AI-based intelligent generation method, device, and equipment for enterprise training courseware

Abstract

This invention discloses an intelligent generation method, apparatus, and device for enterprise training courseware based on artificial intelligence. First, an initial demand dataset is obtained. Then, the initial demand dataset is input into a first pre-trained intelligent agent to perform courseware topic selection and courseware generation task decomposition, obtaining current courseware topic output information and current courseware generation task decomposition information. This is then input into a second pre-trained intelligent agent to generate courseware content, obtaining the current courseware output dataset. Finally, the current courseware output dataset is input into a third pre-trained intelligent agent for content review, obtaining the reviewed courseware output dataset. This invention can combine an initial demand dataset including initial user basic information and training task basic information, sequentially inputting it into multiple pre-trained intelligent agents to perform courseware topic selection, courseware generation task decomposition, courseware content generation, and review, respectively, resulting in training courseware content that is highly relevant to the training task and closely aligned with its content, thus improving the efficiency of training courseware generation.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and in particular to a method, apparatus, and equipment for intelligently generating enterprise training courseware based on artificial intelligence. Background Technology

[0002] In corporate training scenarios, training content is the core vehicle connecting training objectives with the improvement of employee capabilities. The efficiency, quality, and relevance of training content production determine the overall quality of the training. For example, in the financial industry, whether they are business personnel or employees of other companies, they need to master not only professional knowledge in the financial industry but also the professional knowledge of the industries of the clients they work with.

[0003] Taking a property insurance agent as an example, if their client comes from the chemical industry, the agent needs to possess not only insurance expertise but also expertise in the chemical industry (especially safety expertise) to identify key factors and unreasonable aspects during the insurance application and claims process. The financial industry has very high requirements for knowledge diversity, professionalism, rigor, and timeliness, which places higher demands on corporate training departments in creating training courses.

[0004] When disseminating professional knowledge to company employees, a common approach is for training instructors to create training materials (which can be in common presentation formats, video formats, or other formats) based on a relevant training topic, and then conduct corresponding training courses to explain the professional knowledge. When creating training materials, instructors often need to develop a course outline based on their own experience, and then collect and organize relevant knowledge data for the specific training content based on the outline, ultimately forming a training material. This manual method of preparing training materials requires a high level of expertise from the training instructor, but the efficiency and accuracy of collecting and organizing relevant knowledge data based on the course outline are low. This leads to low efficiency in generating training materials and a low degree of compatibility between the relevant knowledge data and the course outline.

[0005] Currently, some training courseware generation methods combining artificial intelligence tools have emerged. These typically require users to first upload training content as an attachment, then input prompts in the corresponding dialog box of the AI ​​tool (such as a large language model) to specify the conditions to be met in the training courseware generation task (such as using a specific page style, one or more font types, etc.), ultimately generating a presentation-format training courseware. However, this method generally only allows for intelligent adjustments to non-core content outside the training material itself. It cannot assist users in quickly collecting and organizing relevant knowledge data for the initial training content, leading to low efficiency in the entire training courseware generation process and a low degree of compatibility between relevant knowledge data and the course outline. Summary of the Invention

[0006] This invention provides an intelligent generation method, apparatus, and device for enterprise training courseware based on artificial intelligence. It aims to address the shortcomings of existing training courseware generation methods that combine AI tools. These methods typically require users to first upload training content in document format as attachments and then input prompts to guide the generation of presentation-format training courseware. They can only intelligently adjust non-core content outside the training material and cannot assist users in quickly collecting and organizing relevant knowledge data for the initial training content. This results in low efficiency throughout the training courseware generation process and a low degree of compatibility between relevant knowledge data and the course outline.

[0007] In a first aspect, embodiments of the present invention provide an intelligent generation method for enterprise training courseware based on artificial intelligence, comprising:

[0008] In response to a courseware intelligent generation command sent by a user terminal, an initial demand dataset corresponding to the courseware intelligent generation command is obtained; wherein, the initial demand dataset includes the initial user basic information of the training instructor and the basic information of the training task; the basic information of the training task includes at least the specific training scenario, the type of training audience, and the initial input data of the courseware.

[0009] The initial requirement dataset is input into the first pre-trained agent to decompose the courseware topic selection and courseware generation task, and the current courseware topic output information and the corresponding current courseware generation task decomposition information are obtained; wherein, the current courseware generation task decomposition information includes at least the current courseware chapter requirement information, the current courseware section requirement information and the current courseware content requirement information;

[0010] The current courseware generation task decomposition information is input into the second pre-trained agent to generate courseware content, and the current courseware output dataset is obtained.

[0011] The current courseware output dataset is input into the third pre-trained agent for content review, and the reviewed courseware output dataset is obtained and sent to the user terminal.

[0012] Secondly, embodiments of the present invention also provide an intelligent generation device for enterprise training courseware based on artificial intelligence, which includes a unit that performs the method described in the first aspect above.

[0013] Thirdly, embodiments of the present invention also provide a computer device, which includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the method described in the first aspect above.

[0014] Fourthly, embodiments of the present invention also provide a computer-readable storage medium storing a computer program, the computer program including program instructions that, when executed by a processor, can implement the method described in the first aspect above.

[0015] This invention provides an artificial intelligence-based method, apparatus, and device for intelligent generation of enterprise training courseware. The method includes: responding to a courseware intelligent generation instruction sent by a user terminal, obtaining an initial requirement dataset corresponding to the courseware intelligent generation instruction; wherein, the initial requirement dataset includes initial user basic information of the training instructor and basic information of the training task; the basic information of the training task includes at least the specific training scenario, the type of training audience, and initial input data of the courseware; inputting the initial requirement dataset into a first pre-trained intelligent agent to perform courseware topic selection and courseware generation task decomposition, obtaining current courseware topic output information and corresponding current courseware generation task decomposition information; wherein, the current courseware generation task decomposition information includes at least current courseware chapter requirement information, current courseware section requirement information, and current courseware content requirement information; inputting the current courseware generation task decomposition information into a second pre-trained intelligent agent to generate courseware content, obtaining a current courseware output dataset; inputting the current courseware output dataset into a third pre-trained intelligent agent to perform content review, obtaining a reviewed courseware output dataset, and sending it to the user terminal. The embodiments of the present invention can combine an initial demand dataset, which includes initial user basic information and training task basic information, and input it sequentially into multiple pre-trained agents to perform courseware topic selection, courseware generation task decomposition, courseware content generation and review, respectively, to obtain training courseware content that is highly relevant to the training task and fits the content, thereby improving the efficiency of training courseware generation. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a schematic diagram illustrating an application scenario of the AI-based intelligent generation method for enterprise training courseware provided in an embodiment of the present invention.

[0018] Figure 2 A flowchart illustrating the intelligent generation method for enterprise training courseware based on artificial intelligence provided in an embodiment of the present invention;

[0019] Figure 3 A schematic diagram of a sub-process of the intelligent generation method for enterprise training courseware based on artificial intelligence provided in an embodiment of the present invention;

[0020] Figure 4 This is another schematic diagram of a sub-process of the intelligent generation method for enterprise training courseware based on artificial intelligence provided in an embodiment of the present invention;

[0021] Figure 5 A schematic block diagram of an intelligent enterprise training courseware generation device based on artificial intelligence provided in an embodiment of the present invention;

[0022] Figure 6 A schematic block diagram of a computer device provided for an embodiment of the present invention. Detailed Implementation

[0023] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] It should be understood that, when used in this specification and the appended claims, the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0025] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0026] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0027] Please also refer to Figure 1 and Figure 2 ,in Figure 1 This is a schematic diagram illustrating a scenario of the intelligent generation method for enterprise training courseware based on artificial intelligence, according to an embodiment of the present invention. Figure 2 This is a flowchart illustrating the intelligent generation method for enterprise training courseware based on artificial intelligence provided in an embodiment of the present invention. Figure 1 As shown, the intelligent generation method for enterprise training courseware based on artificial intelligence provided in this embodiment of the invention is applied to server 10, and user terminal 20 is communicatively connected to server 10. Figure 2 As shown, the method includes the following steps S110-S140.

[0028] S110. In response to the courseware intelligent generation instruction sent by the user terminal, obtain the initial requirement dataset corresponding to the courseware intelligent generation instruction.

[0029] The initial requirements dataset includes the initial user basic information of the training instructors and the basic information of the training tasks; the basic information of the training tasks includes at least the specific training scenario, the type of training audience, and the initial input data of the courseware.

[0030] In this embodiment, the technical solution is described with the server as the execution entity. An enterprise courseware intelligent generation platform is deployed on the server. Users log in to the enterprise courseware intelligent generation platform by entering their login information (such as user registration account and password) through a user terminal. They can first enter the initial requirement dataset in the dialog box of the user interaction interface provided by the enterprise courseware intelligent generation platform. It is important to note that, unlike existing methods where users first upload training content in document format as attachments and then input prompts in the corresponding dialog box of an AI tool (such as a large language model) to create a training presentation file, the initial requirement dataset input by the user includes more dimensions of information. Specifically, it includes basic information about the initial user and basic information about the training task. The basic information about the initial user includes at least the instructor's anonymized name, unique user ID within the company, area of ​​expertise (such as chemical engineering, mechanical engineering, computer science, medicine, etc.), professional level (such as beginner level, junior expert, intermediate expert, senior expert, etc.), work location, age, and gender. The basic information about the training task includes at least the specific training scenario (such as introductory basic training scenario, basic skills enhancement training scenario, intermediate advanced training scenario, etc.), type of training audience (such as internal training within the same company, external training, etc.), and initial input data for the courseware (which can be a short, concise text, a longer, detailed text, or an empty set). Furthermore, the initial user information obtained can be used to build user profiles. During each stage of training courseware production on the enterprise courseware intelligent generation platform, corresponding user profile dimensions can be extracted as decision-making basis. At the same time, user actions at each stage of training courseware production, such as topic preferences, modification traces, and review feedback, can update the user profile in reverse. This enables the platform to provide personalized training courseware production support throughout the entire process driven by user profiles. The various intelligent agents in the enterprise courseware intelligent generation platform can automatically fill in the gaps in the user's weak capabilities in topic selection, framework design, and content production.

[0031] To enable users to provide more accurate initial requirement datasets, the platform can obtain these datasets through a multi-turn dialogue after the user opens the user interface of the enterprise courseware intelligent generation platform on their terminal. For example, the platform might initially ask, "Do you currently have any enterprise courseware intelligent generation needs?" After the user replies with "Yes," the platform might then prompt, "Please provide the lecturer's anonymized name, unique user ID within the company, area of ​​expertise, professional level, work location, age, gender, specific training scenario, target audience type, and initial input data for the courseware." Once the user inputs the corresponding data or uploads the relevant files, the initial requirement dataset is obtained. The platform can further ask the user, "Do you want to start this enterprise courseware intelligent generation?" and initiate the subsequent enterprise courseware intelligent generation process after the user replies with "Yes."

[0032] S120. Input the initial demand dataset into the first pre-trained agent to perform courseware topic selection and courseware generation task decomposition, and obtain the current courseware topic output information and the corresponding current courseware generation task decomposition information.

[0033] The task breakdown information for generating the current courseware includes at least the chapter requirements, section requirements, and content requirements of the current courseware.

[0034] In this embodiment, when creating training courseware, it is necessary to first determine the topic selection information corresponding to the training theme. The enterprise courseware intelligent generation platform can assist users (i.e., training instructors) in selecting courseware topics in at least two ways. The first way is to recommend topics when no clear courseware topic can be extracted from the initial demand dataset. The second way is to provide multiple approximate recommended courseware topics when a clear courseware topic can be extracted from the initial demand dataset, combined with the current courseware topic. After the user determines the output information of the current courseware topic, the first pre-trained agent can further generate the current courseware generation task decomposition information by combining the current courseware topic output information, the type of training audience, and the actual needs of the training courses of the enterprise or organization to which the training audience belongs (e.g., the enterprise to which the training audience belongs has an upper limit requirement on the word count of each chapter of the training courseware).

[0035] Because the first pre-trained agent integrates courseware topic selection and courseware generation task decomposition functions, its deployment embeds at least two agents: a topic recommendation agent and a task decomposition agent, using a chain-like topology. Specifically, the input data of the topic recommendation agent serves as the input data for the task decomposition agent. More specifically, the overall architecture of the first pre-trained agent includes at least a control layer, an agent layer, a communication layer, a knowledge layer, and an execution layer. The main model in the control layer is used for task decomposition, allocation, coordination, result fusion, and agent lifecycle management of the initial requirement dataset. The agent layer includes at least the aforementioned topic recommendation agent and task decomposition agent. The communication layer handles information exchange between the main model and agents, and between agents themselves. The knowledge layer provides knowledge base support for the main model, topic recommendation agent, and task decomposition agent. The execution layer inputs the current courseware topic output information and the corresponding current courseware generation task decomposition information from the task decomposition agent to the other connected agents. Among them, the topic recommendation agent can also analyze content gaps based on the employee job competency model (that is, by identifying the knowledge points corresponding to the instructor's weak content production ability through the instructor's user profile in the initial user basic information), and recommend topics based on the homogeneity analysis of existing courses; moreover, if it is identified that the instructor has professional experience, it will automatically trigger knowledge extraction and recommend training courseware topics with a higher level of professionalism to the user through the topic recommendation agent.

[0036] In one embodiment, such as Figure 3 As shown, step S120 includes:

[0037] S121. If it is determined that the initial input data of the courseware in the initial demand dataset is a brief text type or an empty set, then the initial demand dataset is input into the first pre-trained agent for courseware topic recommendation processing to obtain the current courseware topic output information.

[0038] S122. If it is determined that the initial input data of the courseware in the initial demand dataset is a detailed text material type, then the initial input data of the courseware is input into the first pre-trained agent for courseware topic recommendation processing to obtain the current courseware topic output information;

[0039] S123. If a selection instruction is detected for one of the current courseware topic selection results in the current courseware topic selection output information, then the current target courseware topic selection result is obtained accordingly.

[0040] S124. Input the current target courseware topic selection results and the initial requirement dataset into the first pre-trained agent to decompose the courseware generation task and obtain the current courseware generation task decomposition information.

[0041] In one embodiment, the first pre-trained agent includes at least a topic recommendation agent and a task decomposition agent, and the topic recommendation agent and the task decomposition agent adopt a chain topology structure.

[0042] The step of inputting the initial demand dataset into the first pre-trained agent for courseware topic recommendation processing to obtain the current courseware topic output information includes:

[0043] The initial demand dataset is input into the topic recommendation agent in the first pre-trained agent to perform courseware topic recommendation processing, and the current courseware topic output information is obtained.

[0044] The initial input data of the courseware is input into the first pre-trained agent for courseware topic recommendation processing, resulting in the current courseware topic output information, including:

[0045] The initial input data of the courseware is input into the topic recommendation agent in the first pre-trained agent to perform courseware topic recommendation processing, and the current courseware topic output information is obtained.

[0046] The step of inputting the current target courseware topic selection results and the initial requirement dataset into the first pre-trained agent to decompose the courseware generation task, and obtaining the current courseware generation task decomposition information, includes:

[0047] The current target courseware topic selection results, the initial demand dataset, and the obtained actual training course demand information of the training audience are input into the task decomposition agent in the first pre-trained agent to obtain the current courseware generation task decomposition information.

[0048] In this embodiment, if the initial input data of the courseware in the initial demand dataset is determined to be either brief text or an empty set (e.g., if the total number of characters in the initial input data is less than 100, it can be determined to be brief text; if the initial input data does not contain any information, it is considered an empty set), it indicates that the user's current direction for courseware topic selection is not yet clear. At this time, the initial demand dataset can be input into the topic recommendation agent in the first pre-trained agent for courseware topic recommendation processing to obtain the current courseware topic output information. For example, the current courseware topic output information includes multiple training hot topics related to the actual needs of the training courses of the training audience's company, organization, etc. The user can select one current courseware topic result as the current target courseware topic result.

[0049] If it is determined that the initial input data of the courseware in the initial demand dataset is of the detailed text material type (the total number of words in the initial input data is greater than or equal to 100 words, which can be determined as the detailed text material type), it means that the user's current direction for courseware topic selection is clear. At this time, the initial input data of the courseware is still input into the topic recommendation agent in the first pre-trained agent for courseware topic recommendation processing to obtain the current courseware topic output information. For example, the current courseware topic output information includes multiple candidate courseware topic results whose semantic similarity with the courseware topic provided by the initial input data exceeds the first similarity threshold, as well as the courseware topic provided by the initial input data.

[0050] After the current courseware topic selection output information is displayed in the dialog box of the user interface of the enterprise courseware intelligent generation platform, the user can select one of the current courseware topic selection results as the current target courseware topic selection result. After the topic selection for the training courseware is determined, the actual training course needs information of the enterprise, organization, etc. of the training audience can also be obtained. Through the task decomposition intelligent agent, the current courseware chapter requirement information (which limits the upper limit of the number of chapters in the training courseware), the current courseware section requirement information (which limits the upper limit of the number of sections in each chapter of the training courseware), and the current courseware content requirement information (which limits the upper limit of the word count of the specific course content in each section of each chapter of the training courseware) are extracted from the actual training course needs information.

[0051] S130. Input the current courseware generation task decomposition information into the second pre-trained agent to generate courseware content and obtain the current courseware output dataset.

[0052] In this embodiment, when the first pre-trained agent outputs the current courseware generation task decomposition information, it is used as input data to the second pre-trained agent to generate courseware content, resulting in the current courseware output dataset. It is important to note that the knowledge base connected to the second pre-trained agent is obtained through knowledge extraction and data reconstruction based on the database of the knowledge engine uploaded by enterprise users to the enterprise courseware intelligent generation platform. The knowledge data stored in the knowledge base connected to the second pre-trained agent can serve as a local knowledge base for courseware content generation. Of course, this knowledge base can also serve as a local knowledge base for other agents within the enterprise courseware intelligent generation platform. In practice, in addition to inputting the task breakdown information of the current courseware generation into the second pre-trained agent, the learning behavior preferences corresponding to the training audience type and some suitable teaching methodologies (such as the BOPPPS teaching model, the 5E teaching method, etc.; where the BOPPPS teaching model includes Bridge and introduction, Objective (learning objectives), Pre-assessment (pre-class assessment), Participatory Learning (participatory learning), Post-assessment (post-class quiz), and Summary; the 5E teaching method consists of engagement, exploration, explanation, extension, and evaluation) are also input into the second pre-trained agent as auxiliary data for courseware content generation. Moreover, the generated current courseware output dataset can be optimized according to user instructions before subsequent content review.

[0053] In one embodiment, such as Figure 4 As shown, step S130 includes:

[0054] S131. Input the current courseware chapter requirement information from the current courseware generation task decomposition information into the chapter generation agent in the second pre-trained agent to obtain the current chapter output data;

[0055] S132. Input the current courseware section requirement information from the current courseware generation task decomposition information into the section generation agent in the second pre-trained agent to obtain the current section output data.

[0056] S133. Input the current courseware content requirement information from the current courseware generation task decomposition information into the content generation agent in the second pre-trained agent to obtain the current courseware content output data.

[0057] S134. Based on the target section to which the current courseware content output sub-data included in the current courseware content output data belongs in the current section output data, and the target chapter to which the current chapter output data belongs, each current courseware content output sub-data is grouped into the target section of the corresponding target chapter to obtain the current courseware output dataset.

[0058] In one embodiment, step S133 includes:

[0059] The current courseware content requirement information in the current courseware generation task decomposition information is input into the content generation agent, so that the content generation agent can obtain courseware content from the locally connected knowledge base or network database and generate the current courseware content output data.

[0060] In this embodiment, existing technologies that generate training courseware using artificial intelligence tools often result in low compatibility between relevant knowledge data and the course outline. The difference lies in the fact that this application first selects courseware topics and then combines the output information of the selected topics with the actual needs of the training course to determine the current courseware generation task decomposition information. Subsequently, the current courseware chapter requirements, current courseware section requirements, and current courseware content requirements information from the current courseware generation task decomposition information can be used as the current course outline to guide the generation of the current courseware output dataset. This results in a high degree of fit between the current courseware output dataset and the current courseware generation task decomposition information.

[0061] The second pre-trained agent also embeds chapter generation agents, section generation agents, and content generation agents, and adopts a chain-like topology (i.e., connected sequentially according to the order of chapter generation agents, section generation agents, and content generation agents). Similar to the first pre-trained agent, its overall architecture includes at least a control layer, agent layer, communication layer, knowledge layer, and execution layer. Each agent in the second pre-trained agent generates corresponding training courseware content. Finally, based on the target section in the current section output data and the target chapter in the current chapter output data, each current courseware content output sub-data is grouped into the corresponding target section within the target chapter, resulting in the current courseware output dataset. The second pre-trained agent can also be equipped with a counter to count whether all chapters in the current chapter output data and all sections in the current section output data have corresponding current courseware content output sub-data to complete content filling.

[0062] In this case, determining the training audience type based on the actual needs information of the training course indicates an external training type. Since the local knowledge base may not yet contain strong relevant information about the corresponding enterprise, group, or organization, the content generation agent can legally and compliantly obtain courseware content (such as the latest products and parameter introductions corresponding to the enterprise of the external training audience) from a networked database and generate the current courseware content output data. It is evident that the above data acquisition method using the content generation agent allows for multi-source acquisition of target data, ensuring that the current courseware content output data is not disconnected from the enterprise's actual business and can also obtain the latest training-related data from external sources, avoiding the problem of limited courseware content obtained from a single knowledge base.

[0063] In one embodiment, prior to step S130, the method further includes:

[0064] If a currently uploaded courseware file is detected and it is determined that the currently uploaded courseware file is a non-duplicate courseware file, then the professional field, training audience type, and specific training scenario in the currently uploaded courseware file are obtained based on the preset courseware file parsing strategy, and the current courseware extraction result is obtained.

[0065] Based on a preset file reconstruction strategy, the currently uploaded courseware file is enhanced with domain knowledge, structured fragmentation, and document vectorization data acquisition to obtain the current file reconstruction result.

[0066] After establishing a mapping relationship between the current courseware extraction result, the current file reconstruction processing result, and the currently uploaded courseware file, all are stored in the knowledge base to update the knowledge base.

[0067] In this embodiment, when building the knowledge base in the knowledge engine on the server, the data source for its construction is courseware files uploaded by multiple users. For example, taking the server extracting knowledge from a currently uploaded courseware file to form one of the knowledge files in the knowledge base as an example, the server's knowledge base already stores knowledge data corresponding to other uploaded courseware files. When the server receives the currently uploaded courseware file, it needs to parse and process the file. When parsing the currently uploaded courseware file, semantic understanding can be performed through courseware file parsing strategies to obtain the professional field (such as chemical industry, finance, automobile, medical equipment, electronic components, etc.), training audience type (such as internal training within the same company, external training, etc.), and specific training scenario (such as introductory basic training scenario, primary ability improvement training scenario, intermediate advanced improvement training scenario, etc.) in the currently uploaded courseware file for detection, and obtain the current courseware extraction result.

[0068] After the file extraction and parsing of the uploaded courseware file is completed, the extraction result is obtained. Then, the file structure of the uploaded courseware file can be reconstructed, specifically through domain knowledge enhancement, structured segmentation, and document vectorization data acquisition. Domain knowledge enhancement specifically involves parsing the format of the uploaded courseware file (e.g., determining if it is a PDF, Word, TXT document), then performing word segmentation (e.g., dictionary-based segmentation) to obtain segmentation results including multiple keywords. Next, some keywords in the segmentation results are standardized and converted based on a pre-defined domain keyword library (where domain keywords are generally represented by Chinese terms). Finally, the domain theme and core intent of the uploaded courseware file can be identified to obtain a deeper understanding of the domain semantics. The document after standardization based on the domain keyword library constitutes the domain knowledge enhancement result. The uploaded courseware file is structured and segmented. For example, the file can be divided based on its text layout, such as into chapters or sections, resulting in the current structured segmentation. When obtaining document vectorization data from the uploaded courseware file, the chapter titles can be retrieved first, followed by the semantic model corresponding to each title, which is then used as the document vectorization data. After knowledge extraction and file reconstruction are completed, the extracted and reconstructed results are mapped to the uploaded courseware file and stored in a knowledge base, enabling continuous updates to the knowledge base.

[0069] Of course, upon detecting the currently uploaded courseware file, a duplicate file check can be performed first. This involves determining whether the currently uploaded courseware file is a previously uploaded file. If it is confirmed to be a previously uploaded file, a duplicate file upload notification is sent to the user's terminal (this notification confirms whether to continue uploading the file or choose a different courseware file). If it is confirmed that the currently uploaded courseware file is not a previously uploaded file, it indicates that the duplicate file detection has passed, and a file upload success notification can be sent to the user's terminal. This file duplicate detection effectively avoids processing identical courseware files repeatedly, reducing the consumption of server system resources.

[0070] S140. Input the current courseware output dataset into the third pre-trained agent for content review, obtain the reviewed courseware output dataset, and send it to the user terminal.

[0071] In one embodiment, step S140 includes:

[0072] The current courseware output dataset is input into the third pre-trained agent to detect whether there are grammatical errors in the courseware content and whether it meets the preset courseware content constraints through the enterprise-specific review rules corresponding to the third pre-trained agent. Modification suggestion data corresponding to the current courseware output dataset is obtained and combined with the current courseware output dataset to form the reviewed courseware output dataset.

[0073] In this embodiment, the current courseware output dataset output by the second pre-trained agent may still contain some content that needs adjustment or correction, requiring content review and verification. This content can then be input into the third pre-trained agent for further review. The third pre-trained agent can embed enterprise-specific review rules, such as ensuring the courseware content is free of grammatical and common-sense errors and meets the constraints of the PAS methodology (PAS stands for Problem, Agitate, and Solve, encompassing three core modules: problem, agitation, and solution) or the five-star teaching method (including five core environments: focusing on problems, activating prior knowledge, demonstrating new knowledge, applying new knowledge, and integrating knowledge). After the third pre-trained agent reviews the current courseware output dataset and provides corresponding modification suggestions, a reviewed courseware output dataset is generated and sent to the user terminal (or alternatively, it can be displayed on the user interface of the enterprise courseware intelligent generation platform for the user). When users view the approved courseware output dataset through their user terminals or online, they can make corresponding modifications based on the modification suggestions to obtain the final courseware output dataset, which can be used for subsequent training.

[0074] In one embodiment, the method further includes the following after step S140:

[0075] If the final courseware output dataset corresponding to the approved courseware output dataset is received, the final courseware output dataset is input into the fourth pre-trained agent to generate training speech and training script, thereby obtaining the current training courseware speech data and the current training courseware script data.

[0076] In this embodiment, the first to fourth pre-trained agents used in this application can all adopt a large language model, and multiple agents can be embedded in it to complete different functions according to actual needs. After the user reviews the approved courseware output dataset and makes corresponding modifications according to the modification suggestions to obtain the final courseware output dataset (which is displayed as a presentation to the trainees), in order to facilitate the user as an instructor to have more reference materials during the actual training, the final courseware output dataset can also be input into the fourth pre-trained agent to generate training speech and training script, thus obtaining the current training courseware speech data and the current training courseware script data. Specifically, the fourth pre-trained agent integrates a courseware presentation script generation agent and a training courseware script generation agent. The final courseware output dataset is input into the courseware presentation script generation agent to generate current training courseware presentation script data, which guides instructors in actually explaining the training courseware based on the final courseware output dataset. The final courseware output dataset can also be input into the training courseware script generation agent to generate current training courseware script data, which guides instructors in controlling the explanation order, explanation duration, and whether to include interactive communication segments in the final courseware output dataset. Therefore, this method effectively combines the final courseware output dataset to generate presentation scripts and presentation materials, guiding instructors to conduct more efficient knowledge explanations during training.

[0077] In practice, the final courseware output dataset, the current training courseware speech manuscript data, the current training courseware script data, the selected digital human image data, and voice style data can be combined to generate a current digital human explanation video. This serves as an extended explanation method corresponding to the final courseware output dataset. Other users can then repeatedly view this digital human explanation video to learn about related fields and professional knowledge.

[0078] As can be seen, the implementation of this method can combine an initial demand dataset, which includes initial user basic information and training task basic information, and input it sequentially into multiple pre-trained agents to perform courseware topic selection, courseware generation task decomposition, courseware content generation and review, respectively, to obtain training courseware content that is highly relevant to the training task and fits the content, thereby improving the efficiency of training courseware generation.

[0079] Figure 5 This is a schematic block diagram of an intelligent enterprise training courseware generation device based on artificial intelligence, provided in an embodiment of the present invention. Figure 5As shown, corresponding to the above-described AI-based intelligent generation method for enterprise training courseware, the present invention also provides an AI-based intelligent generation device 100 for enterprise training courseware. This AI-based intelligent generation device 100 includes units for executing the above-described AI-based intelligent generation method for enterprise training courseware. Please refer to... Figure 5 The AI-based intelligent enterprise training courseware generation device 100 includes: an initial demand data acquisition unit 110, a courseware generation task decomposition unit 120, a courseware content generation unit 130, and a courseware content review unit 140.

[0080] The initial requirement data acquisition unit 110 is used to acquire the initial requirement dataset corresponding to the courseware intelligent generation instruction sent by the user terminal in response to the courseware intelligent generation instruction.

[0081] The initial requirements dataset includes the initial user basic information of the training instructors and the basic information of the training tasks; the basic information of the training tasks includes at least the specific training scenario, the type of training audience, and the initial input data of the courseware.

[0082] In this embodiment, the technical solution is described with the server as the execution entity. An enterprise courseware intelligent generation platform is deployed on the server. Users log in to the enterprise courseware intelligent generation platform by entering their login information (such as user registration account and password) through a user terminal. They can first enter the initial requirement dataset in the dialog box of the user interaction interface provided by the enterprise courseware intelligent generation platform. It is important to note that, unlike existing methods where users first upload training content in document format as attachments and then input prompts in the corresponding dialog box of an AI tool (such as a large language model) to create a training presentation file, the initial requirement dataset input by the user includes more dimensions of information. Specifically, it includes basic information about the initial user and basic information about the training task. The basic information about the initial user includes at least the instructor's anonymized name, unique user ID within the company, area of ​​expertise (such as chemical engineering, mechanical engineering, computer science, medicine, etc.), professional level (such as beginner level, junior expert, intermediate expert, senior expert, etc.), work location, age, and gender. The basic information about the training task includes at least the specific training scenario (such as introductory basic training scenario, basic skills enhancement training scenario, intermediate advanced training scenario, etc.), type of training audience (such as internal training within the same company, external training, etc.), and initial input data for the courseware (which can be a short, concise text, a longer, detailed text, or an empty set).

[0083] To enable users to provide more accurate initial requirement datasets, the platform can obtain these datasets through a multi-turn dialogue after the user opens the user interface of the enterprise courseware intelligent generation platform on their terminal. For example, the platform might initially ask, "Do you currently have any enterprise courseware intelligent generation needs?" After the user replies with "Yes," the platform might then prompt, "Please provide the lecturer's anonymized name, unique user ID within the company, area of ​​expertise, professional level, work location, age, gender, specific training scenario, target audience type, and initial input data for the courseware." Once the user inputs the corresponding data or uploads the relevant files, the initial requirement dataset is obtained. The platform can further ask the user, "Do you want to start this enterprise courseware intelligent generation?" and initiate the subsequent enterprise courseware intelligent generation process after the user replies with "Yes."

[0084] The courseware generation task decomposition unit 120 is used to input the initial requirement dataset into the first pre-trained agent to perform courseware topic selection and courseware generation task decomposition, and obtain the current courseware topic output information and the corresponding current courseware generation task decomposition information.

[0085] The task breakdown information for generating the current courseware includes at least the chapter requirements, section requirements, and content requirements of the current courseware.

[0086] In this embodiment, when creating training courseware, it is necessary to first determine the topic selection information corresponding to the training theme. The enterprise courseware intelligent generation platform can assist users (i.e., training instructors) in selecting courseware topics in at least two ways. The first way is to recommend topics when no clear courseware topic can be extracted from the initial demand dataset. The second way is to provide multiple approximate recommended courseware topics when a clear courseware topic can be extracted from the initial demand dataset, combined with the current courseware topic. After the user determines the output information of the current courseware topic, the first pre-trained agent can further generate the current courseware generation task decomposition information by combining the current courseware topic output information, the type of training audience, and the actual needs of the training courses of the enterprise or organization to which the training audience belongs (e.g., the enterprise to which the training audience belongs has an upper limit requirement on the word count of each chapter of the training courseware).

[0087] Because the first pre-trained agent integrates courseware topic selection and courseware generation task decomposition functions, its deployment embeds at least two agents: a topic recommendation agent and a task decomposition agent, using a chain-like topology. Specifically, the input data of the topic recommendation agent serves as the input data for the task decomposition agent. More specifically, the overall architecture of the first pre-trained agent includes at least a control layer, an agent layer, a communication layer, a knowledge layer, and an execution layer. The main model in the control layer is used for task decomposition, allocation, coordination, result fusion, and agent lifecycle management of the initial requirement dataset. The agent layer includes at least the aforementioned topic recommendation agent and task decomposition agent. The communication layer handles information exchange between the main model and agents, and between agents themselves. The knowledge layer provides knowledge base support for the main model, topic recommendation agent, and task decomposition agent. The execution layer inputs the current courseware topic output information and the corresponding current courseware generation task decomposition information from the task decomposition agent to the other connected agents.

[0088] In one embodiment, the courseware generation task decomposition unit 120 is specifically used for:

[0089] If it is determined that the initial input data of the courseware in the initial demand dataset is a brief text type or an empty set, then the initial demand dataset is input into the first pre-trained agent for courseware topic recommendation processing to obtain the current courseware topic output information;

[0090] If it is determined that the initial input data of the courseware in the initial demand dataset is of the detailed text material type, then the initial input data of the courseware is input into the first pre-trained agent for courseware topic recommendation processing to obtain the current courseware topic output information;

[0091] If a selection instruction is detected for one of the current courseware topic selection results in the current courseware topic selection output information, then the current target courseware topic selection result is obtained accordingly.

[0092] The current target courseware topic selection results and the initial requirement dataset are input into the first pre-trained agent to decompose the courseware generation task, thereby obtaining the current courseware generation task decomposition information.

[0093] In one embodiment, the first pre-trained agent includes at least a topic recommendation agent and a task decomposition agent, and the topic recommendation agent and the task decomposition agent adopt a chain topology structure.

[0094] The step of inputting the initial demand dataset into the first pre-trained agent for courseware topic recommendation processing to obtain the current courseware topic output information includes:

[0095] The initial demand dataset is input into the topic recommendation agent in the first pre-trained agent to perform courseware topic recommendation processing, and the current courseware topic output information is obtained.

[0096] The initial input data of the courseware is input into the first pre-trained agent for courseware topic recommendation processing, resulting in the current courseware topic output information, including:

[0097] The initial input data of the courseware is input into the topic recommendation agent in the first pre-trained agent to perform courseware topic recommendation processing, and the current courseware topic output information is obtained.

[0098] The step of inputting the current target courseware topic selection results and the initial requirement dataset into the first pre-trained agent to decompose the courseware generation task, and obtaining the current courseware generation task decomposition information, includes:

[0099] The current target courseware topic selection results, the initial demand dataset, and the obtained actual training course demand information of the training audience are input into the task decomposition agent in the first pre-trained agent to obtain the current courseware generation task decomposition information.

[0100] In this embodiment, if the initial input data of the courseware in the initial demand dataset is determined to be either brief text or an empty set (e.g., if the total number of characters in the initial input data is less than 100, it can be determined to be brief text; if the initial input data does not contain any information, it is considered an empty set), it indicates that the user's current direction for courseware topic selection is not yet clear. At this time, the initial demand dataset can be input into the topic recommendation agent in the first pre-trained agent for courseware topic recommendation processing to obtain the current courseware topic output information. For example, the current courseware topic output information includes multiple training hot topics related to the actual needs of the training courses of the training audience's company, organization, etc. The user can select one current courseware topic result as the current target courseware topic result.

[0101] If it is determined that the initial input data of the courseware in the initial demand dataset is of the detailed text material type (the total number of words in the initial input data is greater than or equal to 100 words, which can be determined as the detailed text material type), it means that the user's current direction for courseware topic selection is clear. At this time, the initial input data of the courseware is still input into the topic recommendation agent in the first pre-trained agent for courseware topic recommendation processing to obtain the current courseware topic output information. For example, the current courseware topic output information includes multiple candidate courseware topic results whose semantic similarity with the courseware topic provided by the initial input data exceeds the first similarity threshold, as well as the courseware topic provided by the initial input data.

[0102] After the current courseware topic selection output information is displayed in the dialog box of the user interface of the enterprise courseware intelligent generation platform, the user can select one of the current courseware topic selection results as the current target courseware topic selection result. After the topic selection for the training courseware is determined, the actual training course needs information of the enterprise, organization, etc. of the training audience can also be obtained. Through the task decomposition intelligent agent, the current courseware chapter requirement information (which limits the upper limit of the number of chapters in the training courseware), the current courseware section requirement information (which limits the upper limit of the number of sections in each chapter of the training courseware), and the current courseware content requirement information (which limits the upper limit of the word count of the specific course content in each section of each chapter of the training courseware) are extracted from the actual training course needs information.

[0103] The courseware content generation unit 130 is used to input the current courseware generation task decomposition information into the second pre-trained agent to generate courseware content and obtain the current courseware output dataset.

[0104] In this embodiment, when the first pre-trained agent outputs the current courseware generation task decomposition information, it is used as input data to the second pre-trained agent to generate courseware content, resulting in the current courseware output dataset. It is important to note that the knowledge base connected to the second pre-trained agent is obtained through knowledge extraction and data reconstruction based on the database of the knowledge engine uploaded by enterprise users to the enterprise courseware intelligent generation platform. The knowledge data stored in the knowledge base connected to the second pre-trained agent can serve as a local knowledge base for courseware content generation. Of course, this knowledge base can also serve as a local knowledge base for other agents within the enterprise courseware intelligent generation platform.

[0105] In one embodiment, the courseware content generation unit 130 is specifically used for:

[0106] The current courseware chapter requirement information in the current courseware generation task decomposition information is input into the chapter generation agent in the second pre-trained agent to obtain the current chapter output data;

[0107] The current courseware section requirement information in the current courseware generation task decomposition information is input into the section generation agent in the second pre-trained agent to obtain the current section output data;

[0108] The current courseware content requirement information in the current courseware generation task decomposition information is input into the content generation agent in the second pre-trained agent to obtain the current courseware content output data;

[0109] Based on the target sections to which the current courseware content output sub-data belongs in the current section output data and the target chapters to which they belong in the current chapter output data, each current courseware content output sub-data is grouped into the target sections of the corresponding target chapters to obtain the current courseware output dataset.

[0110] In one embodiment, the step of inputting the current courseware content requirement information from the current courseware generation task decomposition information into the content generation agent in the second pre-trained agent to obtain the current courseware content output data includes:

[0111] The current courseware content requirement information in the current courseware generation task decomposition information is input into the content generation agent, so that the content generation agent can obtain courseware content from the locally connected knowledge base or network database and generate the current courseware content output data.

[0112] In this embodiment, existing technologies that generate training courseware using artificial intelligence tools often result in low compatibility between relevant knowledge data and the course outline. The difference lies in the fact that this application first selects courseware topics and then combines the output information of the selected topics with the actual needs of the training course to determine the current courseware generation task decomposition information. Subsequently, the current courseware chapter requirements, current courseware section requirements, and current courseware content requirements information from the current courseware generation task decomposition information can be used as the current course outline to guide the generation of the current courseware output dataset. This results in a high degree of fit between the current courseware output dataset and the current courseware generation task decomposition information.

[0113] The second pre-trained agent also embeds chapter generation agents, section generation agents, and content generation agents, and adopts a chain-like topology (i.e., connected sequentially according to the order of chapter generation agents, section generation agents, and content generation agents). Similar to the first pre-trained agent, its overall architecture includes at least a control layer, agent layer, communication layer, knowledge layer, and execution layer. Each agent in the second pre-trained agent generates corresponding training courseware content. Finally, based on the target section in the current section output data and the target chapter in the current chapter output data, each current courseware content output sub-data is grouped into the corresponding target section within the target chapter, resulting in the current courseware output dataset. The second pre-trained agent can also be equipped with a counter to count whether all chapters in the current chapter output data and all sections in the current section output data have corresponding current courseware content output sub-data to complete content filling.

[0114] In this case, determining the training audience type based on the actual needs information of the training course indicates an external training type. Since the local knowledge base may not yet contain strong relevant information about the corresponding enterprise, group, or organization, the content generation agent can legally and compliantly obtain courseware content (such as the latest products and parameter introductions corresponding to the enterprise of the external training audience) from a networked database and generate the current courseware content output data. It is evident that the above data acquisition method using the content generation agent allows for multi-source acquisition of target data, ensuring that the current courseware content output data is not disconnected from the enterprise's actual business and can also obtain the latest training-related data from external sources, avoiding the problem of limited courseware content obtained from a single knowledge base.

[0115] In one embodiment, the AI-based intelligent generation device for enterprise training courseware 100 further includes:

[0116] The courseware extraction unit is used to detect the professional field, training audience type, and specific training scenario in the currently uploaded courseware file based on a preset courseware file parsing strategy if the currently uploaded courseware file is detected and it is determined that the currently uploaded courseware file is a non-duplicate courseware file, and obtain the current courseware extraction result.

[0117] The courseware file reconstruction processing unit is used to perform domain knowledge enhancement, structured fragmentation, and document vectorization data acquisition on the currently uploaded courseware file based on a preset file reconstruction and rebuilding strategy, so as to obtain the current file reconstruction processing result;

[0118] The knowledge storage unit is used to establish a mapping relationship between the current courseware extraction result, the current file reconstruction processing result, and the currently uploaded courseware file, and store them in the knowledge base to update the knowledge base.

[0119] In this embodiment, when building the knowledge base in the knowledge engine on the server, the data source for its construction is courseware files uploaded by multiple users. For example, taking the server extracting knowledge from a currently uploaded courseware file to form one of the knowledge files in the knowledge base as an example, the server's knowledge base already stores knowledge data corresponding to other uploaded courseware files. When the server receives the currently uploaded courseware file, it needs to parse and process the file. When parsing the currently uploaded courseware file, semantic understanding can be performed through courseware file parsing strategies to obtain the professional field (such as chemical industry, finance, automobile, medical equipment, electronic components, etc.), training audience type (such as internal training within the same company, external training, etc.), and specific training scenario (such as introductory basic training scenario, primary ability improvement training scenario, intermediate advanced improvement training scenario, etc.) in the currently uploaded courseware file for detection, and obtain the current courseware extraction result.

[0120] After the file extraction and parsing of the uploaded courseware file is completed, the extraction result is obtained. Then, the file structure of the uploaded courseware file can be reconstructed, specifically through domain knowledge enhancement, structured segmentation, and document vectorization data acquisition. Domain knowledge enhancement specifically involves parsing the format of the uploaded courseware file (e.g., determining if it is a PDF, Word, TXT document), then performing word segmentation (e.g., dictionary-based segmentation) to obtain segmentation results including multiple keywords. Next, some keywords in the segmentation results are standardized and converted based on a pre-defined domain keyword library (where domain keywords are generally represented by Chinese terms). Finally, the domain theme and core intent of the uploaded courseware file can be identified to obtain a deeper understanding of the domain semantics. The document after standardization based on the domain keyword library constitutes the domain knowledge enhancement result. The uploaded courseware file is structured and segmented. For example, the file can be divided based on its text layout, such as into chapters or sections, resulting in the current structured segmentation. When obtaining document vectorization data from the uploaded courseware file, the chapter titles can be retrieved first, followed by the semantic model corresponding to each title, which is then used as the document vectorization data. After knowledge extraction and file reconstruction are completed, the extracted and reconstructed results are mapped to the uploaded courseware file and stored in a knowledge base, enabling continuous updates to the knowledge base.

[0121] Of course, upon detecting the currently uploaded courseware file, a duplicate file check can be performed first. This involves determining whether the currently uploaded courseware file is a previously uploaded file. If it is confirmed to be a previously uploaded file, a duplicate file upload notification is sent to the user's terminal (this notification confirms whether to continue uploading the file or choose a different courseware file). If it is confirmed that the currently uploaded courseware file is not a previously uploaded file, it indicates that the duplicate file detection has passed, and a file upload success notification can be sent to the user's terminal. This file duplicate detection effectively avoids processing identical courseware files repeatedly, reducing the consumption of server system resources.

[0122] The courseware content review unit 140 is used to input the current courseware output dataset into the third pre-trained agent for content review, obtain the reviewed courseware output dataset, and send it to the user terminal.

[0123] In one embodiment, the courseware content review unit 140 is specifically used for:

[0124] The current courseware output dataset is input into the third pre-trained agent to detect whether there are grammatical errors in the courseware content and whether it meets the preset courseware content constraints through the enterprise-specific review rules corresponding to the third pre-trained agent. Modification suggestion data corresponding to the current courseware output dataset is obtained and combined with the current courseware output dataset to form the reviewed courseware output dataset.

[0125] In this embodiment, the current courseware output dataset output by the second pre-trained agent may still contain some content that needs adjustment or correction, requiring content review and verification. This content can then be input into the third pre-trained agent for further review. The third pre-trained agent can embed enterprise-specific review rules, such as ensuring the courseware content is free of grammatical and common-sense errors and meets the constraints of the PAS methodology (PAS stands for Problem, Agitate, and Solve, encompassing three core modules: problem, agitation, and solution) or the five-star teaching method (including five core environments: focusing on problems, activating prior knowledge, demonstrating new knowledge, applying new knowledge, and integrating knowledge). After the third pre-trained agent reviews the current courseware output dataset and provides corresponding modification suggestions, a reviewed courseware output dataset is generated and sent to the user terminal (or alternatively, it can be displayed on the user interface of the enterprise courseware intelligent generation platform for the user). When users view the approved courseware output dataset through their user terminals or online, they can make corresponding modifications based on the modification suggestions to obtain the final courseware output dataset, which can be used for subsequent training.

[0126] In one embodiment, the AI-based intelligent generation device for enterprise training courseware 100 further includes:

[0127] The script and presentation unit is used to input the final courseware output dataset corresponding to the approved courseware output dataset into the fourth pre-trained agent to generate training presentations and training scripts, thereby obtaining the current training courseware presentation data and the current training courseware script data.

[0128] In this embodiment, the first to fourth pre-trained agents used in this application can all adopt a large language model, and multiple agents can be embedded in it to complete different functions according to actual needs. After the user reviews the approved courseware output dataset and makes corresponding modifications according to the modification suggestions to obtain the final courseware output dataset (which is displayed as a presentation to the trainees), in order to facilitate the user as an instructor to have more reference materials during the actual training, the final courseware output dataset can also be input into the fourth pre-trained agent to generate training speech and training script, thus obtaining the current training courseware speech data and the current training courseware script data. Specifically, the fourth pre-trained agent integrates a courseware presentation script generation agent and a training courseware script generation agent. The final courseware output dataset is input into the courseware presentation script generation agent to generate current training courseware presentation script data, which guides instructors in actually explaining the training courseware based on the final courseware output dataset. The final courseware output dataset can also be input into the training courseware script generation agent to generate current training courseware script data, which guides instructors in controlling the explanation order, explanation duration, and whether to include interactive communication segments in the final courseware output dataset. Therefore, this method effectively combines the final courseware output dataset to generate presentation scripts and presentation materials, guiding instructors to conduct more efficient knowledge explanations during training.

[0129] In practice, the final courseware output dataset, the current training courseware speech manuscript data, the current training courseware script data, the selected digital human image data, and voice style data can be combined to generate a current digital human explanation video. This serves as an extended explanation method corresponding to the final courseware output dataset. Other users can then repeatedly view this digital human explanation video to learn about related fields and professional knowledge.

[0130] As can be seen, the implementation of this device can combine an initial demand dataset, which includes initial user basic information and training task basic information, and input it sequentially into multiple pre-trained agents to perform courseware topic selection, courseware generation task decomposition, courseware content generation and review, respectively, to obtain training courseware content that is highly relevant to the training task and fits the content, thereby improving the efficiency of training courseware generation.

[0131] The aforementioned AI-based intelligent generation device for enterprise training courseware can be implemented as a computer program, which can, for example... Figure 6 It runs on the computer device shown.

[0132] Please see Figure 6 , Figure 6This is a schematic block diagram of a computer device provided in an embodiment of the present invention. This computer device integrates any of the intelligent enterprise training courseware generation devices based on artificial intelligence provided in the embodiments of the present invention.

[0133] See Figure 6 The computer device 400 includes a processor 402, a memory, and a network interface 405 connected via a system bus 401. The memory may include a storage medium 403 and internal memory 404.

[0134] The storage medium 403 may store an operating system 4031 and a computer program 4032. The computer program 4032 includes program instructions that, when executed, cause the processor 402 to execute an artificial intelligence-based intelligent generation method for enterprise training courseware.

[0135] The processor 402 provides computing and control capabilities to support the operation of the entire computer device.

[0136] The internal memory 404 provides an environment for the computer program 4032 in the storage medium 403 to run. When the computer program 4032 is executed by the processor 402, the processor 402 can execute the above-mentioned intelligent generation method of enterprise training courseware based on artificial intelligence.

[0137] This network interface 405 is used for network communication with other devices. Those skilled in the art will understand that... Figure 6 The structure shown is merely a block diagram of a portion of the structure related to the present invention and does not constitute a limitation on the computer device to which the present invention is applied. A specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0138] The processor 402 is used to run the computer program 4032 stored in the memory to implement the above-mentioned intelligent generation method for enterprise training courseware based on artificial intelligence.

[0139] It should be understood that, in this embodiment of the invention, the processor 402 may be a Central Processing Unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0140] It will be understood by those skilled in the art that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program includes program instructions and can be stored in a storage medium, which is a computer-readable storage medium. The program instructions are executed by at least one processor in the computer system to implement the process steps of the embodiments of the above methods.

[0141] Therefore, the present invention also provides a computer-readable storage medium. This computer-readable storage medium stores a computer program, wherein the computer program includes program instructions. When executed by a processor, the program instructions cause the processor to perform the aforementioned artificial intelligence-based intelligent generation method for enterprise training courseware.

[0142] The storage medium can be any computer-readable storage medium that can store program code, such as a USB flash drive, external hard drive, read-only memory (ROM), magnetic disk, or optical disk.

[0143] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0144] In the several embodiments provided by this invention, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For example, the division of each unit is merely a logical functional division, and there may be other division methods in actual implementation. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed.

[0145] The steps in the method of this invention can be adjusted, merged, or reduced in order according to actual needs. The units in the device of this invention can be merged, divided, or reduced according to actual needs. Furthermore, the functional units in the various embodiments of this invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0146] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, a terminal, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention.

[0147] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. An artificial intelligence-based enterprise courseware intelligent generation method, characterized in that, include: In response to a courseware intelligent generation command sent by a user terminal, an initial requirement dataset corresponding to the courseware intelligent generation command is obtained; wherein, the initial requirement dataset includes the initial user basic information of the training instructor and the basic information of the training task; the basic information of the training task includes the specific training scenario, the type of training audience, and the initial input data of the courseware. The initial requirement dataset is input into the first pre-trained agent to decompose the courseware topic selection and courseware generation tasks, and the current courseware topic output information and the corresponding current courseware generation task decomposition information are obtained; wherein, the current courseware generation task decomposition information includes the current courseware chapter requirement information, the current courseware section requirement information and the current courseware content requirement information; The current courseware generation task decomposition information is input into the second pre-trained agent to generate courseware content, and the current courseware output dataset is obtained. The current courseware output dataset is input into the third pre-trained agent for content review, and the reviewed courseware output dataset is obtained and sent to the user terminal. The step of inputting the current courseware generation task decomposition information into the second pre-trained agent to generate courseware content and obtain the current courseware output dataset includes: The current courseware chapter requirement information in the current courseware generation task decomposition information is input into the chapter generation agent in the second pre-trained agent to obtain the current chapter output data; The current courseware section requirement information in the current courseware generation task decomposition information is input into the section generation agent in the second pre-trained agent to obtain the current section output data; The current courseware content requirement information in the current courseware generation task decomposition information is input into the content generation agent in the second pre-trained agent to obtain the current courseware content output data. Based on the target sections to which the current courseware content output sub-data belongs in the current section output data and the target chapters to which they belong in the current chapter output data, each current courseware content output sub-data is grouped into the target sections of the corresponding target chapters to obtain the current courseware output dataset.

2. The method according to claim 1, characterized in that, The step of inputting the initial demand dataset into the first pre-trained agent to decompose the courseware topic selection and courseware generation task, and obtaining the current courseware topic output information and the corresponding current courseware generation task decomposition information, includes: If it is determined that the initial input data of the courseware in the initial demand dataset is a brief text type or an empty set, then the initial demand dataset is input into the first pre-trained agent for courseware topic recommendation processing to obtain the current courseware topic output information; If it is determined that the initial input data of the courseware in the initial demand dataset is of the detailed text material type, then the initial input data of the courseware is input into the first pre-trained agent for courseware topic recommendation processing to obtain the current courseware topic output information; If a selection instruction is detected for one of the current courseware topic selection results in the current courseware topic selection output information, then the current target courseware topic selection result is obtained accordingly. The current target courseware topic selection results and the initial requirement dataset are input into the first pre-trained agent to decompose the courseware generation task, thereby obtaining the current courseware generation task decomposition information.

3. The method according to claim 2, characterized in that, The first pre-trained agent includes a topic recommendation agent and a task decomposition agent, and the topic recommendation agent and the task decomposition agent adopt a chain topology structure. The step of inputting the initial demand dataset into the first pre-trained agent for courseware topic recommendation processing to obtain the current courseware topic output information includes: The initial demand dataset is input into the topic recommendation agent in the first pre-trained agent to perform courseware topic recommendation processing, and the current courseware topic output information is obtained. The initial input data of the courseware is input into the first pre-trained agent for courseware topic recommendation processing, resulting in the current courseware topic output information, including: The initial input data of the courseware is input into the topic recommendation agent in the first pre-trained agent to perform courseware topic recommendation processing, and the current courseware topic output information is obtained. The step of inputting the current target courseware topic selection results and the initial requirement dataset into the first pre-trained agent to decompose the courseware generation task, and obtaining the current courseware generation task decomposition information, includes: The current target courseware topic selection results, the initial demand dataset, and the obtained actual training course demand information of the training audience are input into the task decomposition agent in the first pre-trained agent to obtain the current courseware generation task decomposition information.

4. The method according to claim 1, characterized in that, The step of inputting the current courseware content requirement information from the current courseware generation task decomposition information into the content generation agent in the second pre-trained agent to obtain the current courseware content output data includes: The current courseware content requirement information in the current courseware generation task decomposition information is input into the content generation agent, so that the content generation agent can obtain courseware content from the locally connected knowledge base or network database and generate the current courseware content output data.

5. The method according to claim 1, characterized in that, The step of inputting the current courseware output dataset into a third pre-trained agent for content review, resulting in a reviewed courseware output dataset, includes: The current courseware output dataset is input into the third pre-trained agent to detect whether there are grammatical errors in the courseware content and whether it meets the preset courseware content constraints through the enterprise-specific review rules corresponding to the third pre-trained agent. Modification suggestion data corresponding to the current courseware output dataset is obtained and combined with the current courseware output dataset to form the reviewed courseware output dataset.

6. The method according to any one of claims 1-5, characterized in that, After the step of inputting the current courseware output dataset into the third pre-trained agent for content review, obtaining the reviewed courseware output dataset, and sending it to the user terminal, the method further includes: If the final courseware output dataset corresponding to the approved courseware output dataset is received, the final courseware output dataset is input into the fourth pre-trained agent to generate training speech and training script, thereby obtaining the current training courseware speech data and the current training courseware script data.

7. An intelligent enterprise courseware generation device based on artificial intelligence, characterized in that, It includes a unit for performing the AI-based intelligent generation method for enterprise courseware as described in any one of claims 1-6.

8. A computer device, characterized in that, The computer device includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, it implements the intelligent generation method for enterprise courseware based on artificial intelligence as described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which includes program instructions that, when executed by a processor, can implement the AI-based intelligent generation method for enterprise courseware as described in any one of claims 1-6.

Images