Teaching design generation, optimization and reflection method and system based on AI chain driving

Abstract

The invention belongs to the technical field of education, and discloses a teaching design generation, optimization and reflection method and system based on AI chain driving, and the method comprises the steps: screening an applicable module from a reasoning module and a teaching design module according to a target teaching scene, converting the general description of the applicable module into task-oriented fine description, and carrying out the task-oriented fine description; after an instruction template is generated, specific teaching content is filled; an optimization module is selected to carry out secondary conversion, a suggestion template is generated, optimization suggestions are filled, and targeted improvement is carried out on instruction content; and the system verifies the optimization effect through a dual evaluation mechanism: when the content variation exceeds a preset threshold value, performing joint verification on the optimization result by adopting global scoring and local scoring, and when the scores of the global scoring and the local scoring both reach the standard, deploying the optimized instruction template to a teaching agent. According to the method, the AI automation advantage and the reflection mechanism are fused, the whole-process intelligence from template configuration, content filling to quality evaluation is realized, and the teaching design efficiency and adaptability are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the field of educational technology, and in particular to a method and system for utilizing artificial intelligence technology to assist in the generation, optimization and reflection of teaching designs. Background Art

[0002] At present, when teachers formulate teaching designs, they mainly rely on their personal teaching experience. This experience is reflected in whether the teaching design has constructed a clear knowledge framework, whether it fully covers the teaching elements, and whether each element is designed in depth and detail. For experienced teachers, when they construct teaching designs, they can pay attention to the knowledge framework of the teaching design, fully cover the teaching elements, and design each element in depth and detail. Their teaching designs usually maintain a high degree of logical coherence, the knowledge points are organically connected, and the teaching content is arranged in order from simple to complex, from basic to application. Each teaching element is fully developed. However, for new teachers, designing a high-quality teaching plan is a very challenging task. With the continuous development of artificial intelligence, it is possible to use these advanced technologies to assist and construct teaching designs. Although it can make up for the problem of ignoring the construction of the knowledge framework and the lack of teaching elements and the depth of the elements in the teaching design due to the lack of teacher experience. However, when the large language model (LLM) is directly used to construct the teaching design, its performance is often unsatisfactory.

[0003] This is mainly because LLM is a general model, and it is still challenging to use the general knowledge of LLM to complete too many tasks at one time. Therefore, the current teaching design construction methods have the following problems: (1) it is difficult to effectively organize teaching content and ignore the construction and connection of knowledge framework; (2) it fails to conduct in-depth and detailed design of teaching elements; (3) teaching elements are not fully covered in teaching design. Summary of the invention

[0004] The purpose of the embodiments of the present application is to provide a method and system for generating, optimizing and reflecting teaching designs based on an AI chain drive, aiming to solve the problem in the current teaching design construction method that it is difficult to conduct a systematic evaluation from three aspects: the accuracy of knowledge transfer, the coverage of teaching elements and the overall quality of the teaching design.

[0005] In order to solve the above technical problems, this application is implemented as follows: In a first aspect, an embodiment of the present application provides a method for generating, optimizing and reflecting teaching designs based on AI chain drive, the method comprising: A module suitable for a target teaching scenario task is selected from predefined reasoning modules and teaching design modules to obtain a selected first module; the selected first module is converted from a general description into a detailed description for the target teaching scenario task to obtain an adapted first module; the detailed description of the adapted first module is converted into an instruction template; based on the teaching design module and the target teaching scenario task, the instruction template is filled with specific instruction content using LLM to obtain a filled instruction template; Selecting a module suitable for optimizing the filled instruction template from the reasoning module to obtain a selected second module; converting the selected second module from a general description into a detailed description for a target teaching scenario task to obtain an adapted second module; converting the detailed description of the adapted second module into a suggestion template; using LLM to fill the suggestion template with specific suggestion content to obtain a filled suggestion template; optimizing each suggestion in the filled suggestion template to obtain an optimized instruction template; Determine whether the content quantity of the filled instruction template and the optimized instruction template is greater than the preset change requirement; if it is greater than the preset change requirement, use the global scoring mechanism and the local scoring mechanism to obtain the global score and the local part respectively; configure the optimized instruction template corresponding to the time when the global score and the local part are both greater than the corresponding threshold into the teaching intelligent body to obtain the final teaching design.

[0006] As an optional implementation of the first aspect of the present application, in the step of converting the detailed description of the adapted first module into an instruction template: the instruction template includes role settings, target audiences, and multiple teaching behaviors, the teaching behaviors represent instruction tags of the teaching process, and each instruction tag represents an independent teaching link; each instruction tag includes specific commands and rules, the commands are used to define specific execution behaviors in the teaching process, and are arranged in a logical order; the rules ensure that the implementation of the instructions complies with the predetermined logical order and teaching standards.

[0007] As an optional implementation manner of the first aspect of the present application, in the steps of obtaining the global score and the local score respectively using a global scoring mechanism and a local scoring mechanism: the local scoring mechanism is used to evaluate the improvement of the sub-parts of the instruction template, and the scoring is performed according to the bonus items and the deduction items, the sub-parts include role settings, target audiences, and multiple teaching behaviors, the bonus items include innovation and significant improvement, and the deduction items include redundancy and reduced relevance; the global scoring mechanism is used to evaluate the integrity, relevance, accuracy and creativity of the overall structure and content of the instruction template, and to give a comprehensive score of 0 to 100.

[0008] As an optional implementation of the first aspect of the present application, in the step of converting the adapted detailed description of the second module into a suggestion template: the suggestion template is presented in the form of key-value pairs for LLM to understand and execute gradual filling of key values.

[0009] As an optional implementation of the first aspect of the present application, the step of optimizing each suggestion in the filled suggestion template to obtain an optimized instruction template includes: optimizing each suggestion in the filled suggestion template through a guidance model, wherein the guidance model includes GPT-4o; wherein the optimization process includes modifying ambiguous instructions, eliminating redundant instructions, and adding new instructions.

[0010] As an optional implementation of the first aspect of the present application, it is determined whether the content quantity of the filled instruction template and the optimized instruction template is greater than the preset change requirement; if it is greater than the preset change requirement, the steps of using the global scoring mechanism and the local scoring mechanism to obtain the global score and the local score respectively also include: if it is less than the preset change requirement, reselecting the reasoning module and the teaching design module for iteration.

[0011] As an optional implementation of the first aspect of the present application, the optimized instruction template corresponding to the situation where the global score and the local part are both greater than the corresponding threshold value is configured into the teaching intelligent body, and before the step of obtaining the final teaching design, it also includes: classifying according to the size relationship between the global score and the local part and the corresponding threshold value; if the global score and the local part are both less than the corresponding threshold value, reselecting the reasoning module and the teaching design module; if the local part is greater than the corresponding threshold value and the global score is less than the corresponding threshold value, reselecting the reasoning module and the teaching design module; if the global score is greater than the corresponding threshold value and the local part is less than the corresponding threshold value, re-optimizing and generating the instruction template.

[0012] In a second aspect, the embodiment of the present application provides a teaching design generation, optimization and reflection system based on AI chain drive, the system includes a teaching design generation module, a teaching design optimization module and a teaching design reflection module, wherein: The teaching design generation module includes: A first module selection unit, used to select a module suitable for a target teaching scenario task from predefined reasoning modules and teaching design modules to obtain a selected first module; A first module adaptation unit, used to convert the selected first module from a general description into a detailed description for a target teaching scenario task, so as to obtain an adapted first module; a first module structuring unit, configured to convert the adapted first module's detailed description into an instruction template; A first content filling unit is used to fill the instruction template with specific instruction content using LLM based on the instruction design module and the target teaching scenario task to obtain a filled instruction template; The teaching design optimization module includes: a second module selection unit, configured to select a module suitable for optimizing the filled instruction template from the reasoning module to obtain a selected second module; A second module adaptation unit, used to convert the selected second module from a general description into a detailed description for a target teaching scenario task, so as to obtain an adapted second module; a second module structuring unit, for converting the adapted second module's detailed description into a suggestion template; A second content filling unit is used to fill the suggestion template with specific suggestion content using the LLM to obtain a filled suggestion template; An optimization filling unit, used for optimizing each suggestion in the filled suggestion template to obtain an optimized instruction template; The teaching design reflection module includes: A judgment classification unit is used to judge whether the content quantity of the filled instruction template and the optimized instruction template is greater than the preset change requirement; if it is greater than the preset change requirement, a global scoring mechanism and a local scoring mechanism are used to obtain a global score and a local score respectively; The configuration unit is used to configure the optimized instruction template corresponding to when the global score and the local score are both greater than the corresponding threshold into the teaching agent to obtain the final teaching design.

[0013] In a third aspect, an embodiment of the present application provides an electronic device, comprising a processor, a memory, and a program or instruction stored in the memory and executable on the processor, wherein the program or instruction, when executed by the processor, implements the steps of the method described in the first aspect.

[0014] In a fourth aspect, an embodiment of the present application provides a readable storage medium, on which a program or instruction is stored, and when the program or instruction is executed by a processor, the steps of the method described in the first aspect are implemented.

[0015] Compared with the prior art, the present invention proposes a teaching design generation, optimization and reflection method based on AI chain drive, which has the following advantages: 1. The teaching design generation method provided by the present invention guides LLM to organize these labels and infer the internal logical relationship between knowledge points through the steps of module selection, module adaptation, module structuring and content filling, thereby effectively solving the problem of unclear knowledge framework in teaching design.

[0016] 2. The teaching design optimization method provided by the present invention introduces a suggestion template, modifies the one-sided or inaccurate instructions in the original instruction template, removes redundant instructions, and adds missing instructions, thereby solving the problem of shallow depth of teaching elements.

[0017] 3. The teaching design reflection method provided by the present invention ensures that the teaching design can maintain the logical coherence of the knowledge framework and comprehensively cover the teaching elements in an overall manner through the adjustment of tactics and strategies, and can also have sufficient depth in details, effectively ensuring the comprehensiveness of the teaching design, thereby improving the quality of the teaching design. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 is a first flow chart of a method for generating, optimizing and reflecting teaching designs based on AI chain drive provided by the first embodiment of the present invention; Figure 2 is a second flow chart of an AI chain-driven instructional design generation, optimization and reflection method provided by the first embodiment of the present invention; Figure 3 It is a structural diagram of an AI chain-driven teaching design generation, optimization and reflection system provided by the second embodiment of the present invention. DETAILED DESCRIPTION

[0019] 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 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.

[0020] The terms "first", "second", etc. in the specification and claims of the present application are used to distinguish similar objects, and are not used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable under appropriate circumstances, so that the embodiments of the present application can be implemented in an order other than those illustrated or described here. In addition, the "and / or" in the specification and claims represents at least one of the connected objects, and the character " / " generally represents that the objects associated with each other are in an "or" relationship.

[0021] In order to illustrate the technical solution described in this application, a specific embodiment is provided below for illustration.

[0022] Example 1 See also Figure 1 and Figure 2 , Figure 1 This is a flowchart of a teaching design generation, optimization and reflection method based on AI chain drive proposed in the first embodiment of this application. Figure 2 A second flow chart of a teaching design generation, optimization and reflection method based on AI chain drive is provided for the first embodiment of the present invention, and the steps of the proposed method are as follows.

[0023] Step S1: Instructional design generation; Step S11: selecting a module suitable for the target teaching scenario task from the predefined reasoning modules and teaching design modules to obtain a selected first module; For example, guide LLM to screen a set of optional reasoning modules and teaching design modules, and select the modules that are most helpful in establishing a clear logical relationship between knowledge points. Given a set of reasoning modules (such as "critical thinking" and "breaking down problems into sub-problems") and a set of teaching design modules (such as "promoting reflective thinking" and "clarifying learning goals"), according to the specific teaching scenario, screen out the first module after selection that is most helpful for task solving. For example, there are 39 reasoning modules and 28 teaching design modules, all of which are basic functional descriptions.

[0024] Step S12: converting the selected first module from a general description into a detailed description for the target teaching scenario task to obtain an adapted first module; It should be noted that after completing the module selection, each selected reasoning module and teaching design module (i.e., the first module after selection) needs to be further adapted to better meet the teaching scenario, and the general description of the first module after selection is converted into specific instructions for the task, that is, the adapted first module is obtained.

[0025] Step S13: converting the adapted detailed description of the first module into an instruction template; In some embodiments, the instruction template includes role settings, target audiences, and multiple teaching behaviors. The teaching behaviors represent instruction tags of the teaching process, and each instruction tag represents an independent teaching link; each instruction tag includes specific commands and rules. The commands are used to define specific execution behaviors in the teaching process and are arranged in a logical order; the rules ensure that the implementation of the instructions complies with the predetermined logical order and teaching standards.

[0026] Specifically, after completing the module adaptation, these adapted first module descriptions are converted into an instruction template. The instruction template includes role settings, target audiences, and multiple instruction tags for the teaching process. Each instruction tag contains specific commands and rules. Commands define the specific execution behaviors in the teaching process and are arranged in a logical order. Rules are used to regulate and constrain the rules of the execution process to ensure that the implementation of instructions complies with the predetermined logical order and teaching standards. In addition, the teaching process is decomposed into multiple instruction tags with a clear order. Each instruction tag represents an independent teaching link, and each tag has a plan for gradually implementing the teaching link. The intelligent agent organizes the content in an orderly manner according to each teaching link to ensure that the knowledge points in the teaching design are naturally connected and logically smooth.

[0027] Step S14: Based on the teaching design module and the target teaching scenario task, use LLM to fill the instruction template with specific instruction content to obtain the filled instruction template; After the module is structured, the specific content of the instruction template is gradually filled in through LLM to obtain a filled instruction template. The content of the instruction template is filled based on the teaching design module and the teaching scenario requirements provided, and the details in each module are fully defined.

[0028] For ease of understanding, the process of steps S11 to S14 is described by way of example: For the teaching requirement of "determining the most effective course introduction method based on a specific knowledge point". First, in the module selection operation, "critical thinking" and "system thinking" were selected from the predefined reasoning modules, and the selected teaching design modules included "clearly defining learning outcomes and educational objectives" and "selecting or developing a teaching framework that supports teaching objectives". Subsequently, these selected modules were adjusted to descriptions that are more relevant to the current task. For example, "critical thinking" was adapted to "encourage course designers to critically evaluate the relevance and depth of each knowledge point to ensure that it is consistent with the course objectives", while "system thinking" focused on guiding designers to understand the broader impact of different content components. For the module adaptation stage, "clarifying learning outcomes" was adapted to "defining learning outcomes: assisting course designers to clarify the learning outcomes that the course aims to achieve", and "teaching framework" was adapted to "helping course designers select or develop a framework that effectively supports the integration of various knowledge points". Then, the first module after adaptation was transformed into a module structure, generating a structured instruction template containing role settings, target audiences, and specific teaching instructions, where the teaching instructions included "analyzing knowledge points" and "determining the course introduction method". Finally, through the instruction template content filling unit, detailed content is filled in for each part of the instruction template to generate a complete filled instruction template, including the role setting as "curriculum designer", the target audience as "educators, subject experts, students", and specific commands and rules.

[0029] Step S2: teaching design optimization; Step S21: selecting a module suitable for optimizing the filled instruction template from the inference module to obtain a selected second module; It can be understood that the filled instruction template obtained in the instructional design generation phase is used as the basis for module selection, and the module that can best help solve the current optimization task is selected from a set of reasoning modules, that is, the selected second module is obtained, such as the "design experiments to help solve the problem" or "list ideas for solving the problem" module.

[0030] Step S22: converting the selected second module from a general description into a detailed description for the target teaching scenario task to obtain an adapted second module; The selected reasoning module (i.e., the second module after selection) is adapted. In order to better meet the specific optimization requirements, the filled instruction template is used as the reference direction of the adaptation operation. After the adaptation operation, the general module description is adjusted to a refined description under a specific task scenario to obtain the adapted second module.

[0031] Step S23: converting the adapted detailed description of the second module into a suggestion template; The descriptions of these adapted reasoning modules (i.e., the adapted second modules) are converted into a suggestion template, which is presented in the form of key-value pairs, similar to the JSON format, to facilitate LLM understanding and further execution.

[0032] Step S24: using LLM to fill the suggestion template with specific suggestion content to obtain a filled suggestion template; After generating the suggestion template, use LLM to gradually fill in the corresponding values ​​according to the keys in the suggestion template until a complete and detailed filled suggestion template is obtained. The model needs to strictly follow the structure of the suggestion template to ensure that each key part is filled with detailed and accurate content. Each key-value pair contains enough details to support its optimization task.

[0033] Step S25: Optimizing each suggestion in the filled suggestion template to obtain an optimized instruction template; After obtaining the filled suggestion template, the initially generated instruction template is optimized based on each suggestion provided in the suggestion template through the guidance model to improve its quality, wherein the guidance model includes but is not limited to GPT-4o and DeepSeek-R1.

[0034] Furthermore, the optimization process includes: (1) Modify ambiguous instructions: In the initially generated instruction template, some instructions may affect the quality of the teaching design due to ambiguous descriptions or insufficient coverage. Modify the instructions in the filled instruction template according to the suggestions in the filled suggestion template.

[0035] (2) Eliminate redundant instructions: Unnecessary repeated instructions may appear in the suggested template, which increases the complexity of the teaching process. These redundant parts are identified and deleted through the guidance of the suggested template.

[0036] (3) Adding new instructions: The initially generated instruction template may contain missing teaching elements. Add new instructions to the filled instruction template based on the suggestions in the filled suggestion template.

[0037] For ease of understanding, the process of steps S21 to S25 is described by way of example: For the initially generated instruction template, first, when selecting the reasoning module, the second module after selection includes "How to design an experiment to help solve the problem?" and "List ideas for solving this problem and apply them one by one to see if there is any progress". The former is adapted to "How can I design an experiment to determine the most effective course introduction method for this specific knowledge point?", and the latter is adapted to "List ideas for introducing this knowledge point in the course". Then, the adapted second module is transformed into a suggestion template, which contains suggestions such as "Design an experiment to help determine the most effective course introduction method for this specific knowledge point", "Break the course introduction process into smaller and more manageable steps", "Integrate feedback to improve the course introduction method", and "Simplify the course introduction to make it more attractive and effective for learners". After the suggestion template content is filled, a complete filled suggestion template is formed. Subsequently, the filled suggestion template is optimized based on these suggestions. According to the suggestion "design a controlled experiment to compare different course introduction methods for specific knowledge points", the "consider the most effective course introduction method for a certain knowledge point" in the filled suggestion template was modified to "design a controlled experiment to determine the most effective course introduction method for each knowledge point"; according to the suggestion "create a step-by-step guide for implementing course introduction methods", the two redundant contents in the filled suggestion template were merged into "break down the course introduction process into smaller, manageable steps for effective implementation". The content of interactive teaching activities that attract learners is missing in the filled suggestion template. Based on the suggestion "develop a simplified course introduction process that focuses on key concepts and attracts learners through interactive activities", a new instruction was added: "consider incorporating gamification or interactive activities to attract learners through interactive experiences."

[0038] Step S3: Reflection on teaching design; Step S31: determining whether the content quantity of the filled instruction template and the optimized instruction template is greater than the preset change requirement; Step S32: if it is greater than the preset change requirement, then a global scoring mechanism and a local scoring mechanism are used to obtain a global score and a local score respectively; In some embodiments, a local scoring mechanism is used to evaluate the improvement of sub-parts of the instruction template, and the scoring is based on bonus items and deduction items. The sub-parts include role settings, target audiences, and multiple teaching behaviors. Bonus items include innovation and significant improvement, and deduction items include redundancy and reduced relevance. The global scoring mechanism is used to evaluate the integrity, relevance, accuracy, and creativity of the overall structure and content of the instruction template, and give a comprehensive score of 0 to 100.

[0039] Specifically, by comparing the filled instruction template and the optimized instruction template, the effectiveness of the optimization process is evaluated, and it is determined whether iterative optimization is needed. It is determined whether the content quantity of the optimized instruction template and the instruction template filled in the first stage is greater than the preset change requirement. If it is greater than the preset change requirement, the global scoring mechanism and the local scoring mechanism are used to obtain the global score and the local score respectively. If it is less than the preset change requirement, the reasoning module and the teaching design module are reselected for iteration to systematically evaluate the improvement quality of the optimized template at different levels. This scoring mechanism not only focuses on the overall changes, but also deeply analyzes the specific details of each part of the template to ensure the balance between the whole and the details. The scoring mechanism includes: (1) Local scoring mechanism: This mechanism aims to evaluate the improvement of each label in the instruction template (such as role setting, target audience, and multiple teaching behaviors). In local scoring, we define two types of scoring criteria: plus points (Grade A) and minus points (Grade B). Plus points are used to reward those parts that show innovation and significant improvement in content. In contrast, parts that have been slightly improved but have not yet achieved significant improvement will be given 1-2 points. Minus points are for parts that have changed less, have not shown effective improvement, or even have deteriorated after changes. Such instructions may have weakened their relevance to task requirements or introduced unnecessary complexity. Minus points are divided into -1, -2, and -3 points according to the severity of the regression, so as to quantify the regression of content and deviation from task objectives.

[0040] (2) Global scoring mechanism: It focuses more on the overall improvement quality of the teaching template, aiming to evaluate the overall changes between the new and old templates from a global perspective. The global scoring criteria involve comprehensive evaluation of multiple dimensions, including completeness, relevance, accuracy, and creativity. The completeness dimension emphasizes whether the template comprehensively covers all aspects of the teaching task, including objectives, content, methods, and evaluation, to ensure that no key teaching elements are omitted. The relevance dimension evaluates the consistency between the template content and the task requirements, ensuring that the design always revolves around the teaching objectives and actual tasks and avoids irrelevant content. The accuracy dimension ensures that the teaching content and strategies provided in the template are scientific and reasonable, and accurately reflect the teaching objectives and methods. Finally, the creativity dimension evaluates whether the template proposes innovative solutions in terms of teaching strategies and activity design. In addition, changes in the number and content of instructions are also key dimensions for evaluating template changes. The former measures whether the template has added new teaching steps or plans after optimization to better achieve the task objectives. The added steps can make the task execution more detailed and efficient, thereby improving the practicality and effectiveness of the template. The latter compares the content of the new and old templates and evaluates the impact of these changes on achieving the task objectives. Adding or modifying plans can make the task execution more precise and further optimize the teaching design. After a comprehensive evaluation of these dimensions, the total score will quantitatively reflect the overall improvement of the template, ranging from 0 to 100.

[0041] Step S33: configure the optimized instruction template corresponding to the global and local parts that are greater than the corresponding threshold into the teaching agent to obtain the final teaching design.

[0042] The results are classified according to the preset thresholds based on the global and local scoring results. This classification enables the instruction templates to be effectively assigned to different adjustment paths, and the following four strategies are implemented based on the classification results: Strategy 1: If both the global score and the local score are less than the threshold, it means that the template not only lacks reasonable support for the teaching objectives in the overall direction, but also the details of each module are not sufficient to effectively achieve the task objectives. In this case, a thorough reconstruction is required. We should start from the strategic level, re-select modules, and identify and define the reasoning modules and teaching design modules suitable for the current teaching objectives from scratch.

[0043] Strategy 2: If the local score is greater than the threshold but the global score is less than the threshold, it means that the labels of the instruction template have high quality in their respective detail processing, but there are still major defects in the overall logic or lack of teaching elements. At this time, it is necessary to adjust the strategic level, focusing on adjusting the overall direction. The reasoning module and the instructional design module should be reselected and organized.

[0044] Strategy 3: If the global score is greater than the threshold but the local score is less than the threshold, it means that the instruction template is relatively correct in the overall direction and can basically meet the teaching task objectives, but the description of the specific instructions in the instruction template is still not detailed enough and the practicality of the specific operation guidance is not strong. Tactical adjustments are needed to re-optimize and generate instruction templates based on the adapted modules, strengthen the integrity and pertinence of its content at the detail level, and add more detailed operation steps, teaching scenarios and interactive activities.

[0045] Strategy 4: If both the global score and the local score are greater than the threshold, it means that the instruction template has reached a high standard in all aspects of the overall design and specific implementation. In this case, the instruction template can be applied to the agent configuration.

[0046] For the above four strategies, the agent is configured using the instruction template that meets the threshold requirements (i.e., strategy four), generating an instructional design with a clear knowledge framework, comprehensive coverage of teaching elements, and in-depth design.

[0047] For ease of understanding, the process of steps S31 to S33 is described by way of example: The initial instruction template in the generation phase was compared with the template in the optimization phase. It shows that the optimized template has added a lot of details that are helpful for teaching, and the number of instructions has also increased. Next, the two templates were compared locally. In the role setting part, the optimized template described the role more specifically and clearly, and received a Grade A evaluation. In the target audience part, the optimized template further clarified the audience range and refined the object description, and also received a Grade A evaluation. Similar detailed evaluations were also conducted for other parts, such as the commands and rules in the teaching instructions. In terms of global comparison, the first is the change in quantity. Compared with the template in the generation phase, the optimized template has 18 more instructions and 3 instruction labels. Secondly, in terms of content changes, the optimized template is also significantly different from the template in the generation phase. The instructions in the original template are described in more detail, and the depth of the content is expanded. In addition, the optimized template has also been significantly improved in terms of task relevance, comprehensiveness and creativity. Based on the above comparison results, the local scoring template and the global scoring template are used to assign scores to the optimized instruction template to quantify its improvement effect. After the score distribution, the final global score was 90 and the local score was 0.607, both exceeding the preset thresholds of 85 and 0.5. This shows that the optimized instruction template has achieved the expected improvement effect in both details and overall direction. Therefore, the optimized instruction template was integrated into the agent configuration and interacted with the agent to generate the final teaching design.

[0048] Example 2 See also Figure 3 , which is a structural diagram of a teaching design generation, optimization and reflection system based on AI chain drive proposed in the second embodiment of the present application, wherein the system comprises a teaching design generation module 100, a teaching design optimization module 200 and a teaching design reflection module 300, wherein: The teaching design generation module 100 includes: A first module selection unit 101 is used to select a module suitable for a target teaching scenario task from predefined reasoning modules and teaching design modules to obtain a selected first module; A first module adaptation unit 102, configured to convert the selected first module from a general description into a detailed description for a target teaching scenario task, so as to obtain an adapted first module; A first module structuring unit 103, configured to convert the adapted first module's detailed description into an instruction template; A first content filling unit 104 is used to fill the instruction template with specific instruction content using LLM based on the instruction design module and the target teaching scenario task to obtain a filled instruction template; The teaching design optimization module 200 includes: A second module selection unit 201, configured to select a module suitable for optimizing the filled instruction template from the reasoning module to obtain a selected second module; A second module adaptation unit 202, configured to convert the selected second module from a general description into a detailed description for a target teaching scenario task, so as to obtain an adapted second module; A second module structuring unit 203, configured to convert the adapted fine description of the second module into a suggestion template; A second content filling unit 204 is used to fill the suggestion template with specific suggestion content using LLM to obtain a filled suggestion template; The optimization filling unit 205 is used to optimize each suggestion in the filled suggestion template to obtain an optimized instruction template; The teaching design reflection module 300 includes: The judgment and classification unit 301 is used to judge whether the content quantity of the filled instruction template and the optimized instruction template is greater than the preset change requirement; if it is greater than the preset change requirement, a global scoring mechanism and a local scoring mechanism are used to obtain a global score and a local score respectively; and classification is performed according to the size relationship between the global score and the local score and the corresponding threshold respectively; The configuration unit 302 is used to configure the optimized instruction template corresponding to when the global score and the local score are both greater than the corresponding threshold into the teaching agent to obtain the final teaching design.

[0049] In the embodiment of the present application, a teaching design generation, optimization and reflection system based on AI chain drive can be a device, or a component, integrated circuit, or chip in a terminal. The device can be a mobile electronic device or a non-mobile electronic device. Exemplarily, the mobile electronic device can be a mobile phone, a tablet computer, a laptop computer, a PDA, a car electronic device, a wearable device, an ultra-mobile personal computer (UMPC), a netbook or a personal digital assistant (PDA), etc., and the non-mobile electronic device can be a server, a network attached storage (NAS), a personal computer (PC), etc., which is not specifically limited in the embodiment of the present application.

[0050] In the embodiment of the present application, a teaching design generation, optimization and reflection system based on AI chain drive can be a device with an operating system. The operating system can be an Android operating system, an iOS operating system, or other possible operating systems, which are not specifically limited in the embodiment of the present application.

[0051] The teaching design generation, optimization and reflection system based on AI chain drive provided in the embodiment of the present application can achieve Figure 1 In the method embodiment, each process of implementing the teaching design generation, optimization and reflection method based on AI chain drive will not be repeated here to avoid repetition.

[0052] Optionally, an embodiment of the present application also provides an electronic device, including a processor, a memory, and a program or instruction stored in the memory and executable on the processor. When the program or instruction is executed by the processor, each process of the above-mentioned AI chain-driven teaching design generation, optimization and reflection method embodiment is implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

[0053] An embodiment of the present application also provides a readable storage medium, on which a program or instruction is stored. When the program or instruction is executed by a processor, each process of the above-mentioned embodiment of the teaching design generation, optimization and reflection method driven by an AI chain is implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

[0054] The processor is a processor in the electronic device described in the above embodiment. The readable storage medium includes a computer readable storage medium, such as a computer read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0055] It should be noted that, in this article, the terms "comprise", "include" or any other variant thereof are intended to cover non-exclusive inclusion, so that the process, method, article or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise one..." do not exclude the presence of other identical elements in the process, method, article or device including the element. In addition, it should be pointed out that the scope of the methods and devices in the embodiments of the present application is not limited to performing functions in the order shown or discussed, and may also include performing functions in a substantially simultaneous manner or in reverse order according to the functions involved, for example, the described method may be performed in an order different from that described, and various steps may also be added, omitted, or combined. In addition, the features described with reference to certain examples may be combined in other examples.

[0056] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, a magnetic disk, or an optical disk), and includes a number of instructions for a terminal (which can be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods described in each embodiment of the present application.

[0057] The embodiments of the present application are described above in conjunction with the accompanying drawings, but the present application is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present application, ordinary technicians in this field can also make many forms without departing from the purpose of the present application and the scope of protection of the claims, all of which are within the protection of the present application.

Claims

1. A method for generating, optimizing and reflecting instructional designs based on AI chain drive, characterized in that: include: Selecting a module suitable for a target teaching scenario task from predefined reasoning modules and teaching design modules to obtain a selected first module; The selected first module is converted from a general description into a detailed description for a target teaching scenario task to obtain an adapted first module; the detailed description of the adapted first module is converted into an instruction template; based on the teaching design module and the target teaching scenario task, the instruction template is filled with specific instruction content using LLM to obtain a filled instruction template; Selecting a module suitable for optimizing the filled instruction template from the reasoning module to obtain a selected second module; converting the selected second module from a general description into a detailed description for a target teaching scenario task to obtain an adapted second module; converting the detailed description of the adapted second module into a suggestion template; using LLM to fill the suggestion template with specific suggestion content to obtain a filled suggestion template; Optimizing each suggestion in the filled suggestion template to obtain an optimized instruction template; Determine whether the content quantity of the filled instruction template and the optimized instruction template is greater than the preset change requirement; if it is greater than the preset change requirement, use the global scoring mechanism and the local scoring mechanism to obtain the global score and the local part respectively; configure the optimized instruction template corresponding to the time when the global score and the local part are both greater than the corresponding threshold into the teaching intelligent body to obtain the final teaching design.

2. According to claim 1, a method for generating, optimizing and reflecting teaching designs based on AI chain drive, characterized in that: In the step of converting the adapted detailed description of the first module into an instruction template: The instruction template includes role settings, target audiences, and multiple teaching behaviors, wherein the teaching behaviors represent instruction tags of the teaching process, and each instruction tag represents an independent teaching link; Each instruction tag includes specific commands and rules. The commands are used to define specific execution behaviors in the teaching process and are arranged in a logical order; the rules ensure that the implementation of the instructions complies with the predetermined logical order and teaching standards.

3. According to claim 2, a method for generating, optimizing and reflecting teaching designs based on AI chain drive, characterized in that: In the steps of using the global scoring mechanism and the local scoring mechanism to obtain the global score and the local score respectively: The local scoring mechanism is used to evaluate the improvement of the sub-parts of the instruction template, and the scoring is performed according to the bonus items and the deduction items, the sub-parts include role setting, target audience and multiple teaching behaviors, the bonus items include innovation and significant improvement, and the deduction items include redundancy and reduced relevance; The global scoring mechanism is used to evaluate the integrity, relevance, accuracy and creativity of the overall structure and content of the instruction template, and to give a comprehensive score from 0 to 100.

4. According to claim 1, a method for generating, optimizing and reflecting teaching designs based on AI chain drive, characterized in that: In the step of converting the adapted fine description of the second module into a suggestion template: The suggestion template is presented in the form of key-value pairs, so that LLM can understand and execute step-by-step filling of key values.

5. According to claim 1, a method for generating, optimizing and reflecting teaching designs based on AI chain drive, characterized in that: The step of optimizing each suggestion in the filled suggestion template to obtain an optimized instruction template comprises: Optimizing each suggestion in the filled suggestion template by a guidance model, wherein the guidance model includes GPT-4o; The optimization process includes modifying ambiguous instructions, eliminating redundant instructions, and adding new instructions.

6. According to claim 1, a method for generating, optimizing and reflecting teaching designs based on AI chain drive, characterized in that: The step of judging whether the content quantity of the filled instruction template and the optimized instruction template is greater than the preset change requirement; if it is greater than the preset change requirement, adopting the global scoring mechanism and the local scoring mechanism to obtain the global score and the local score respectively also includes: If it is less than the preset change requirement, the reasoning module and teaching design module are reselected for iteration.

7. The method for generating, optimizing and reflecting teaching designs based on AI chain drive according to claim 1 is characterized in that: Before configuring the optimized instruction template corresponding to when the global score and the local score are both greater than the corresponding threshold into the teaching agent to obtain the final teaching design, the step also includes: Classification is performed according to the magnitude relationship between the global score and the local score and the corresponding thresholds; If both the global score and the local score are less than the corresponding threshold, the reasoning module and the instructional design module are reselected; If the local score is greater than the corresponding threshold and the global score is less than the corresponding threshold, the reasoning module and the instructional design module are reselected; If the global score is greater than the corresponding threshold and the local score is less than the corresponding threshold, the instruction template is re-optimized and generated.

8. An AI-driven instructional design generation, optimization and reflection system, characterized in that: The system includes an instructional design generation module, an instructional design optimization module and an instructional design reflection module, wherein: The teaching design generation module includes: A first module selection unit, used to select a module suitable for a target teaching scenario task from predefined reasoning modules and teaching design modules to obtain a selected first module; A first module adaptation unit, used to convert the selected first module from a general description into a detailed description for a target teaching scenario task, so as to obtain an adapted first module; a first module structuring unit, configured to convert the adapted first module's detailed description into an instruction template; A first content filling unit is used to fill the instruction template with specific instruction content using LLM based on the instruction design module and the target teaching scenario task to obtain a filled instruction template; The teaching design optimization module includes: a second module selection unit, configured to select a module suitable for optimizing the filled instruction template from the reasoning module to obtain a selected second module; A second module adaptation unit, used to convert the selected second module from a general description into a detailed description for a target teaching scenario task, so as to obtain an adapted second module; a second module structuring unit, for converting the adapted second module's detailed description into a suggestion template; A second content filling unit is used to fill the suggestion template with specific suggestion content using the LLM to obtain a filled suggestion template; An optimization filling unit, used for optimizing each suggestion in the filled suggestion template to obtain an optimized instruction template; The teaching design reflection module includes: A judgment classification unit is used to judge whether the content quantity of the filled instruction template and the optimized instruction template is greater than the preset change requirement; if it is greater than the preset change requirement, a global scoring mechanism and a local scoring mechanism are used to obtain a global score and a local score respectively; The configuration unit is used to configure the optimized instruction template corresponding to when the global score and the local score are both greater than the corresponding threshold into the teaching agent to obtain the final teaching design.

9. An electronic device, characterized in that: It includes a processor, a memory, and a program or instruction stored in the memory and executable on the processor. When the program or instruction is executed by the processor, the steps of a teaching design generation, optimization and reflection method based on AI chain drive are implemented as described in any one of claims 1 to 7.

10. A readable storage medium, characterized in that: The readable storage medium stores programs or instructions, and when the programs or instructions are executed by the processor, the steps of the teaching design generation, optimization and reflection method based on AI chain drive as described in any one of claims 1 to 7 are implemented.

Images