BOPPPS teaching design generation method based on teacher and student agent comment interaction

Abstract

The invention provides a BOPPPS teaching design generation method based on teacher and student agent comment interaction, and the method is characterized in that the real teaching interaction process between a teacher and students is simulated through the cooperation of multiple agents, and the automatic generation and personalized optimization of the teaching design content are realized. The method comprises the following steps: constructing a student agent corresponding to each student according to the knowledge level and interest preference of each student so as to obtain a student agent set; constructing a teacher intelligent agent according to a teaching outline corresponding to each class hour in the current teaching course and the learning condition data of the current teaching class; comment interaction between the teacher agent and the student agent is carried out, and a complete teaching design scheme of each class hour in the current teaching course is obtained according to a comment interaction result; therefore, compared with the prior art, the method has higher adaptability and acceptability, the real teaching scene is simulated by making full use of multi-agent cooperation, and therefore teaching design content better meeting the actual requirements of students can be generated.

Description

Technical Field

[0001] This invention relates to the fields of natural language processing and multi-agent collaboration, and in particular to a BOPPPS instructional design generation method based on teacher-student agent feedback interaction. Background Technology

[0002] The BOPPPS teaching model is a widely used structured teaching model in higher education, comprising six core components: Bridge-in, Objectives, Pre-assessment, Participatory Learning, Post-assessment, and Summary. This model is goal-oriented and student-centered, emphasizing full student participation and timely feedback. Currently, an increasing number of university teachers are integrating the BOPPPS teaching model into their teaching practice to improve overall teaching effectiveness.

[0003] In the early days, university teachers often manually conceived and wrote instructional designs based on the BOPPPS teaching model based on their teaching experience. This process was not only time-consuming and labor-intensive but also had a high professional threshold, often hindering the large-scale implementation of the BOPPPS teaching model in actual teaching. With the widespread adoption of the BOPPPS teaching model, the automated generation of instructional designs based on it has gradually become an important research hotspot in the field of educational technology. With the continuous evolution of machine learning and deep learning, especially the powerful capabilities of large language models in general language understanding and generation in recent years, solid technical support has been provided for the automated generation of complex instructional designs. Currently, some researchers and teaching platforms are actively exploring the application of large language models to the generation of BOPPPS instructional designs, attempting to automatically cover the six core stages from introduction to summary, striving to reduce the burden of lesson preparation for university teachers while improving the standardization and implementation effectiveness of BOPPPS instructional designs.

[0004] However, current mainstream methods still primarily rely on the interaction between university teachers and large language models, typically with teachers designing prompts based on their teaching experience to drive the generation process. While this teacher-experience-centric model has some guiding significance, it lacks student participation and feedback mechanisms, making it difficult to dynamically respond to individual differences among students in terms of knowledge level, interests, and preferences. If the instructional design lacks a precise grasp of students' real needs, it becomes difficult to rationally arrange the teaching pace and key content, thus affecting the relevance and effectiveness of the teaching. Summary of the Invention

[0005] In view of the shortcomings of the prior art described above, the purpose of this invention is to provide a BOPPPS instructional design generation method based on teacher-student intelligent agent feedback interaction, which is used to solve the problem that existing instructional designs lack a precise grasp of students' real needs, make it difficult to reasonably arrange the teaching pace and key content, thereby affecting the pertinence and effectiveness of teaching.

[0006] To achieve the above and other related objectives, the present invention provides the following technical solution:

[0007] A BOPPPS instructional design generation method based on teacher-student intelligent agent feedback interaction includes the following steps: obtaining each student's knowledge level and interest preferences, and constructing a student intelligent agent corresponding to each student based on their knowledge level and interest preferences, thereby obtaining a set of student intelligent agents; obtaining the teaching syllabus corresponding to each class hour in the current teaching course and the learning situation data of the current teaching class, and constructing a teacher intelligent agent based on the teaching syllabus corresponding to each class hour in the current teaching course and the learning situation data of the current teaching class; conducting feedback interaction between the teacher intelligent agent and the student intelligent agent, and obtaining a complete instructional design scheme for each class hour in the current teaching course based on the feedback interaction results.

[0008] In one embodiment of the present invention, the step of constructing a student intelligent agent corresponding to each student based on each student's knowledge level and interest preferences to obtain a set of student intelligent agents includes: forming a text description of each student's knowledge level and interest preferences as system prompt words for a large language model; constructing a student intelligent agent corresponding to each student based on the large language model and each student's knowledge level and interest preferences to obtain a set of student intelligent agents; wherein the student intelligent agents are able to perform differentiated understanding and feedback on the generated instructional design content based on the configured knowledge level and interest preferences.

[0009] In one embodiment of the present invention, each student's knowledge level includes the student's final exam score in the preceding courses related to the current teaching course and the student's post-test score in the previous class hour of the current teaching course, wherein the current teaching course is divided into H class hours; each student's interests and preferences are collected through a questionnaire.

[0010] In one embodiment of the present invention, the step of constructing a teacher agent based on the teaching syllabus corresponding to each class hour in the current teaching course and the learning data of the current teaching class includes: forming a text description of the teaching syllabus corresponding to each class hour in the current teaching course and the learning data of the current teaching class as system prompt words of a large language model; constructing a teacher agent through the large language model and based on the teaching syllabus and learning data; wherein the teacher agent is capable of generating instructional design content for each class hour in the current teaching course based on the configured teaching syllabus and learning data.

[0011] In one embodiment of the present invention, the teaching syllabus includes the teaching scope, core knowledge points, and ability goals that students need to achieve in each class hour of the current teaching course; the learning data of the current teaching class includes the set of interests and preferences of all students, the average final exam scores of all students in the preceding courses most relevant to the current teaching course, and the average post-test scores of all students in the previous class hour of the current teaching course.

[0012] In one embodiment of the present invention, the step of conducting the review and feedback interaction between the teacher agent and the student agent, and obtaining the complete teaching design scheme for each class hour in the current teaching course based on the review and feedback interaction results, includes: obtaining the teaching design content of the previous BOPPPS stage; the teacher agent generating the teaching design content of the current BOPPPS stage based on the teaching design content of the previous BOPPPS stage; obtaining feedback information from all student agents on the teaching design content of the current BOPPPS stage; confirming the magnitude relationship between the feedback information of all student agents and a specified threshold; and obtaining the complete teaching design scheme for each class hour in the current teaching course based on the confirmation result.

[0013] In one embodiment of the present invention, confirming the relationship between the feedback information of all student agents and a specified threshold, and obtaining a complete instructional design scheme for each class hour in the current teaching course based on the confirmation result, includes: determining whether the feedback information of all student agents is lower than the specified threshold; if so, the teacher agent makes targeted modifications to the instructional design content of the current BOPPPS stage based on the common problems extracted from the feedback information, and re-accepts feedback from student agents until the acceptance rate is not lower than the specified threshold, thereby proceeding to the generation of instructional design content for the next BOPPPS stage; if not, the teacher agent generates instructional design content for the next BOPPPS stage based on the instructional design content of the current BOPPPS stage, until a complete instructional design scheme for each class hour in the current teaching course is obtained.

[0014] In one embodiment of the present invention, if the acceptance ratio remains below a specified threshold and the maximum number of modification rounds is reached, a step rollback mechanism is triggered. The teacher agent then rolls back to the previous teaching step and, based on the common problems extracted from the feedback information of the previous teaching step, makes targeted modifications to the instructional design content of the BOPPPS step in the previous teaching step. The process then proceeds sequentially to generate the instructional design content of the current BOPPPS step.

[0015] As described above, the BOPPPS instructional design generation method based on teacher-student intelligent agent interaction of the present invention has the following beneficial effects: The present invention constructs student intelligent agents based on each student's knowledge level and interest preferences, and constructs teacher intelligent agents based on the teaching syllabus and student learning data. The teacher intelligent agent can generate the instructional design content of the current BOPPPS stage based on the instructional design content of the previous BOPPPS stage, and then obtain feedback information from all student intelligent agents on the instructional design content of the current BOPPPS stage, and judge the relationship between the feedback information and a specified threshold, and then perform a cycle of correction and feedback until a complete instructional design scheme for each class hour in the current teaching course is obtained. Compared with the prior art, the present invention has higher adaptability and acceptability. The present invention makes full use of multi-agent collaboration to simulate real teaching scenarios, thereby generating instructional design content that is more in line with the actual needs of students. Attached Figure Description

[0016] Figure 1 The diagram shown is a general flowchart of the BOPPPS instructional design generation method based on teacher-student intelligent agent feedback interaction disclosed in this embodiment of the invention.

[0017] Figure 2 The diagram shown is a detailed flowchart of the BOPPPS instructional design generation method based on teacher-student intelligent agent feedback interaction disclosed in this embodiment of the invention. Detailed Implementation

[0018] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. It should be noted that, unless otherwise specified, the following embodiments and features described herein can be combined with each other.

[0019] The purpose of this invention is to simulate real teaching scenarios using multi-agent collaboration, and to provide a BOPPPS instructional design generation method based on teacher-student agent feedback interaction. Through teacher agent explanation and student agent feedback, instructional designs tailored to students' actual needs are generated. The process is as follows: Figure 1 As shown, the details are as follows:

[0020] Step 101: Obtain the knowledge level and interest preferences of each student, and construct a student agent corresponding to each student based on the knowledge level and interest preferences, thereby obtaining a set of student agents.

[0021] Specifically, each student's knowledge level and interests are described in text and used as system prompts for the large language model. The large language model is then used to construct student agents corresponding to each student based on their knowledge level and interests, thus obtaining a set of student agents.

[0022] More specifically, let the current teaching course be C, which is divided into H class hours. The current task is to generate the instructional design for the h-th class hour (1≤h≤H).

[0023] Let the set of students in the teaching class be S = {s1, s2, ... s}. M}, where M is the total number of students in the class, and for each student s m (1≤m≤M), obtain student s m The knowledge level and interests of students are used to form textual descriptions, which serve as system prompts for the large language model. These prompts are then processed through the large language model and based on student data. m Based on students' knowledge level and interests, a corresponding student intelligent agent A(s) is constructed. m ); students m The student's knowledge level is composed of their final exam scores in prerequisite courses related to the current course C, and their post-test score in the (h-1)th class hour (the post-test score can be omitted when h=1); it should be noted that N prerequisite courses most relevant to the current course are selected, and student s is recorded. m Final exam grades in these prerequisite courses in, (1≤m≤M, 1≤n≤N) represents the student s m Final exam score in the nth prerequisite course; student s m The post-test score from the previous lesson is recorded as follows: (1≤h≤H);

[0024] Student S m Interests and preferences were collected through questionnaires;

[0025] The final set of student agents is A = {A(s1), A(s2), ..., A(s...}. m Each student intelligent agent A(s) m (1≤m≤M) can provide differentiated understanding and feedback on the generated instructional design content based on the configured knowledge level and interest preferences.

[0026] Step 102: Obtain the teaching syllabus for each class hour in the current teaching course and the learning data of the current teaching class, and construct a teacher agent based on the teaching syllabus for each class hour in the current teaching course and the learning data of the current teaching class.

[0027] Specifically, the teaching syllabus corresponding to each class hour in the current teaching course and the learning data of the current class are formed into textual descriptions, which serve as system prompts for the large language model. The teacher's intelligent agent is then constructed through the large language model based on the teaching syllabus and learning data.

[0028] More specifically, the teaching syllabus fragment corresponding to the current class hour and the learning situation data of the current class are obtained and formed into a text description, which serves as the system prompt words of the large language model. The teacher intelligent agent A(t) is constructed through the large language model based on the teaching syllabus fragment and the learning situation data, so that the teacher intelligent agent A(t) can generate the teaching design content for each class hour based on the configured teaching syllabus and learning situation data.

[0029] The syllabus fragment provides the teaching scope, core knowledge points, and ability goals that students are expected to achieve in the current class period. The learning data for the current class consists of the set of interests and preferences of all students, the average final exam scores of all students in the N prerequisite courses most relevant to the current course C, and the average post-test score of all students in the (h-1)th class period (the average post-test score can be omitted when h=1). It should be noted that the set of interests and preferences of all students is a text set formed after deduplication, merging, and frequency statistics of the interest and preference data of all students in the class. Where Q represents the total number of interests in the interest preference set, and I q (1≤q≤Q) represents the q-th interest preference. (1≤q≤Q) represents the frequency of the q-th interest preference; the average final exam score of all students in the N prerequisite courses most relevant to the current course C. in, (1≤n≤N) represents the average final exam score of all students in the teaching class for the nth prior course; the average post-test score of all students in the previous class period (i.e., the h-1th class period) is denoted as (1≤h≤H).

[0030] Step 103: Conduct feedback and discussion interactions between the teacher's and student's agents, and obtain a complete instructional design plan for each class hour in the current teaching course based on the feedback and discussion interaction results.

[0031] Specifically, firstly, the instructional design content of the previous BOPPPS stage is obtained, and the teacher agent A(t) generates the instructional design content of the current BOPPPS stage based on the instructional design content of the previous BOPPPS stage; then, the feedback information of all student agents on the instructional design content of the current BOPPPS stage is obtained; then, the relationship between the feedback information of all student agents and the specified threshold is confirmed, and the complete instructional design scheme for each class hour in the current teaching course is obtained based on the confirmation result.

[0032] More specifically, let the instructional design content corresponding to the six stages of BOPPPS be as follows: in, (1≤h≤H,1≤k≤6) represents the instructional design content generated by the teacher agent and student agent during the interactive review process for the k (1≤k≤6) BOPPPS stage in the h-th class period;

[0033] For any k-th (1≤k≤6) BOPPPS step, the following process will be followed for presentation and discussion:

[0034] First, the teacher agent A(t) designs the instructional content for the (k-1)th BOPPPS segment. (When k=1, prerequisite dependencies are omitted), generate the instructional design content for the k-th BOPPPS stage.

[0035] Subsequently, all student agents evaluated the instructional design content. Provide feedback and form a feedback set. in, Represents the student intelligent agent A(s) m ) content of instructional design Feedback information, including feedback results (i.e., acceptance or rejection of the instructional design content) and corresponding feedback opinions;

[0036] If the student's intelligent agent A(s) m ) to receive instructional design content If the acceptance rate is not lower than the specified threshold α, the process proceeds to the (k+1)th BOPPPS stage for instructional design generation; if the acceptance rate is lower than the specified threshold α, the teacher agent A(t) will, based on the feedback set... The common problems extracted from the teaching design content Targeted corrections are made, and feedback from the student agent is accepted again. This correction-feedback loop continues until the acceptance rate is not lower than a specified threshold α or the maximum number of modification rounds β is reached. If the acceptance rate is still lower than the specified threshold α after reaching the maximum number of modification rounds β, a step rollback mechanism is triggered. The teacher agent rolls back to the (k-1)th teaching step and re-adjusts based on the feedback set of the (k-1)th teaching step. The common problems extracted from the teaching design content Make targeted revisions, and then proceed step by step to the instructional design content according to the established process. The generation of;

[0037] The above process continues until the instructional design for all six stages is completed, ultimately outputting the complete instructional design plan for the h-th class hour.

[0038] For the overall process in practical applications, please refer to [link / reference]. Figure 2This example is from a university course called "Algorithms and Data Structures," which is divided into 64 class hours. The current task is to generate a teaching design for the 4th class hour. The specific operation is as follows:

[0039] Step 1: Constructing a student's intelligent agent based on their knowledge level and interests:

[0040] The set of students in the teaching class is S = {s1, s2, ... s} 30 The class has a total of 30 students. For each student... m (1≤m≤30), obtain student s m The knowledge level and interests of students are used to form textual descriptions, which serve as system prompts for the large language model. These prompts are then processed through the large language model and based on student data. m Based on students' knowledge level and interests, a corresponding student intelligent agent A(s) is constructed. m );

[0041] Student S m The student's knowledge level is comprised of their final exam scores in prerequisite courses related to the current course C, and their post-test score in the third class period. It should be noted that the three prerequisite courses most relevant to the current course were selected: *Introduction to Computers*, *Computer Organization*, and *Object-Oriented Programming*. Student s... m Final exam grades in these prerequisite courses in, (1≤m≤30, 1≤n≤3) represents the number of students s. m Final exam score in the nth prerequisite course; student s m The post-test score in the third class period is recorded as (1≤h≤64), for example, student s1's G(s1)={80,90,85}, student s1's It is 80;

[0042] Student S m Interest preferences were collected through questionnaires. For example, student s1's interest preferences are {movies, music}.

[0043] The final set of student agents is A = {A(s1), A(s2), ..., A(s...}. m Each student intelligent agent A(s) m (1≤m≤M) can provide differentiated understanding and feedback on the generated instructional design content based on the configured knowledge level and interest preferences;

[0044] Step 2: Constructing a teacher agent based on the teaching syllabus and student learning data:

[0045] Obtain the teaching syllabus fragment corresponding to the current class hour and the learning situation data of the current class, and form a text description as the system prompt words of the large language model. Construct a teacher agent A(t) based on the teaching syllabus fragment and learning situation data through the large language model, so that the teacher agent A(t) can generate teaching design content based on the configured teaching syllabus and learning situation data.

[0046] The syllabus fragments provide the teaching scope, core knowledge points, and ability goals that students need to achieve in the current class period. For example, the teaching scope of the 4th class period is the basic operations of linear lists (search, insert, delete).

[0047] The current class's learning data consists of the set of all students' interests and preferences, the average final exam scores of all students in the three prerequisite courses most relevant to the current course, and the average post-test scores of all students in the third class period. It should be noted that the set of all students' interests and preferences is a text set I = {{movies, 60%}, {music, 20%}, {swimming, 10%}} formed after deduplication, merging, and frequency statistics of the interest and preference data of all students in the class. The average final exam scores of all students in the three prerequisite courses most relevant to the current course... Average post-test score of all students in the third class period

[0048] Step 3: Conduct feedback and discussion interactions between teacher and student agents:

[0049] Let the instructional design content corresponding to the six stages of BOPPPS be as follows: in, This indicates that the teacher agent A(t) represents the instructional design content generated in the k-th (1≤k≤6) BOPPPS stage during the 4th class period;

[0050] Taking the second BOPPPS stage as an example:

[0051] Teacher agent A(t) based on the instructional design content of the previous stage. Generate instructional design content for the current stage.

[0052] All student agents A(s) m The instructional design content for the current stage Provide feedback and form a feedback set. in, Represents the student intelligent agent A(s) mFeedback information on the instructional design content of the second segment of the fourth lesson, including the feedback result (i.e., acceptance or rejection of the instructional design content) and corresponding feedback opinions, for example, student s1's feedback on the instructional design content. The feedback message was "Rejected; the terminology is unclear and the meaning of abstract data types cannot be understood."

[0053] If the student's intelligent agent receives the instructional design content If the acceptance rate is not lower than the specified threshold of 0.6, the process proceeds to the fifth BOPPPS stage for instructional design generation; if the acceptance rate is lower than the specified threshold of 0.6, the teacher agent A(t) will, based on the feedback set... The common problems extracted (e.g., unclear terminology) affect the content of instructional design. Make targeted corrections and accept feedback from student agents again. This correction-feedback loop continues until the acceptance rate is not lower than the specified threshold of 0.6 or the number of modification rounds reaches the maximum of 3.

[0054] If, after reaching the maximum number of modification rounds (3), the acceptance rate remains below the specified threshold of 0.6, a rollback mechanism is triggered. The teacher agent reverts to the third teaching round and re-adjusts the feedback set from that round. The common problems extracted from the teaching design content Make targeted revisions, and then proceed step by step to the instructional design content according to the established process. The generation of;

[0055] The above process continues until the instructional design for all six stages is completed, ultimately outputting the complete instructional design plan for the fourth class period. For example, "Remember the scene in *Avengers: Infinity War* where Thanos snapped his fingers and half of all life in the universe disappeared? Imagine if we used a sequential list to record all the superheroes, and now we need to delete half of them. How would we quickly accomplish this? Besides deletion, we can also freely add heroes or find the position of a specific character—this is actually the core operation of a linear list. In this lesson, we'll start with this kind of scenario from the movie to learn about insertion, deletion, searching, and traversal of linear lists."

[0056] In summary, this invention constructs a student agent based on each student's knowledge level and interests, and a teacher agent based on the teaching syllabus and student learning data. The teacher agent can generate the teaching design content for the current BOPPPS session based on the teaching design content of the previous BOPPPS session. It then obtains feedback information from all student agents regarding the teaching design content of the current BOPPPS session, judges the relationship between the feedback information and a specified threshold, and performs a cycle of correction and feedback until a complete teaching design scheme for each class hour in the current course is obtained. Compared with existing technologies, this invention has higher adaptability and acceptability. It fully utilizes multi-agent collaboration to simulate real teaching scenarios, thereby generating teaching design content that better meets students' actual needs. Therefore, this invention can utilize multi-agent collaboration to simulate the real teaching interaction process between teachers and students, realizing the automated generation and personalized optimization of teaching design content.

[0057] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the invention. All equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in this invention should still be covered by the claims of this invention.

Claims

1. A BOPPPS instructional design generation method based on teacher-student intelligent agent feedback interaction, characterized in that, The method includes the following steps: The knowledge level and interest preferences of each student are obtained, and a student agent corresponding to each student is constructed based on the knowledge level and interest preferences, thereby obtaining a set of student agents; Obtain the teaching syllabus for each class hour in the current teaching course and the learning situation data of the current teaching class, and construct a teacher agent based on the teaching syllabus for each class hour in the current teaching course and the learning situation data of the current teaching class; Conduct the feedback interaction between the teacher agent and the student agent, and obtain the complete teaching design scheme for each class hour in the current teaching course based on the feedback interaction results.

2. The BOPPPS instructional design generation method based on teacher-student intelligent agent feedback interaction as described in claim 1, characterized in that: The process involves constructing a student agent corresponding to each student based on their knowledge level and interests, thereby obtaining a set of student agents, including: Each student's knowledge level and interests are described in text and used as system prompts for the large language model. Based on each student's knowledge level and interests, a student agent corresponding to each student is constructed using the large language model, thus obtaining a set of student agents. These student agents are able to provide differentiated understanding and feedback on the generated instructional design content based on their configured knowledge level and interests.

3. The BOPPPS instructional design generation method based on teacher-student intelligent agent feedback interaction as described in claim 2, characterized in that: Each student's knowledge level includes their final exam scores in the prerequisite courses related to the current course and their post-test scores in the previous class hour of the current course. The current course is divided into H class hours. Each student's interests and preferences are collected through a questionnaire.

4. The BOPPPS instructional design generation method based on teacher-student intelligent agent feedback interaction as described in claim 2, characterized in that: The construction of the teacher agent based on the teaching syllabus corresponding to each class hour in the current teaching course and the learning data of the current class includes: The teaching syllabus and student learning data for each class hour in the current teaching course are formed into a text description and used as system prompts for the large language model. A teacher agent is constructed based on the teaching syllabus and student learning data through the large language model. The teacher agent can generate the teaching design content for each class hour in the current teaching course based on the configured teaching syllabus and student learning data.

5. The BOPPPS instructional design generation method based on teacher-student intelligent interaction and review as described in claim 4, characterized in that: The syllabus includes the scope of instruction, core knowledge points, and ability goals that students need to achieve in each class hour of the current course; the learning data of the current class includes the set of interests and preferences of all students, the average final exam scores of all students in the prerequisite courses most relevant to the current course, and the average post-test scores of all students in the previous class hour of the current course.

6. The BOPPPS instructional design generation method based on teacher-student intelligent agent feedback interaction as described in claim 1, characterized in that: The process involves conducting feedback interactions between the teacher and student agents, and based on the feedback interaction results, obtaining a complete instructional design scheme for each class hour in the current teaching course, including: The teacher agent obtains the instructional design content of the previous BOPPPS session and generates the instructional design content of the current BOPPPS session based on the instructional design content of the previous BOPPPS session. Obtain feedback from all student agents regarding the instructional design content of the current BOPPPS segment; Confirm the relationship between the feedback information of all student agents and the specified threshold, and obtain a complete instructional design scheme for each class hour in the current teaching course based on the confirmation results.

7. The BOPPPS instructional design generation method based on teacher-student intelligent agent feedback interaction as described in claim 6, characterized in that: The process involves confirming the relationship between the feedback information of all student agents and a specified threshold, and based on the confirmation results, obtaining a complete instructional design scheme for each class hour in the current teaching course, including: Determine whether the feedback information of all student agents is below a specified threshold; If so, the teacher agent will make targeted modifications to the instructional design content of the current BOPPPS stage based on the common problems extracted from the feedback information, and will accept feedback from the student agent again until the acceptance rate is not lower than the specified threshold, thereby entering the generation of instructional design content for the next BOPPPS stage. If not, the teacher agent generates the instructional design content for the next BOPPPS segment based on the instructional design content of the current BOPPPS segment, until a complete instructional design scheme for each class hour in the current teaching course is obtained.

8. The BOPPPS instructional design generation method based on teacher-student intelligent agent feedback interaction as described in claim 7, characterized in that: If the acceptance rate remains below the specified threshold and the maximum number of modification rounds is reached, a rollback mechanism is triggered. The teacher agent then rolls back to the previous teaching stage, re-adjusts the instructional design content of the BOPPPS stage based on the common problems extracted from the feedback information of the previous teaching stage, and then proceeds sequentially to generate the instructional design content of the current BOPPPS stage.

Images