Education industry content recommendation method and device based on large language model

Abstract

The invention provides an education industry content recommendation method and device based on a large language model. The method comprises the steps of obtaining user information of each user and to-be-processed education data; processing the education data by using a pre-trained large language model to obtain a plurality of entity triples; constructing a knowledge graph according to the plurality of entity triads; constructing a user portrait of each user according to the user information of each user; according to the knowledge graph and the user portrait, generating recommended education content corresponding to each user; and sending the recommended education content corresponding to each user to each user. By applying the method provided by the embodiment of the invention, the required education content can be accurately recommended to the user.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a method and device for recommending content in the education industry based on a large language model. Background Art

[0002] With the development of Internet technology and the popularization of online education, personalized recommendation of educational content has become one of the important means to improve learning effects. Traditional educational content recommendation systems mainly rely on static data and simple learning records for recommendation.

[0003] However, this recommendation method often fails to accurately capture the real needs of learners, resulting in insufficient personalization of recommendation results and poor learning experience. Summary of the invention

[0004] The technical problem to be solved by this application is to provide a method and device for recommending content in the education industry based on a large language model, which can accurately recommend the educational content required by users. The specific solution is as follows:

[0005] A method for recommending content in the education industry based on a large language model, the method comprising:

[0006] Obtain user information of each user and educational data to be processed;

[0007] Processing the education data using a pre-trained large language model to obtain a plurality of entity triples;

[0008] Construct a knowledge graph according to the multiple entity triples;

[0009] Building a user profile for each user based on the user information of each user;

[0010] Generate recommended educational content corresponding to each of the users based on the knowledge graph and the user portrait;

[0011] The recommended educational content corresponding to each of the users is sent to each of the users.

[0012] In the above method, optionally, the educational data is processed using a preset large language model to obtain a plurality of entity triples, including:

[0013] The education data is input into the large language model, so that when the education data is not text data, the large language model converts the education data into text content, and performs semantic analysis on the text content to obtain multiple entity triples.

[0014] In the above method, optionally, generating recommended educational content corresponding to each user according to the knowledge graph and the user portrait includes:

[0015] Acquire learning behavior data of each user, wherein the learning behavior includes at least one of learning time, completion degree, interaction information, and feedback information;

[0016] Determine preference information of each user according to the learning behavior data of each user;

[0017] Recommended educational content is generated for each user based on the preference information, user portrait and knowledge graph of each user.

[0018] In the above method, optionally, sending the recommended educational content corresponding to each user to each user includes:

[0019] Generating personalized learning path information for each user based on the recommended educational content for each user;

[0020] The learning path information of each of the users is sent to each of the users.

[0021] The above method, optionally, after sending the learning path information of each user to each user, further comprises:

[0022] Acquire learning feedback information of each user; the learning feedback information of each user includes at least one of learning behavior, learning performance, user satisfaction and content suitability corresponding to the learning path of each user;

[0023] The knowledge graph is updated according to the learning feedback information of each of the users to generate new recommended educational content according to the updated knowledge graph.

[0024] A content recommendation device for the education industry based on a large language model, comprising:

[0025] A first acquisition unit, used to acquire user information of each user and education data to be processed;

[0026] An execution unit, configured to process the education data using a pre-trained large language model to obtain a plurality of entity triples;

[0027] A first construction unit, configured to construct a knowledge graph according to the plurality of entity triples;

[0028] A second construction unit, configured to construct a user portrait of each user according to the user information of each user;

[0029] A generating unit, configured to generate recommended educational content corresponding to each of the users according to the knowledge graph and the user portrait;

[0030] The sending unit is used to send the recommended educational content corresponding to each of the users to each of the users.

[0031] In the above device, optionally, the execution unit includes:

[0032] An input subunit is used to input the educational data into the large language model, so that when the educational data is not text data, the large language model converts the educational data into text content and performs semantic analysis on the text content to obtain multiple entity triples.

[0033] In the above device, optionally, the generating unit comprises:

[0034] An acquisition subunit, configured to acquire learning behavior data of each user, wherein the learning behavior includes at least one of learning time, completion degree, interaction information, and feedback information;

[0035] a determination subunit, configured to determine preference information of each of the users according to the learning behavior data of each of the users;

[0036] The first generating subunit is used to generate recommended educational content for each user based on the preference information, user portrait and the knowledge graph of each user.

[0037] In the above device, optionally, the sending unit includes:

[0038] A second generating subunit, configured to generate personalized learning path information for each user according to the recommended educational content for each user;

[0039] The sending subunit is used to send the learning path information of each user to each user.

[0040] The above device may optionally further include:

[0041] A second acquisition unit is used to acquire learning feedback information of each user; the learning feedback information of each user includes at least one of the learning behavior, learning performance, user satisfaction and content suitability corresponding to the learning path of each user;

[0042] An updating unit is used to update the knowledge graph according to the learning feedback information of each user, so as to generate new recommended educational content according to the updated knowledge graph.

[0043] Based on the implementation of the present application, a method and system for recommending content in the education industry based on a large language model is provided, the method comprising: obtaining user information of each user and the education data to be processed; processing the education data using a pre-trained large language model to obtain a plurality of entity triples; constructing a knowledge graph based on the plurality of entity triples; constructing a user portrait of each user based on the user information of each user; generating recommended education content corresponding to each user based on the knowledge graph and the user portrait; and sending the recommended education content corresponding to each user to each user. Applying the method provided in the embodiment of the present application, it is possible to accurately recommend the education content required by the user. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0045] Figure 1 A method flow chart of a method for recommending content in the education industry based on a large language model provided in this application;

[0046] Figure 2 A flowchart of a process for generating recommended educational content corresponding to each user provided by the present application;

[0047] Figure 3 A schematic diagram of the structure of a content recommendation device for the education industry based on a large language model provided in this application;

[0048] Figure 4 A schematic diagram of the structure of an electronic device provided in this application. DETAILED DESCRIPTION

[0049] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0050] In this application, the terms "comprises", "comprising" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprising a ..." does not exclude the presence of other identical elements in the process, method, article or device comprising the element.

[0051] The embodiment of the present application provides a method for recommending content in the education industry based on a large language model. The method can be applied to electronic devices. The method flow chart of the method is as follows: Figure 1 As shown, specifically including:

[0052] S101: Obtain user information of each user and educational data to be processed.

[0053] In this embodiment, the user may be the object of receiving education.

[0054] Optionally, the user information of the user may include data such as basic user information, learning style, ability level, interest points, etc. The basic user information may include user identification and age, etc.

[0055] In this embodiment, the educational data may be one or more of structured text, unstructured text, images, audio, video, etc. in the education industry.

[0056] S102: Using a pre-trained large language model to process the education data, to obtain a plurality of entity triples.

[0057] In this embodiment, the entity triple may include an entity, a target entity associated with the entity or an attribute of the entity, and an association relationship between the entity and the target entity or the attribute.

[0058] In one embodiment provided in the present application, the process of processing the education data using a preset large language model to obtain multiple entity triples includes: inputting the education data into the large language model so that the large language model converts the education data into text content when the education data is not text data, and performing semantic analysis on the text content to obtain multiple entity triples.

[0059] Optionally, after the educational data is input into the large language model, when the educational data is image data, image recognition technology is used to identify the image data to obtain image recognition results, which include at least one of objects and scenes in the image data, and the image recognition results are converted into understandable text content; when the educational data is video content, frame analysis and time series processing can be performed on the video content to extract key frames and actions in the video content, and the key frames and actions are converted into structured text content; after obtaining the text content, semantic analysis is performed on the text content to obtain multiple entity triples.

[0060] S103: Construct a knowledge graph based on the multiple entity triples.

[0061] In this embodiment, knowledge graph construction technology can be used to generate a knowledge graph based on the extracted entity triples. The knowledge graph includes multiple nodes, each node represents a knowledge point or concept, and the edges between nodes represent the relationship between the nodes.

[0062] In some embodiments, logical and semantic connections between knowledge points may be determined, such as causal logic, conditional logic, and contrastive logic.

[0063] S104: Construct a user profile of each user according to the user information of each user.

[0064] In this embodiment, a user profile of each user may be constructed based on data such as basic information, learning style, ability level, and interests in the user information of the user.

[0065] S105: Generate recommended educational content corresponding to each of the users based on the knowledge graph and the user portrait.

[0066] In this embodiment, the user's interest tags, learning style, ability level and other data can be obtained from the user portrait, and the interest tags, learning style, ability level and other data can be converted into query statements, which can be suitable for searching in the knowledge graph; the query statement is used to perform a search in the knowledge graph to retrieve educational resources that match the user portrait information; the retrieved educational resources are sorted, and the sorting basis includes at least one of the relevance score of the resource, the user's previous response score to similar resources, and the freshness score of the resource; the top N educational resources after sorting are selected as recommended educational content.

[0067] S106: Send the recommended educational content corresponding to each of the users to each of the users.

[0068] In this embodiment, the recommended educational content corresponding to each of the users is sent to each of the users, including: generating personalized learning path information for each of the users based on the recommended educational content for each of the users; and sending the learning path information for each of the users to each of the users.

[0069] In this embodiment, the user's personalized learning path information is generated based on the user's recommended educational content, current learning progress, learning goals and user preferences; the personalized learning path information learning path information includes a series of learning tasks arranged by priority, and each learning task contains recommended educational resources and their target completion time.

[0070] Optionally, the personalized learning path information also includes evaluation criteria for measuring the user's mastery level after completing the learning task.

[0071] In this embodiment, the generated personalized learning path information can be stored in a user account database and associated with a corresponding user identifier; the personalized learning path information can be sent to an application on the user's mobile device through electronic communication means; the user accesses the personalized learning path information by logging into a personal account and starts to perform learning tasks according to the instructions.

[0072] Optionally, the user's performance can be continuously monitored, and the personalized learning path information can be updated based on the user's feedback or behavior.

[0073] In an embodiment provided in the present application, based on the above solution, optionally, a process of generating recommended educational content corresponding to each user according to the knowledge graph and the user portrait is as follows: Figure 2 As shown, including:

[0074] S201: Acquire learning behavior data of each user, where the learning behavior includes at least one of learning time, completion degree, interaction information, and feedback information.

[0075] In this embodiment, the user's login time, access frequency, length of stay, etc. are recorded, and the user's learning time is determined based on the user's login time, access frequency, and length of stay.

[0076] Optionally, record the user's progress in the course, such as chapter completion, assignment submission status, etc., and determine the degree of completion based on the user's progress in the course, such as chapter completion, assignment submission status, etc.

[0077] In this embodiment, the user's interactions on the platform, such as watching videos, reading articles, participating in discussions, etc., are recorded to obtain the user's interaction information.

[0078] Optionally, record users' evaluations of courses, teachers, or other user-generated content to obtain feedback information.

[0079] S202: Determine preference information of each user according to the learning behavior data of each user.

[0080] In this embodiment, a clustering algorithm can be used to cluster the learning behavior data of each user to obtain the user grouping results, and a personalized label is assigned to each user based on the grouping results. The personalized label at least reflects the user's preferred course type, learning time preference, and learning method preference. The personalized label is the user's preference information.

[0081] S203: Generate recommended educational content for each user based on the preference information, user portrait and knowledge graph of each user.

[0082] In this embodiment, preference information and user portrait information can be used to retrieve educational resources that are highly relevant to the user's learning needs in the knowledge graph, and then the retrieved educational resources are sorted. The sorting criteria include at least one of the resource's relevance score, the user's previous response score to similar resources, and the resource's freshness score. The top N educational resources after sorting are selected as recommended educational content.

[0083] In an embodiment provided by the present application, based on the above solution, optionally, after sending the learning path information of each user to each user, the method further includes:

[0084] Acquire learning feedback information of each user; the learning feedback information of each user includes at least one of learning behavior, learning performance, user satisfaction and content suitability corresponding to the learning path of each user;

[0085] The knowledge graph is updated according to the learning feedback information of each of the users to generate new recommended educational content according to the updated knowledge graph.

[0086] In this embodiment, based on the learning feedback information, the relationship between the corresponding nodes and edges in the knowledge graph is updated, and the weight values ​​of the educational resources are adjusted; the relevance scores of the educational resources are recalculated using the updated knowledge graph; new recommended educational content is generated based on the updated knowledge graph and the new scoring results; the new recommended educational content is presented to the user, and further feedback from the user is collected; and the above process is continuously iterated to continuously optimize the recommendation results.

[0087] In one embodiment provided in the present application, there is also provided an education industry content recommendation system based on a large language model, including a first module, a second module, a third module and a fourth module; the functions of each module are as follows:

[0088] The first module will receive content inputs in various formats from the education field, including but not limited to text, images and video materials; then, it will pre-process the content, including removing noise, standardizing formats and languages; then, it will apply natural language processing technology and use advanced large language models to perform deep semantic analysis of the text content, identify key concepts, entities and their attributes, and extract information such as people, places, and organizations in the text through named entity recognition (NER) technology; in addition, image recognition technology will be used to analyze image content, identify objects and scenes in the image, and convert them into understandable labels; video content will be converted into structured data through frame analysis and time series processing to extract key frames and actions.

[0089] Furthermore, the first module integrates the extracted information into a graph form through knowledge graph construction technology; in this process, knowledge point association and mapping technology will analyze and determine the logical and semantic connections between knowledge points, such as cause and effect, conditions, comparison, etc., to ensure the richness and accuracy of the knowledge graph; the dynamic update and maintenance mechanism of the knowledge graph will ensure that the graph can continue to evolve with the addition of new content, while correcting errors and optimizing content through user feedback and expert review.

[0090] In addition, the semantic representation of content will be achieved through vectorization technology, which will convert text, image and video content into vector form to facilitate efficient similarity comparison and retrieval; deep analysis technology will identify themes and assess the difficulty of educational content, providing appropriate learning materials for students at different learning stages; interactive content understanding allows users to interact with the system, and through user queries and feedback, the system can further optimize the accuracy of content parsing; finally, security and privacy protection measures will ensure that the security and privacy of user data are fully protected during content analysis and processing.

[0091] In this embodiment, the first module will be able to deeply understand the educational content and build a comprehensive, dynamic and accurate knowledge graph.

[0092] The second module can first conduct user portrait analysis, and build a unique portrait of each student by collecting data such as students' basic information, learning style, ability level and interests; then, use learning history tracking technology to record and analyze students' learning behaviors on the platform, including learning time, completion, interaction and feedback, etc., to evaluate students' learning effects and preferences; then, apply RAG technology, combined with knowledge graphs and user portraits, to retrieve and recommend educational content and resources that match students' learning needs and interests; further, generate personalized learning paths, dynamically adjust learning plans according to students' learning history and current status, and ensure the consistency and adaptability of learning content; at the same time, design a dynamic adjustment mechanism for learning paths, and adjust learning plans in real time according to students' learning progress and feedback to adapt to students' changes; in addition, realize the visualization of learning paths so that students can clearly understand their learning progress and future learning plans; finally, ensure the scalability and flexibility of learning path planning to adapt to the needs of different subjects, different syllabuses and different educational stages.

[0093] During the process of generating personalized learning paths, the system will continuously collect students' learning data, and through data analysis and mining, continuously optimize user portraits and learning path recommendation algorithms to improve the accuracy and effectiveness of personalized recommendations; at the same time, the system will provide a feedback mechanism to encourage students to make suggestions and evaluations on learning content and paths to further enhance the learning experience and system performance.

[0094] In addition, the system will take into account students' individual differences, such as cognitive ability, learning motivation and emotional state, and adjust learning content and methods through emotional analysis and motivational incentive technology to stimulate students' learning interest and motivation; at the same time, the system will follow the principles of educational psychology and cognitive science to ensure the scientificity and effectiveness of the learning path.

[0095] In this embodiment, the second module will be able to provide each student with a tailor-made study plan, achieving truly personalized education and helping students learn and grow more effectively.

[0096] The third module can first identify the structure and semantics of educational content based on the content generation algorithm of the artificial intelligence large language model, and generate high-quality text, image and video materials; then, integrate adaptive learning technology to dynamically adjust the difficulty and complexity of generated content according to students' learning progress and feedback, ensuring that the content always matches students' abilities; in addition, through multimedia content adaptation technology, educational content is generated so that the generated educational content can adapt to different learning scenarios and devices, providing rich visual and auditory experience;

[0097] Furthermore, a real-time feedback mechanism will be established to collect students' interactive data during the learning process, such as the accuracy of answering questions, learning time, user satisfaction, etc., to guide the further generation and optimization of content; personalized educational content will be customized according to students' learning history and preferences, including personalized exercises, case studies and interactive simulations; ensure the diversity and innovation of content generation, avoid content repetition and monotony, and stimulate students' interest and participation in learning.

[0098] Synchronize with the personalized learning path generated by the second module to ensure that content generation is consistent with students' learning goals and stages; implement strict content quality control and review processes to ensure that the generated educational content is accurate and meets educational standards; integrate various artificial intelligence technologies such as natural language processing, image recognition and video analysis to continuously optimize content generation algorithms and improve the relevance and attractiveness of content.

[0099] Display a user interface that enables students to easily access and use the generated educational content, while providing personalized recommendations and prompts; during the content generation and adaptation process, protect the security and privacy of student data and comply with relevant laws and regulations; ensure the scalability of the system to support the growing number of users and diverse educational needs; the system will continue to learn the latest educational theories and technological developments, and continuously iterate and improve content generation algorithms.

[0100] Finally, a feedback loop mechanism is designed, and the system will continuously collect students' usage feedback and learning outcomes to further optimize the content generation logic and form a closed-loop adaptive learning system to achieve the best educational results and learning experience.

[0101] The fourth module uses RAG technology to retrieve and recommend supplementary materials and exercises based on students' learning performance and historical data, combined with the educational content generated in the third module; then, it implements a learning effect evaluation mechanism to track students' learning behavior and grades, evaluate learning effects, and adjust recommendation strategies in a timely manner; then, it builds a feedback loop mechanism to collect students' feedback on recommended content, including satisfaction, difficulty appropriateness, etc., in order to optimize the recommended educational content.

[0102] Furthermore, the recommended content is displayed through the user interface so that students can easily access the recommended materials, and feedback channels are provided so that students can provide feedback; data security and privacy protection measures are integrated to ensure the security and privacy of student data, and relevant laws and regulations are complied with when collecting and processing feedback information.

[0103] In this embodiment, the performance and accuracy of the recommendation system are improved by utilizing artificial intelligence technologies such as natural language processing, image recognition, and video analysis; and the recommendation algorithm and learning effect evaluation mechanism are continuously iterated and improved by continuously learning the latest educational concepts and technological developments.

[0104] Finally, the knowledge graph is optimized, and the results of learning effect evaluation and student feedback are fed back to Module 1 to optimize the knowledge graph and improve the accuracy and richness of knowledge representation; the learning path is adjusted, and the learning path is dynamically adjusted according to the learning effect evaluation and student feedback to ensure that the learning plan matches the students' actual learning progress and abilities.

[0105] By implementing the above technical solutions, Module 4 will be able to provide students with a personalized and efficient learning experience, promote their in-depth learning and understanding, and provide educators with valuable feedback and data support.

[0106] and Figure 1 Corresponding to the method described above, the embodiment of the present application also provides an educational industry content recommendation device based on a large language model, which is applied to an electronic device for recommending content in an educational industry. Figure 3 The specific implementation of the method in the embodiment of the present invention is shown in the structural diagram of the device. Figure 3 As shown, including:

[0107] A first acquisition unit 301 is used to acquire user information of each user and education data to be processed;

[0108] An execution unit 302 is used to process the education data using a pre-trained large language model to obtain a plurality of entity triples;

[0109] A first construction unit 303, configured to construct a knowledge graph according to the plurality of entity triples;

[0110] A second constructing unit 304 is used to construct a user portrait of each user according to the user information of each user;

[0111] A generating unit 305, configured to generate recommended educational content corresponding to each of the users according to the knowledge graph and the user portrait;

[0112] The sending unit 306 is used to send the recommended educational content corresponding to each of the users to each of the users.

[0113] In an embodiment provided in the present application, based on the above solution, optionally, the execution unit 302 includes:

[0114] An input subunit is used to input the educational data into the large language model, so that when the educational data is not text data, the large language model converts the educational data into text content and performs semantic analysis on the text content to obtain multiple entity triples.

[0115] In an embodiment provided in the present application, based on the above solution, optionally, the generating unit 305 includes:

[0116] An acquisition subunit, configured to acquire learning behavior data of each user, wherein the learning behavior includes at least one of learning time, completion degree, interaction information, and feedback information;

[0117] a determination subunit, configured to determine preference information of each of the users according to the learning behavior data of each of the users;

[0118] The first generating subunit is used to generate recommended educational content for each user based on the preference information, user portrait and the knowledge graph of each user.

[0119] In an embodiment provided in the present application, based on the above solution, optionally, the sending unit 306 includes:

[0120] A second generating subunit, configured to generate personalized learning path information for each user according to the recommended educational content for each user;

[0121] The sending subunit is used to send the learning path information of each user to each user.

[0122] In an embodiment provided in the present application, based on the above solution, optionally, it further includes:

[0123] A second acquisition unit is used to acquire learning feedback information of each user; the learning feedback information of each user includes at least one of the learning behavior, learning performance, user satisfaction and content suitability corresponding to the learning path of each user;

[0124] An updating unit is used to update the knowledge graph according to the learning feedback information of each user, so as to generate new recommended educational content according to the updated knowledge graph.

[0125] The specific principles and execution processes of each unit and module in the education industry content recommendation device based on a large language model disclosed in the above-mentioned embodiment of the present application are the same as the education industry content recommendation method based on a large language model disclosed in the above-mentioned embodiment of the present application. Please refer to the corresponding parts of the education industry content recommendation method based on a large language model provided in the above-mentioned embodiment of the present application, and no further details will be given here.

[0126] An embodiment of the present application also provides a storage medium, which includes stored instructions, wherein when the instructions are executed, the device where the storage medium is located is controlled to execute the above-mentioned education industry content recommendation method based on a large language model.

[0127] The present application also provides an electronic device, whose structural diagram is as follows: Figure 4As shown, it specifically includes a memory 401 and one or more instructions 402, wherein the one or more instructions 402 are stored in the memory 401 and are configured to be executed by one or more processors 403 to perform the following operations:

[0128] Obtain user information of each user and educational data to be processed;

[0129] Processing the education data using a pre-trained large language model to obtain a plurality of entity triples;

[0130] Construct a knowledge graph according to the multiple entity triples;

[0131] Building a user profile for each user based on the user information of each user;

[0132] Generate recommended educational content corresponding to each of the users based on the knowledge graph and the user portrait;

[0133] The recommended educational content corresponding to each of the users is sent to each of the users.

[0134] It should be noted that each embodiment in this specification is described in a progressive manner, and each embodiment focuses on the differences from other embodiments, and the same or similar parts between the embodiments can be referred to each other. For the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0135] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are merely used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations.

[0136] For the convenience of description, the above device is described in various units according to their functions. Of course, when implementing the present application, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0137] It can be known from the description of the above implementation methods that those skilled in the art can clearly understand that the present application can be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present application can be essentially or partly contributed to the prior art in the form of a software product, which can be stored in a storage medium such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes several instructions for enabling a computer device (which can be a personal computer, a server, etc.) to execute the methods described in the various embodiments of the present application or certain parts of the embodiments.

[0138] The above is a detailed introduction to the content recommendation method for the education industry based on a large language model provided by the present application. This article uses specific examples to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea; at the same time, for general technical personnel in this field, according to the idea of ​​the present application, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims

1. A content recommendation method for the education industry based on a large language model, characterized in that: include: Obtain user information of each user and educational data to be processed; Processing the education data using a pre-trained large language model to obtain a plurality of entity triples; Construct a knowledge graph according to the multiple entity triples; Building a user profile for each user based on the user information of each user; Generate recommended educational content corresponding to each of the users based on the knowledge graph and the user portrait; The recommended educational content corresponding to each of the users is sent to each of the users.

2. The method according to claim 1, characterized in that The educational data is processed by using a preset large language model to obtain a plurality of entity triples, including: The education data is input into the large language model, so that when the education data is not text data, the large language model converts the education data into text content, and performs semantic analysis on the text content to obtain multiple entity triples.

3. The method according to claim 1, characterized in that Generating recommended educational content corresponding to each user according to the knowledge graph and the user portrait includes: Acquire learning behavior data of each user, wherein the learning behavior includes at least one of learning time, completion degree, interaction information, and feedback information; Determine preference information of each user according to the learning behavior data of each user; Recommended educational content is generated for each user based on the preference information, user portrait and knowledge graph of each user.

4. The method according to claim 1, characterized in that: The sending the recommended educational content corresponding to each of the users to each of the users includes: Generating personalized learning path information for each user based on the recommended educational content for each user; The learning path information of each of the users is sent to each of the users.

5. The method according to claim 4, characterized in that After sending the learning path information of each user to each user, the method further includes: Acquire learning feedback information of each user; the learning feedback information of each user includes at least one of learning behavior, learning performance, user satisfaction and content suitability corresponding to the learning path of each user; The knowledge graph is updated according to the learning feedback information of each of the users to generate new recommended educational content according to the updated knowledge graph.

6. A content recommendation device for the education industry based on a large language model, characterized in that: include: A first acquisition unit, used to acquire user information of each user and education data to be processed; An execution unit, configured to process the education data using a pre-trained large language model to obtain a plurality of entity triples; A first construction unit, configured to construct a knowledge graph according to the plurality of entity triples; A second construction unit, configured to construct a user portrait of each user according to the user information of each user; A generating unit, configured to generate recommended educational content corresponding to each of the users according to the knowledge graph and the user portrait; The sending unit is used to send the recommended educational content corresponding to each of the users to each of the users.

7. The device according to claim 6, characterized in that The execution unit comprises: An input subunit is used to input the educational data into the large language model, so that when the educational data is not text data, the large language model converts the educational data into text content and performs semantic analysis on the text content to obtain multiple entity triples.

8. The device according to claim 6, characterized in that The generating unit comprises: An acquisition subunit, configured to acquire learning behavior data of each user, wherein the learning behavior includes at least one of learning time, completion degree, interaction information, and feedback information; a determination subunit, configured to determine preference information of each of the users according to the learning behavior data of each of the users; The first generating subunit is used to generate recommended educational content for each user based on the preference information, user portrait and the knowledge graph of each user.

9. The device according to claim 6, characterized in that The sending unit comprises: A second generating subunit, configured to generate personalized learning path information for each user according to the recommended educational content for each user; The sending subunit is used to send the learning path information of each user to each user.

10. The device according to claim 9, characterized in that Also includes: A second acquisition unit is used to acquire learning feedback information of each user; the learning feedback information of each user includes at least one of the learning behavior, learning performance, user satisfaction and content suitability corresponding to the learning path of each user; An updating unit is used to update the knowledge graph according to the learning feedback information of each user, so as to generate new recommended educational content according to the updated knowledge graph.

Images