on specific search keywords; tags and categories are performed according to specific scenarios; comment content is generated by combining the characteristics of insurance products and note content; finally, a review and revision mechanism is implemented, which uses an expert-accumulated knowledge base for automated review and revision.

[0158] Optionally, notes can be crawled from the platform; the notes can be categorized and labeled using a traditional classification model; and scores can be generated based on expert knowledge, experience, and rules, with the highest-scoring comment being selected.

[0159] Considering the large number of model parameters in the content generation model and the limited computing resources on the client side, the content publishing method proposed in the embodiments of this specification can be applied to, for example... Figure 3 The content publishing system shown is not limited to this. See also Figure 3 , Figure 3 This specification illustrates an architecture diagram of a content publishing system according to an embodiment of the present specification. The content publishing system may include a client 302 and a server 304. Client 302 is used to send the target project's marketing information and project reference content to server 304. The server-side 304 is used to input project marketing information and project reference content into the content generation model to obtain the content to be published. The project marketing information is used to constrain the content generation direction of the target project; the content to be published is sent to the client-side 302. Client 302 is also used to receive content to be published sent by server 304; and to publish the content to be published to the content association area of ​​the project reference content.

[0160] like Figure 3As shown, the content generation model is deployed in server 304. Server 304 can connect to one or more clients 302 via a local area network (LAN), wide area network (WAN), internet connection, or other types of data network. Data transmitted by client 302 may require encoding, transcoding, compression, or other processing before being published to server 304. Client 302 can also interact with users through a graphical user interface to invoke the content generation model, thereby implementing the content publishing method provided in this embodiment. Multiple clients 302 can establish communication connections through server 304. In the content publishing scenario, server 304 provides content publishing services between multiple clients 302. Multiple clients 302 can act as senders or receivers, communicating through server 304. Users can interact with server 304 through client 302 to receive data sent by other clients 302, or send data to other clients 302, etc. In a content publishing scenario, a user can publish the target project's marketing information and reference content to the server 304 via client 302. The server 304 then generates the content to be published based on the marketing information and reference content and pushes the content to be published to other clients that have established communication.

[0161] Client 302 can be a browser, application (APP), or web application such as HyperText Markup Language 5 (H5) application, or a lightweight application (also known as a mini-program), or cloud application, etc. Client 302 can be developed based on the software development kit (SDK) of the corresponding service provided by server 304, such as based on the Real-Time Communication (RTC) SDK. Client 302 can be deployed in electronic devices and depends on the device to run or on certain APPs on the device. Electronic devices may have a display screen and support information browsing, such as personal mobile terminals such as mobile phones, tablets, and personal computers (PCs). Various other types of applications can also be configured in electronic devices, such as human-computer interaction applications, model training applications, text processing applications, web browser applications, shopping applications, search applications, instant messaging tools, email clients, social platform software, etc.

[0162] Server-side 304 can include servers providing various services, such as servers providing communication services to multiple clients, servers supporting backend training of models used on clients, and servers processing data sent by clients. It should be noted that server-side 304 can be implemented as a distributed server cluster composed of multiple servers, or as a single server. The server can also be a server in a distributed system, or a server integrated with blockchain. The server can also be a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms, or an intelligent cloud computing server or intelligent cloud host with artificial intelligence technology.

[0163] It is worth noting that the content publishing method provided in the embodiments of this specification is generally executed by the server. However, in other embodiments of this specification, if the client's runtime resources can meet the deployment and runtime conditions of the content generation model, the client can also have similar functions to the server, thereby executing the content publishing method provided in the embodiments of this specification. In other embodiments, the content publishing method provided in the embodiments of this specification can also be executed jointly by the client and the server.

[0164] Corresponding to the above method embodiments, this specification also provides embodiments of a content publishing device. Figure 4 A schematic diagram of a content publishing device according to one embodiment of this specification is shown. Figure 4 As shown, the device includes: Module 402 is configured to acquire project marketing information and project reference content of the target project. The generation module 404 is configured to input project marketing information and project reference content into the content generation model to obtain content to be published, wherein the project marketing information is used to constrain the content generation direction of the target project; The publishing module 406 is configured to publish the content to be published to the content association area of ​​the project reference content.

[0165] Optionally, module 402 is also configured to acquire project marketing information and project topics for the target project; and retrieve project reference content from the content publishing platform based on the project topics.

[0166] Optionally, the acquisition module 402 is further configured to retrieve candidate reference content from the content publishing platform based on the project topic, wherein the candidate reference content is content related to the project topic in the content publishing platform; identify the category of the candidate reference content to obtain the candidate content category, wherein the candidate content category is used to reflect the publishing scenario corresponding to the candidate reference content; and filter out project reference content from the candidate reference content based on the candidate content category, wherein the candidate content category corresponding to the project reference content conforms to the publishing scenario of the target project.

[0167] Optionally, the acquisition module 402 is further configured to input the candidate reference content into the content parsing model to obtain the title and interactive text of the candidate reference content; and based on the title and interactive text, identify the category of the candidate reference content to obtain the candidate content category.

[0168] Optionally, the candidate content categories include advertising categories and non-advertising categories; the acquisition module 402 is also configured to determine the candidate reference content of non-advertising categories as project reference content; and / or, remove the content to be deleted from the candidate reference content to obtain the project reference content, wherein the content to be deleted belongs to the advertising category.

[0169] Optionally, module 402 is also configured to retrieve the content index corresponding to the project topic from the content publishing platform based on the login credentials of the content publishing platform; and to obtain the project reference content based on the content index.

[0170] Optionally, the generation module 404 is also configured to input generation prompt information, project marketing information, and project reference content into the content generation model to obtain content to be published, wherein the generation prompt information is used to guide the content generation model to generate content to be published based on the project marketing information and project reference content.

[0171] Optionally, the generation module 404 is also configured to obtain the reference content category of the project reference content; and to build and generate prompt information based on the reference content category.

[0172] Optionally, the publishing module 406 is also configured to evaluate the content to be published using quality assessment criteria and / or quality assessment models to obtain assessment indicators, wherein the assessment indicators are used to reflect the quality of the content to be published; determine the target content to be published based on the assessment indicators; and publish the target content to the content association area of ​​the project reference content.

[0173] Optionally, the publishing module 406 is also configured to obtain content modification suggestions if the evaluation of the content to be published is determined to be unsuccessful based on the evaluation metrics; and to generate target publishing content based on the content modification suggestions.

[0174] The scheme described in this specification involves obtaining the target project's marketing information and reference content; inputting the marketing information and reference content into a content generation model to obtain content to be published, wherein the marketing information constrains the content generation direction of the target project; and publishing the content to be published in the content association area of ​​the reference content. By inputting the target project's marketing information and reference content together into the content generation model, it is effectively ensured that the generated content to be published not only conforms to the target project's core marketing strategy but also possesses an expression form and context highly consistent with the reference content, thereby significantly improving the relevance, consistency, and conversion effect of the content to be published. Simultaneously, accurately publishing the content to be published in the content association area achieves scenario-based matching and efficient reach of content delivery, enhancing the overall synergy and intelligence level of project marketing communication, and further improving user engagement and project promotion efficiency.

[0175] The above is an illustrative scheme of a content publishing device according to this embodiment. It should be noted that the technical solution of this content publishing device and the technical solution of the content publishing method described above belong to the same concept. For details not described in detail in the technical solution of the content publishing device, please refer to the description of the technical solution of the content publishing method described above.

[0176] Figure 5 A structural block diagram of a computing device according to one embodiment of this specification is shown. The components of the computing device 500 include, but are not limited to, a memory 510 and a processor 520. The processor 520 is connected to the memory 510 via a bus 530, and a database 550 is used to store data.

[0177] The computing device 500 also includes an access device 540, which enables the computing device 500 to communicate via one or more networks 560. Examples of these networks include Public Switched Telephone Network (PSTN), Local Area Network (LAN), Wide Area Network (WAN), Personal Area Network (PAN), or combinations of communication networks such as the Internet. The access device 540 may include one or more of any type of wired or wireless network interface (e.g., Network Interface Card (NIC)), such as an IEEE 402.11 Wireless Local Area Networks (WLAN) interface, a Wi-MAX (World Interoperability for Microwave Access) interface, an Ethernet interface, a Universal Serial Bus (USB) interface, a cellular network interface, a Bluetooth interface, a Near Field Communication (NFC) interface, and so on.

[0178] In one embodiment of this specification, the above-described components of the computing device 500 and Figure 5 Other components, not shown, can also be connected to each other, for example, via a bus. It should be understood that... Figure 5 The block diagram of the computing device shown is for illustrative purposes only and is not intended to limit the scope of this specification. Those skilled in the art can add or replace other components as needed.

[0179] The computing device 500 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. The computing device 500 can also be a mobile or stationary server.

[0180] The processor 520 is used to execute computer programs / instructions, which, when executed by the processor, implement the steps of the above-described content publishing method.

[0181] The above is an illustrative scheme of a computing device according to this embodiment. It should be noted that the technical solution of this computing device and the technical solution of the content publishing method described above belong to the same concept. For details not described in detail in the technical solution of the computing device, please refer to the description of the technical solution of the content publishing method described above.

[0182] An embodiment of this specification also provides a computer-readable storage medium storing a computer program / instructions that, when executed by a processor, implement the steps of the above-described content publishing method.

[0183] The above is an illustrative scheme of a computer-readable storage medium according to this embodiment. It should be noted that the technical solution of this storage medium and the technical solution of the content publishing method described above belong to the same concept. For details not described in detail in the technical solution of the storage medium, please refer to the description of the technical solution of the content publishing method described above.

[0184] An embodiment of this specification also provides a computer program product, including a computer program / instructions that, when executed by a processor, implement the steps of the above-described content publishing method.

[0185] The above is an illustrative scheme 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 content publishing method described above belong to the same concept. For details not described in detail in the technical solution of the computer program product, please refer to the description of the technical solution of the content publishing method described above.

[0186] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0187] Computer instructions include computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. Computer-readable media can include: any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in computer-readable media can be appropriately added or removed according to the requirements of patent practice. For example, in some regions, according to patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.

[0188] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that the embodiments in this specification are not limited to the described order of actions, because according to the embodiments in this specification, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in this specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to the embodiments in this specification.

[0189] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0190] The preferred embodiments disclosed above are merely illustrative of this specification. The optional embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the embodiments described herein. These embodiments are selected and specifically described in this specification to better explain the principles and practical applications of the embodiments, thereby enabling those skilled in the art to better understand and utilize this specification. This specification is limited only by the claims and their full scope and equivalents.

Claims

1. A content publishing method, characterized in that, include: Obtain project marketing information and project reference materials for the target project; The project marketing information and the project reference content are input into the content generation model to obtain the content to be published, wherein the project marketing information is used to constrain the content generation direction of the target project; Publish the content to be published to the content association area of ​​the project reference content.

2. The method according to claim 1, characterized in that, The acquisition of the target project's marketing information and reference content includes: Obtain the project marketing information and project topics of the target project; Based on the project topic, retrieve reference content for the project from the content publishing platform.

3. The method according to claim 2, characterized in that, The step of retrieving reference content for the project from the content publishing platform based on the project topic includes: Based on the project topic, candidate reference content is retrieved from the content publishing platform, wherein the candidate reference content is content related to the project topic in the content publishing platform; Identify the category of the candidate reference content to obtain the candidate content category, wherein the candidate content category is used to reflect the publishing scenario corresponding to the candidate reference content; Based on the candidate content category, the project reference content is selected from the candidate reference content, wherein the candidate content category corresponding to the project reference content conforms to the release scenario of the target project.

4. The method according to claim 3, characterized in that, The process of identifying the category of the candidate reference content and obtaining the candidate content category includes: The candidate reference content is input into the content parsing model to obtain the title and interactive text of the candidate reference content; Based on the title and the interactive text, the category of the candidate reference content is identified, and the candidate content category is obtained.

5. The method according to claim 3, characterized in that, The candidate content categories include advertising categories and non-advertising categories; The step of filtering the project reference content from the candidate reference content based on the candidate content category includes: The candidate reference content in the non-advertising category is determined as the project reference content; and / or, From the candidate reference content, the content to be deleted is removed to obtain the project reference content, wherein the content to be deleted belongs to the advertising category.

6. The method according to claim 2, characterized in that, The step of retrieving reference content for the project from the content publishing platform based on the project topic includes: Based on the login credentials of the content publishing platform, retrieve the content index corresponding to the project topic from the content publishing platform; Based on the content index, obtain the reference content for the project.

7. The method according to claim 1, characterized in that, The step of inputting the project marketing information and the project reference content into the content generation model to obtain the content to be published includes: The generated prompt information, the project marketing information, and the project reference content are input into the content generation model to obtain the content to be published. The generated prompt information is used to guide the content generation model to generate the content to be published based on the project marketing information and the project reference content.

8. The method according to claim 7, characterized in that, Before inputting the generated prompt information, the project marketing information, and the project reference content into the content generation model to obtain the content to be published, the method further includes: Obtain the reference content category of the project's reference content; The generated prompt information is constructed based on the reference content category.

9. The method according to any one of claims 1 to 8, characterized in that, Before publishing the content to be published to the content association area of ​​the project reference content, the method further includes: The content to be published is evaluated using quality assessment standards and / or quality assessment models to obtain assessment indicators, wherein the assessment indicators are used to reflect the quality of the content to be published. Based on the aforementioned evaluation metrics, the target content to be published is determined; The step of publishing the content to be published to the content association area of ​​the project reference content includes: Publish the target content to the content association area of ​​the project reference content.

10. The method according to claim 9, characterized in that, The determination of target content to be published based on the evaluation metrics includes: If the evaluation of the content to be published is determined to be unsuccessful based on the evaluation indicators, suggestions for content modification will be obtained. Based on the content modification suggestions, the target published content is generated.

11. A content publishing device, characterized in that, include: The acquisition module is configured to acquire project marketing information and project reference content for the target project. The generation module is configured to input the project marketing information and the project reference content into the content generation model to obtain the content to be published, wherein the project marketing information is used to constrain the content generation direction of the target project; The publishing module is configured to publish the content to be published to the content association area of ​​the project reference content.

12. A computing device, characterized in that, include: Memory and processor; The memory is used to store computer programs / instructions, and the processor is used to execute the computer programs / instructions, which, when executed by the processor, implement the steps of the method according to any one of claims 1 to 10.

13. A computer-readable storage medium, characterized in that, It stores a computer program / instructions that, when executed by a processor, implement the steps of the method according to any one of claims 1 to 10.

14. A computer program product, characterized in that, Includes a computer program / instructions that, when executed by a processor, implement the steps of the method according to any one of claims 1 to 10.

Images