Online teaching programming system and teaching recommendation method based on large language model assistance

Abstract

The application discloses an online programming system and a teaching recommendation method based on a large language model, and the system comprises a base model acquisition module, a problem decomposition module, an abstract modeling module, an algorithm design module, a code analysis and assistance module and an evaluation module, is used for helping students to complete programming tasks of teaching design online, decomposes, abstracts and designs algorithms for actual problems by using a large language model, and answers problems encountered by students in programming. Based on the above system, the application also designs a teaching recommendation method, collects multi-modal data in the classroom teaching process, establishes a dynamic three-level teaching interaction graph of a space-time environment, knowledge resources and cognitive behavior through human-computer, student-student and teacher-student interaction behaviors, and provides accurate teaching recommendation and intervention strategies in a human-in-the-loop. The application establishes a three-dimensional comprehensive teaching field of learning environment intelligent network computing in a programming scene, and provides guarantee for the development of high-quality modern education.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of artificial intelligence and recommendation algorithm, and particularly relates to an online programming teaching system based on large language model assistance and a teaching recommendation method. BACKGROUND

[0002] With the advent of the era of artificial intelligence, the New Generation Artificial Intelligence Development Plan proposes to establish a learner-centered education environment, provide precise education services, and realize the customization of daily education and lifelong education. Subsequently, the hot development of the cross-field of AI + education, the tool driven by artificial intelligence is a major trend in programming education, such as chat robots, automatic code evaluation tools, etc. Skalka et al. proposed a conceptual framework that combines microlearning and automatic source code evaluation, aiming to provide immediate feedback for students. Malik et al. introduced a specially designed chat robot, aiming to highlight problem-solving strategies, common programming errors, syntax and semantics, and ultimately to help inexperienced learners master multiple skills at the same time.

[0003] However, although the above methods have achieved good results in improving students' programming ability, they often do not pay attention to the cultivation of students' comprehensive quality and thinking ability. At the same time, these tools often focus on a certain stage of programming skills, and cannot track students' multi-modal learning data throughout the whole process, resulting in insufficient continuity and comprehensiveness of teaching support. In addition, the existing methods fail to fully consider the differences between individual learners, lack of mechanisms for teaching adjustment and resource recommendation based on individual learning states, making the allocation of educational resources and teaching intervention strategies unable to achieve true individualization and precision.

[0004] Therefore, it is urgent to develop an integrated online programming teaching system that can comprehensively track students' learning progress, provide continuous support from the beginner to the advanced stage, and generate personalized learning paths and teaching recommendations through dynamic data analysis, truly realizing the learner-centered education mode and promoting the development of programming education towards higher quality and modernization. SUMMARY

[0005] The present application aims to solve the problems existing in the current primary and secondary school information classroom teaching, and provides an online teaching programming system based on large language model assistance and a teaching recommendation method. The method of the present application can timely solve the problems encountered by students during programming, and provide precise teaching recommendation and intervention strategies for students.

[0006] In order to achieve the above-mentioned application purpose, the present application specifically adopts the following technical solutions:

[0007] In a first aspect, the present application provides an online teaching programming system based on large language model assistance, comprising:

[0008] a base model obtaining module configured to obtain a general large language model and fine-tune the general large language model to become a specialized base model for the programming teaching field;

[0009] a problem decomposition module configured to obtain a first user input and convert the first user input into a designed first prompt, input a question to be decomposed and the first prompt into the specialized base model, output a problem decomposition result from the specialized base model, and display the problem decomposition result on a user interface of an online teaching programming system;

[0010] an abstract modeling module configured to obtain a second user input and convert the second user input into a designed second prompt, input a question to be modeled and the second prompt into the specialized base model, output an abstract modeling result from the specialized base model, and display the abstract modeling result on the user interface;

[0011] an algorithm design module configured to obtain a third user input and convert the third user input into a designed third prompt, input a question to be designed and the third prompt into the specialized base model, output an algorithm design result from the specialized base model, and display the algorithm design result on the user interface;

[0012] a code analysis and assistance module configured to intelligently analyze a code completed by a user, obtain error information appearing on a console, input the code completed by the user and the error information into the specialized base model, output an analysis result and a correction suggestion from the specialized base model, and display the analysis result and the correction suggestion on the user interface;

[0013] an evaluation module configured to intelligently evaluate a code submitted by a user, input the code submitted by the user, a standard answer of a programming question, and a scoring rule into the specialized base model, output an evaluation result of the code submitted by the user from the specialized base model, and display the evaluation result on the user interface.

[0014] As a preferred embodiment of the first aspect, in the base model obtaining module, the specific process of fine-tuning the large language model is as follows: the LoRA model is used to fine-tune the large language model, and the pre-designed restriction words are used to constrain the output content and language style of the large language model; a pre-constructed information technology course knowledge base is obtained; the information technology course knowledge base and the LoRA fine-tuned large language model are used as inputs of a large model knowledge chain framework to obtain a hallucination-reduced large language model, which is used as the specialized base model.

[0015] As a preferred embodiment of the first aspect, in the problem decomposition module, the first prompt contains an output format requirement and a word limit of the specialized base model, and indicates that the problem decomposition needs to be completed by the specialized base model.

[0016] As a preferred embodiment of the first aspect, in the abstract modeling module, the second prompt word contains the output of the specialized base model as a calculation model and a word limit, and indicates that the specialized base model needs to provide an explanation of the modeling process.

[0017] As a preferred embodiment of the first aspect, in the algorithm design module, the third prompt word contains the output of the specialized base model as an algorithm flowchart and a flowchart format requirement, and indicates that the specialized base model needs to provide an explanation of the algorithm design process.

[0018] As a preferred embodiment of the first aspect, the code analysis and assistance module includes a code analysis module and an assistance debugging module.

[0019] The code analysis module is used to obtain the code completed by the user, input the code completed by the user into the specialized base model, output the analysis results of the code from the whole to the line by the specialized base model, and display the analysis results on the user interface.

[0020] The assistance debugging module is used to obtain error information output by the console, input the error information into the specialized base model, output error explanation and correction suggestions of the error information by the specialized base model, and display the error explanation and correction suggestions of the error information on the user interface.

[0021] As a preferred embodiment of the first aspect, in the evaluation module, the specialized base model weights the execution time and the code analysis result of the user-submitted code to comprehensively evaluate the efficiency and quality of the user-submitted code, and finally outputs an evaluation result of the user-submitted code.

[0022] In a second aspect, the present application provides a teaching recommendation method based on the online teaching programming system of the first aspect, which includes the following steps:

[0023] S1, obtaining multi-modal data in the online classroom teaching process, including human-computer interaction data, student-student interaction data and teacher-student interaction data;

[0024] S2, data cleaning is performed on the multi-modal data, and the multi-modal data after data cleaning is labeled to label human-computer interaction behavior, student-student interaction behavior and teacher-student interaction behavior, and to obtain processed multi-modal data;

[0025] S3, constructing a dynamic three-level teaching interaction graph coupled with knowledge resources and interaction behaviors based on the processed multi-modal data;

[0026] S4, knowledge blind spot distribution and learning mode evolution analysis are performed on the dynamic three-level teaching interaction graph, the analysis result is input into the trained intelligent agent, the intelligent agent selects teaching resources corresponding to the knowledge point requiring teaching recommendation from the pre-constructed teaching resource recommendation library, and the teaching resources are pushed to the student user in a preset mode.

[0027] As a preferred embodiment of the second aspect, the specific process of step S3 is as follows:

[0028] S31, taking the student requiring teaching recommendation as a student node, obtaining student information from the online teaching programming system, taking the obtained student information as student node features, taking the knowledge points corresponding to the programming questions on the online teaching programming system as knowledge point nodes, taking the teachers as teacher nodes, taking the obtained human-computer interaction behaviors as edges between the student nodes and the knowledge point nodes, taking the obtained student-student interaction behaviors as edges between two student nodes, and taking the obtained teacher-student interaction behaviors as edges between the student nodes and the teacher nodes, to obtain an original three-level teaching interaction graph;

[0029] S32, inputting the original three-level teaching interaction graph into the trained graph neural network, updating the nodes and edges in the original three-level teaching interaction graph, and forming a dynamic three-level teaching interaction graph.

[0030] In a third aspect, the present application provides a computer electronic device, comprising a memory and a processor;

[0031] The memory is used for storing a computer program.

[0032] The processor is used for implementing the teaching recommendation method according to any one of the second aspect when executing the computer program.

[0033] Compared with the prior art, the present application has the following beneficial effects:

[0034] The present application provides an online teaching programming system and a teaching recommendation method based on large language model assistance, which effectively solves the problem of insufficient teaching resources for information courses in primary and secondary schools. With the support of the language model, students can complete programming learning tasks independently without relying on teachers, enhancing their programming ability and problem-solving ability. At the same time, the present application provides customized teaching support for students by building a comprehensive learning community, accurately pushing teaching resources and intervention strategies, solving the problem of insufficient personalized teaching recommendation in existing online teaching programming systems, and having the advantages of efficiency and individualization. BRIEF DESCRIPTION OF DRAWINGS

[0035] Figure 1 The present application provides an online teaching programming system and a teaching recommendation method based on large language model assistance, which effectively solves the problem of insufficient teaching resources for information courses in primary and secondary schools. With the support of the language model, students can complete programming learning tasks independently without relying on teachers, enhancing their programming ability and problem-solving ability. At the same time, the present application provides customized teaching support for students by building a comprehensive learning community, accurately pushing teaching resources and intervention strategies, solving the problem of insufficient personalized teaching recommendation in existing online teaching programming systems, and having the advantages of efficiency and individualization.

[0036] Figure 2 The step flow chart of the teaching recommendation method of the present application;

[0037] Figure 3 The structural schematic diagram of the dynamic three-level teaching interaction diagram of the present application;

[0038] Figure 4 The schematic diagram of the proportion of the data obtained by the present embodiment in each learning stage; DETAILED DESCRIPTION

[0039] In order to make the above objectives, characteristics and advantages of the present application more apparent, obvious and easy to understand, the specific embodiments of the present application will be described in detail below with reference to the drawings. In the following description, a large number of specific details are set forth in order to facilitate a full understanding of the present application. However, the present application can be implemented in many other ways different from those described herein, and those skilled in the art can make similar improvements without departing from the connotation of the present application, so the present application is not limited by the specific embodiments disclosed below. The technical features in each embodiment of the present application can be combined accordingly without conflict.

[0040] In the description of the present application, it should be understood that the terms "first", "second" are only used for distinguishing description purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined with "first", "second" can explicitly or implicitly include at least one of the features.

[0041] The technical problems to be solved by the present application are: 1) the existing programming education tools are mostly single-function modules, lacking integrated support throughout the learning process, and unable to effectively integrate teaching, evaluation and feedback functions; 2) there is insufficient personalized support for students during the learning process, and when there are differences in individual levels of students, the teaching content cannot be adjusted adaptively; 3) the existing systems usually only focus on data collection of programming behavior itself, ignoring multi-modal information in the teaching process, such as teacher-student interaction, environmental information, etc., and cannot form a comprehensive perception of the learning state of students. Based on this, the present application proposes an online teaching programming system based on large language model assistance and a teaching recommendation method.

[0042] As shown in Figure 1 In a preferred implementation of the present application, the above-mentioned online teaching programming system based on large language model assistance includes the following modules:

[0043] Base model acquisition module: used for acquiring a general large language model and fine-tuning it to become a specialized base model for the field of programming teaching;

[0044] The problem decomposition module is configured to obtain the first user input and convert the first user input into a designed first prompt, input a question to be decomposed and the first prompt into the base model, and output a problem decomposition result from the base model and display the problem decomposition result on a user interface of the online teaching programming system.

[0045] The abstract modeling module is configured to obtain the second user input and convert the second user input into a designed second prompt, input a question to be modeled and the second prompt into the base model, and output an abstract modeling result from the base model and display the abstract modeling result on the user interface.

[0046] The algorithm design module is configured to obtain the third user input and convert the third user input into a designed third prompt, input a question to be designed and the third prompt into the base model, and output an algorithm design result from the base model and display the algorithm design result on the user interface.

[0047] The code analysis and assistance module is configured to intelligently analyze a code completed by a user, obtain error information appearing on a console, input the code completed by the user and the error information into the base model, and output an analysis result and a correction suggestion from the base model and display the analysis result and the correction suggestion on the user interface.

[0048] The evaluation module is configured to intelligently evaluate a code submitted by a user, input the code submitted by the user, a standard answer of a programming question, and a scoring rule into the base model, and output an evaluation result of the code submitted by the user from the base model and display the evaluation result on the user interface.

[0049] It should be noted that, in the base model obtaining module of the present application, the specific process of fine-tuning the large language model is as follows: the LoRA model is used to fine-tune the large language model, and the pre-designed restriction words are used to constrain the output content and language style of the large language model, the pre-constructed information technology course knowledge base is obtained, the information technology course knowledge base and the large language model fine-tuned by the LoRA are used as inputs of the large model knowledge chain framework, a hallucination-reduced large language model is obtained, and the hallucination-reduced large language model is used as the base model.

[0050] It should be noted that, in the problem decomposition module of the present application, the first prompt contains an output format requirement and a word limit of the base model, and indicates that the base model needs to complete problem decomposition.

[0051] It should be noted that, in the abstract modeling module of the present application, the second prompt contains an output format requirement and a word limit of the base model, and indicates that the base model needs to provide an explanation of a modeling process.

[0052] It should be noted that in the algorithm design module of the present application, the third prompt word contains the output of the exclusive base model in the form of algorithm flowchart and flowchart format requirements, and points out that the exclusive base model needs to provide explanation of the algorithm design process.

[0053] It should be noted that in the code analysis and assistance module of the present application, it contains a code analysis module and an auxiliary debugging module.

[0054] The code analysis module is used to obtain the code completed by the user, input the code completed by the user into the exclusive base model, output the analysis results of the code from the whole to the line by the exclusive base model, and display the analysis results on the user interface.

[0055] The auxiliary debugging module is used to obtain the error information output by the console, input the error information into the exclusive base model, output the error explanation and correction suggestion of the error information by the exclusive base model, and display the error explanation and correction suggestion of the error information on the user interface.

[0056] It should be noted that in the evaluation module of the present application, the exclusive base model weights the execution time and code analysis result of the user submitted code to comprehensively evaluate the efficiency and quality of the user submitted code, and finally outputs the evaluation result of the user submitted code.

[0057] As shown in Figure 2 In a preferred implementation manner of the present application, based on the online teaching programming system described above, the present application further provides a teaching recommendation method, which includes steps S1-S4, and the implementation process thereof will be described below.

[0058] S1, obtain multi-modal data in the online classroom teaching process, including human-computer interaction data, student-student interaction data and teacher-student interaction data.

[0059] It should be noted that in step S1 of the present application, the human-computer interaction data generated by the student is obtained by the online teaching programming system, such as the code submitted by the student, the questions asked by the student to the exclusive base model, etc. The voice data and video data of the student during the class are collected by the intelligent classroom Internet of Things equipment. The two kinds of data can come from the interaction behavior between students, such as the communication between students, and can also come from the interaction behavior between students and teachers, such as the questions asked by students to teachers.

[0060] S2, data cleaning is performed on the multi-modal data, and the multi-modal data after data cleaning is labeled to label the human-computer interaction behavior, student-student interaction behavior and teacher-student interaction behavior, and obtain the processed multi-modal data.

[0061] It should be noted that in step S2 of the present application, the missing and incomplete data in the collected multi-modal data such as text, voice and video are cleaned, and then the cleaned multi-modal data are standardized and labeled to label the human-computer interaction behavior, the student-student interaction behavior and the teacher-student interaction behavior.

[0062] S3, constructing a dynamic three-level teaching interaction graph coupled with knowledge resources and interaction behaviors based on the processed multi-modal data;

[0063] It should be noted that the specific process of step S3 of the present application is as follows:

[0064] S31, taking a student to be taught as a student node, obtaining student information (such as student gender, grade, etc.) from the online teaching programming system, taking the obtained student information as student node features, taking knowledge points corresponding to programming topics on the online teaching programming system as knowledge point nodes, taking a teacher as a teacher node, taking the obtained human-computer interaction behavior as an edge between the student node and the knowledge point node, taking the obtained student-student interaction behavior as an edge between two student nodes, taking the obtained teacher-student interaction behavior as an edge between the student node and the teacher node, and obtaining an original three-level teaching interaction graph;

[0065] S32, inputting the original three-level teaching interaction graph into the trained graph neural network to update the nodes and edges in the original three-level teaching interaction graph, and forming a dynamic three-level teaching interaction graph.

[0066] S4, performing knowledge blind spot distribution and learning mode evolution analysis on the dynamic three-level teaching interaction graph, inputting the analysis result into the trained intelligent agent, and selecting teaching resources corresponding to the knowledge point to be taught from the pre-constructed teaching resource recommendation library and pushing the teaching resources to student users in a preset manner by the intelligent agent.

[0067] In order to better show the specific implementation and technical effects of the present application, the teaching recommendation method shown in steps S1-S4 in the above preferred implementation mode will be applied to a specific example.

[0068] Embodiment

[0069] The specific implementation process of the teaching recommendation method used in this embodiment is as described above, and will not be repeated here.

[0070] The online teaching programming system based on large language model assistance in this embodiment has been tried in many primary and secondary schools in Zhejiang Province, with nearly five thousand registered users, and a total of 70407 human-computer interaction data collected. At present, more than 200 schools from primary schools, junior high schools, high schools and universities have collected 12183 pieces of multi-modal data through questionnaires, interviews and other investigation methods, and the proportion of each stage is as follows: Figure 4as shown.

[0071] To show the technical effect of the present embodiment, 831 pieces of background data generated by 50 students in a class of a certain middle school in an information class were analyzed, and they were divided into high-efficiency group, medium-efficiency group and low-efficiency group according to the number of successful submission and the correct rate of the questions. After excluding some useless records (such as login), based on the obtained data, the experimental results obtained by using the method of the present application are shown in Table 1.

[0072] Table 1. Experimental results of learning efficiency improvement of a class using the method of the present application

[0073] Query large model Running error Running success Query large model proportion High efficiency group 76 97 107 37.25％ Medium efficiency group 42 92 47 30.21％ Low efficiency group 19 110 5 16.52％

[0074] As shown in Table 1, the high-efficiency group can use the online teaching programming system to assist programming more proficiently and effectively, and the low-efficiency group rarely asks questions to the large language model. It can be seen that the online teaching programming system developed by the present application can improve the learning efficiency of programming learners.

[0075] It can be understood that the teaching recommendation method described in S1-S4 can be realized by a computer program. Therefore, based on the same inventive concept, another preferred embodiment of the present application also provides a computer electronic device corresponding to the teaching recommendation method provided in the above embodiment, which includes a memory and a processor;

[0076] The memory is used to store a computer program.

[0077] The processor is used to implement the teaching recommendation method in the above embodiment when executing the computer program.

[0078] In addition, the logical instructions in the above memory can be realized in the form of a software functional unit and sold or used as an independent product, which can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application or the part of the technical solutions that essentially contribute to the prior art can be embodied in the form of a software product, which is stored in a storage medium and includes a plurality of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the method described in the various embodiments of the present application.

[0079] It can be understood that the processor described above can be a general processor, including a central processing unit (CPU), a network processor (NP), etc.; can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components.

[0080] The above-described embodiments are only a preferred solution of the present application, and are not intended to limit the present application. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the present application. Therefore, all technical solutions obtained by equivalent replacement or equivalent transformation shall fall within the protection scope of the present application.

Claims

1. An online teaching programming system based on a large language model, characterized by: include: Base model acquisition module: used to obtain a general large language model and fine-tune it to make it a dedicated base model for programming teaching; Problem decomposition module: used to obtain the first user input and convert it into a designed first prompt word, input the problem to be decomposed and the first prompt word into the dedicated base model, and the dedicated base model outputs the problem decomposition result and displays it on the user interface of the online teaching programming system; An abstract modeling module is configured to obtain a second user input and convert it into a designed second prompt word, input the topic to be abstractly modeled and the second prompt word into the dedicated base model, and output the abstract modeling result by the dedicated base model and display it on the user interface; Algorithm design module: used to obtain third user input and convert it into a designed third prompt word, input the problem to be algorithmically designed and the third prompt word into the dedicated base model, and have the dedicated base model output the algorithm design result and display it on the user interface; Code analysis and auxiliary module: used to intelligently analyze the code completed by the user, obtain the error information that appears in the console, and input the user's completed code and error information into the dedicated base model. The dedicated base model outputs the analysis results and correction suggestions and displays them on the user interface; Evaluation module: used to intelligently evaluate the code submitted by the user, input the user-submitted code together with the standard answers and scoring rules of the programming questions into the exclusive base model, and the exclusive base model outputs the evaluation results of the user-submitted code and displays it on the user interface.

2. The online teaching programming system according to claim 1, wherein: In the base model acquisition module, the specific process of fine-tuning the large language model is as follows: use the LoRA model to fine-tune the large language model, and use pre-designed restrictive words to constrain the output content and language style of the large language model, obtain a pre-built information technology course knowledge base, and use the information technology course knowledge base and the large language model fine-tuned by LoRA as inputs to the large model knowledge chain framework to obtain a large language model with reduced hallucinations and use it as an exclusive base model.

3. The online teaching programming system according to claim 1, wherein: In the problem decomposition module, the first prompt word includes the output format requirements and word limit of the exclusive base model, and indicates that the exclusive base model is required to complete the problem decomposition.

4. The online teaching programming system according to claim 1, wherein: In the abstract modeling module, the second prompt word includes that the output of the exclusive base model is a computational model and a word limit, and indicates that the exclusive base model is required to provide an explanation of the modeling process.

5. The online teaching programming system according to claim 1, wherein: In the algorithm design module, the third prompt word includes that the output of the exclusive base model is an algorithm flow chart and flow chart format requirements, and indicates that the exclusive base model is required to provide an explanation of the algorithm design process.

6. The online teaching programming system according to claim 1, wherein: The code analysis and auxiliary module includes a code analysis module and an auxiliary debugging module; The code analysis module is used to obtain the code completed by the user, input the code completed by the user into the exclusive base model, and the exclusive base model outputs the analysis results of the code from the whole to the line by line, and displays the analysis results on the user interface; The auxiliary debugging module is used to obtain the error information output by the console, input the error information into the exclusive base model, and the exclusive base model outputs the error explanation and correction suggestions of the error information, and displays the error explanation and correction suggestions of the error information on the user interface.

7. The online teaching programming system according to claim 1, wherein: In the evaluation module, the exclusive base model weights the execution time and code analysis results of the user-submitted code to comprehensively assess the efficiency and quality of the user-submitted code, and finally outputs the evaluation results of the user-submitted code.

8. A teaching recommendation method based on the online teaching programming system according to any one of claims 1 to 7, characterized in that: The following steps are involved: S1. Obtain multimodal data during online classroom teaching, including human-computer interaction data, student-student interaction data, and teacher-student interaction data; S2. Clean the multimodal data and annotate the cleaned multimodal data to identify human-computer interaction behaviors, student-student interaction behaviors, and teacher-student interaction behaviors, thereby obtaining processed multimodal data. S3, construct a dynamic three-level teaching interaction diagram that couples knowledge resources and interactive behaviors based on the processed multimodal data; S4. Conduct knowledge blind spot distribution and learning pattern evolution analysis on the dynamic three-level teaching interaction diagram, input the analysis results into the trained intelligent agent, and the intelligent agent selects teaching resources corresponding to the knowledge points that need to be recommended for teaching from the pre-built teaching resource recommendation library and pushes the teaching resources to student users in a preset manner.

9. The teaching recommendation method according to claim 8, wherein: The specific process of step S3 is as follows: S31. The student to be recommended for teaching is used as a student node. Student information is obtained from the online teaching programming system. The obtained student information is used as a student node feature. The knowledge point corresponding to the programming problem in the online teaching programming system is used as a knowledge point node. The teaching teacher is used as a teacher node. The obtained human-computer interaction behavior is used as an edge between the student node and the knowledge point node. The obtained student-student interaction behavior is used as an edge between two student nodes. The obtained teacher-student interaction behavior is used as an edge between the student node and the teacher node, thereby obtaining an original three-level teaching interaction graph. S32. Input the original three-level teaching interaction graph into the trained graph neural network, update the nodes and edges in the original three-level teaching interaction graph, and form a dynamic three-level teaching interaction graph.

10. A computer electronic device, characterized in that: including memory and processor; The memory is used to store computer programs; The processor is configured to implement the teaching recommendation method according to any one of claims 8 to 9 when executing the computer program.

Images