CTS2 differential intelligent course evaluation method based on multi-modal large model

Abstract

The invention discloses a CTS2 differential intelligent course evaluation method based on a multi-modal large model. According to the invention, a multi-modal large model is introduced, and class evaluation analysis is carried out by fusing various modal data such as texts, audios, images and the like. Compared with a traditional lesson evaluation system which only processes a single mode (such as text or audio), the introduction of the multi-mode data enables the lesson evaluation result to be more comprehensive and accurate. According to the method, classroom teaching activities can be understood from multiple angles, the objectivity and diversity of class evaluation are remarkably improved, various tiny differences in classroom teaching can be accurately recognized and interpreted through the data processing and analysis capability of a multi-modal large model, and differentiated class evaluation results are generated. Experimental data show that the system can significantly improve the accuracy of course evaluation, and particularly in a scientific experiment course, a course evaluation result generated by the system shows high accuracy and consistency in multiple dimensions such as experimental design, teaching methods and student participation.

Description

Technical Field

[0001] The present invention belongs to the technical field of teaching evaluation, and specifically is a CTS2 differentiated intelligent evaluation method based on a multimodal large model. Background Art

[0002] Lesson evaluation is an important means to improve teaching quality. Through lesson evaluation, teachers can understand the strengths and weaknesses of their teaching practice, so as to continuously improve teaching methods and strategies. Although traditional lesson evaluation promotes the professional development of teachers, there are also some problems in practice, such as heavy tasks, unclear goals, and lack of professionalism, which may lead to superficial effects of lesson evaluation activities and difficulty in obtaining in-depth and effective feedback. At present, data-based classroom analysis methods have gradually been applied in the process of teacher teaching evaluation, such as Flanders' classroom interaction analysis system and ST analysis method, which capture the verbal behavior of teachers and students through quantitative and qualitative dimensions and test the impact of social emotional environment on students. After entering the digital age, technologies such as artificial intelligence have been applied to classroom teaching research, forming a new paradigm of classroom intelligent analysis, such as the interactive analysis coding system ITIAS based on information technology and the reliable COP based on big data analysis. This paradigm provides evidence-based classroom diagnostic reports through structured analysis of multimodal data to promote teaching reflection and improvement. Primary and secondary schools mainly use the TEAM Model Socrates AI teacher behavior data analysis system and the "New Learning Road" teacher behavior data analysis system. It mainly analyzes and interprets teachers’ classroom recordings or real-time classes, and provides quantitative classroom data for teachers to conduct further research and reflection.

[0003] However, classroom teaching evaluation plays a key role in the field of education. However, traditional classroom evaluation methods, which mainly rely on manual observation and standardized evaluation criteria, have the following shortcomings:

[0004] 1. Observation-dependent and highly subjective: Traditional evaluation methods often rely on the observation of teachers or evaluators. The evaluation results are easily affected by personal subjective factors and lack objectivity and consistency. This method is not only time-consuming and labor-intensive, but also has low evaluation efficiency.

[0005] 2. Standardized evaluation lacks flexibility: Existing automated evaluation systems are usually based on fixed standardized indicators, which are difficult to fully reflect the diversity and complexity of classroom teaching. These systems are unable to provide personalized evaluations when faced with differences in different subjects, grades, teacher development stages, and cultural backgrounds, resulting in a disconnect between evaluation results and actual teaching needs.

[0006] 3. Limited degree of automation: Although some systems have introduced data-based analysis methods, most lesson evaluation systems are still unable to achieve comprehensive automated analysis of the classroom, especially in the processing and understanding of multimodal data such as audio and video. Summary of the invention

[0007] The purpose of the present invention is to provide a CTS2 differentiated intelligent course evaluation method based on a multimodal large model in order to solve the above-mentioned problems.

[0008] The technical solution adopted by the present invention is as follows: a CTS2 differentiated intelligent course evaluation method based on a multimodal large model, the method comprising the following steps:

[0009] S1: First, collect classroom data and obtain classroom audio and video materials from the classroom recording and broadcasting system;

[0010] S2: Then pre-process the classroom audio and video materials, convert the audio into text, and form classroom dialogue text materials;

[0011] S3: intercept frames from the classroom video at fixed time intervals to form classroom image materials. Store the classroom data before and after processing in objects;

[0012] S4: Conduct CTS2 analysis and construct an indicator system. The teaching and research staff collects differentiated evaluation information on six core elements of differentiated evaluation, namely cultural background, classroom type, teaching topic, teacher development stage, subject, and academic stage, and designs differentiated evaluation indicators and an indicator system based on this.

[0013] S5: The pre-processed text and pictures of the classroom data, as well as all the evaluation index details under the differentiated evaluation system, are input into the multimodal large model through prompt words to generate evaluation results including evaluation index scores, result interpretation, and evidence-based timestamps;

[0014] S6: Finally, the above results are visualized through the system front end. In addition, the accumulation of course evaluation data can also serve as the data basis for tuning multimodal large models.

[0015] In a preferred embodiment, in step S1, the classroom data also includes the teacher's lesson plan document.

[0016] In a preferred implementation, in step S4, the course evaluation indicator pool may store common course evaluation indicators that can be shared among different course evaluation indicator systems.

[0017] In a preferred embodiment, in step S4, each indicator can allow the multimodal large model to combine with CTS2 information, such as the teacher and teaching topic information of each class, to dynamically generate more detailed evaluation criteria to improve the differentiated course evaluation effect.

[0018] In a preferred embodiment, in step S4, the cultural background includes national policies at the macro level and school-based culture at the micro level. Among them, national policies such as the new curriculum standards can be used to guide teaching practice and teaching evaluation, and are important guides for implementing differentiated evaluation. School-based culture refers to the specific spiritual environment and cultural atmosphere of the school, which can indirectly or directly affect teachers' teaching methods and students' learning experience.

[0019] In a preferred embodiment, in step S4, different subjects have their own characteristics and requirements, which have a significant impact on the design, implementation and evaluation of classroom teaching. For example, the language subject emphasizes the cultivation of reading, writing, oral expression and literary analysis skills. The classroom may focus more on the cultivation of text interpretation, creative writing and critical thinking. The science classroom emphasizes experimental operation, observation and recording, and scientific inquiry methods, and encourages students to understand natural phenomena through experiments and exploration.

[0020] In a preferred embodiment, in step S4, in terms of learning stage, students in different learning stages have different characteristics in terms of cognitive development level, learning needs, etc. For example, primary school students are in the concrete operation stage and rely more on intuition and concrete experience; middle school students gradually transition to the formal operation stage and can handle more abstract concepts. The cognitive learning rules of students in different learning stages should be fully respected to promote the improvement of students' literacy.

[0021] In a preferred embodiment, in step S4, in terms of the teacher development stage, "the professional growth of teachers is an ongoing process that requires continuous learning, reflection and practice." Studies have found that the classroom teaching behaviors of groups of teachers at different stages of teacher professional development will show group differences. For example, novice teachers usually perform well in teacher quality, but they still need to be strengthened in the two dimensions of teaching content and classroom operation. Therefore, each teacher has different focuses and needs at different stages of growth.

[0022] In a preferred embodiment, in step S4, the teaching subject refers to the subject of the specific teaching content. Teachers need to use different teaching methods and teaching techniques to achieve teaching objectives according to the specific teaching subject. The evaluation of specific teaching content can ensure that the teaching activities meet the established educational standards and quality requirements, and is a reflection of the teacher's ability to design and integrate teaching content.

[0023] In a preferred embodiment, in step S4, in terms of class type, the types of classroom teaching are divided according to different teaching objectives, content, methods and organizational forms, such as lectures focus on knowledge transfer, discussion classes emphasize student thinking and communication, and experimental classes focus on the cultivation of practical operation ability. Different class types have different focuses on teaching content. Therefore, it is necessary to pay attention to the evaluation differences of different class types in class evaluation.

[0024] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0025] 1. In the present invention, a multimodal large model is introduced, and class evaluation analysis is performed by integrating multiple modal data such as text, audio, and images. Compared with the traditional class evaluation system that only processes a single modality (such as text or audio), the introduction of multimodal data makes the class evaluation results more comprehensive and accurate. It can understand classroom teaching activities from multiple perspectives, such as identifying multidimensional information such as teacher explanations, student behaviors, and classroom interactions, which significantly improves the objectivity and diversity of class evaluation. Through the data processing and analysis capabilities of the multimodal large model, it can accurately identify and interpret various nuances in classroom teaching, and generate differentiated class evaluation results. Experimental data show that the system can significantly improve the accuracy of class evaluation, especially in scientific experiment classes. The class evaluation results generated by the system show a high degree of accuracy and consistency in multiple dimensions such as experimental design, teaching methods, and student participation.

[0026] 2. In the present invention, the CTS2 differentiated course evaluation model is the core innovation of the present invention, which emphasizes the personalization and diversification of course evaluation standards. By comprehensively considering factors such as cultural background, subject, stage of study, and stage of teacher development, the model can provide a course evaluation plan that is highly targeted and in line with actual teaching situations. This differentiated course evaluation breaks through the traditional "one-size-fits-all" fixed standard model, adapts to different teaching environments and needs, and effectively supports the professional development of teachers. In a real classroom environment, the course evaluation results of this system are highly consistent with the manual course evaluation results of multiple experts, proving its reliability and effectiveness in practical applications. In addition, the system's multi-dimensional analysis function can provide teachers with more specific and targeted improvement suggestions, further improving the quality of teaching.

[0027] 3. In the present invention, an intelligent class evaluation system based on a multimodal large model is constructed to improve the automation of class evaluation. By automatically collecting classroom data, preprocessing audio and video information, and combining the intelligent analysis capabilities of the multimodal model, the system can automatically generate class evaluation results, reduce manual intervention, and improve the efficiency and accuracy of class evaluation. At the same time, it provides detailed evidence-based analysis and personalized improvement suggestions. The automated class evaluation function greatly reduces the subjective bias of manual class evaluation and improves the efficiency of class evaluation. Through real-time data collection and analysis, the system can instantly generate teaching feedback, help teachers quickly understand the classroom effect, and adjust teaching strategies in a timely manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 The CTS2 differentiated lesson evaluation system model of the present invention;

[0029] Figure 2 It is the technical framework of the CTS2 differentiated course evaluation system in the present invention;

[0030] Figure 3 Create a new lesson evaluation indicator page in the present invention;

[0031] Figure 4 Create a new lesson review template page in the present invention;

[0032] Figure 5 Creating an automated lesson evaluation task in the present invention;

[0033] Figure 6 This is the lesson review report page in the present invention;

[0034] Figure 7 This is the course portrait information page in the present invention;

[0035] Figure 8 This is the expert manual lesson evaluation interface in the present invention. DETAILED DESCRIPTION

[0036] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0037] Reference Figure 1-8 ,

[0038] Example:

[0039] A CTS2 differentiated intelligent course evaluation method based on a multimodal large model, the method comprising the following steps:

[0040] S1: First, collect classroom data and obtain classroom audio and video materials from the classroom recording and broadcasting system;

[0041] S2: Then pre-process the classroom audio and video materials, convert the audio into text, and form classroom dialogue text materials;

[0042] S3: intercept frames from the classroom video at fixed time intervals to form classroom image materials. Store the classroom data before and after processing in objects;

[0043] S4: Conduct CTS2 analysis and construct an indicator system. The teaching and research staff collects differentiated evaluation information on six core elements of differentiated evaluation, namely cultural background, classroom type, teaching topic, teacher development stage, subject, and academic stage, and designs differentiated evaluation indicators and an indicator system based on this.

[0044] S5: The pre-processed text and pictures of the classroom data, as well as all the evaluation index details under the differentiated evaluation system, are input into the multimodal large model through prompt words to generate evaluation results including evaluation index scores, result interpretation, and evidence-based timestamps;

[0045] S6: Finally, the above results are visualized through the system front end. In addition, the accumulation of course evaluation data can also serve as the data basis for tuning multimodal large models.

[0046] In step S1, the classroom data also includes the teacher's lesson plan document.

[0047] In step S4, the course evaluation indicator pool may store common course evaluation indicators that can be shared among different course evaluation indicator systems.

[0048] In step S4, each indicator can allow the multimodal large model to combine with CTS2 information, such as the teacher and teaching topic information of each class, to dynamically generate more detailed evaluation criteria to improve the effect of differentiated class evaluation.

[0049] In step S4, the cultural background includes national policies at the macro level and school-based culture at the micro level. Among them, national policies such as the new curriculum standards can be used to guide teaching practice and teaching evaluation, and are an important guide for implementing differentiated evaluation. School-based culture refers to the specific spiritual environment and cultural atmosphere of the school, which can indirectly or directly affect teachers' teaching methods and students' learning experience.

[0050] In step S4, different subjects have their own characteristics and requirements, which have a significant impact on the design, implementation and evaluation of classroom teaching. For example, the language subject emphasizes the cultivation of reading, writing, oral expression and literary analysis skills. The classroom may pay more attention to the cultivation of text interpretation, creative writing and critical thinking. Science classes emphasize experimental operation, observation and recording, and scientific inquiry methods, encouraging students to understand natural phenomena through experiments and exploration.

[0051] In step S4, students in different stages have different characteristics in terms of cognitive development level and learning needs. For example, primary school students are in the concrete operation stage and rely more on intuition and concrete experience; middle school students gradually transition to the formal operation stage and can handle more abstract concepts. We should fully respect the cognitive learning rules of students in different stages and promote the improvement of students' literacy.

[0052] In step S4, regarding the teacher development stage, "the professional growth of teachers is an ongoing process that requires continuous learning, reflection and practice." Existing studies have found that the classroom teaching behaviors of groups of teachers at different stages of teacher professional development will show group differences. For example, novice teachers usually perform well in teacher quality, but they still need to improve in the two dimensions of teaching content and classroom operation. Therefore, each teacher has different focuses and needs at different stages of growth.

[0053] In step S4, the teaching topic refers to the topic of the specific teaching content. Teachers need to use different teaching methods and teaching techniques to achieve teaching objectives according to the specific teaching topic. The evaluation of specific teaching content can ensure that the teaching activities meet the established educational standards and quality requirements, and is a reflection of the teacher's ability to design and integrate teaching content.

[0054] In step S4, in terms of class types, the types of classroom teaching are divided according to the teaching objectives, content, methods and organizational forms. For example, lectures focus on knowledge transfer, discussion classes emphasize student thinking and communication, and experimental classes focus on the cultivation of practical operation capabilities. Different class types have different focuses on teaching content. Therefore, it is necessary to pay attention to the evaluation differences of different class types in class evaluation.

[0055] This method can form a multi-modal evaluation basis by creating new indicators, screening indicators and constructing a multi-dimensional evaluation template. The evaluation indicators can set the indicator type, such as Figure 3 When evaluating lessons, the method will use a multimodal large model for targeted evaluation according to different indicator types.

[0056] On the basis of the existing evaluation indicators, a new evaluation template is created. The additional prompt words can be added to explain the cultural environment, course requirements and other background information of the evaluation template, so as to help the multimodal large model evaluate the course more differentiated and accurately. At the same time, this method also supports the division of multiple evaluation indicators into different dimensions, and the evaluation indicators are selected under different dimensions to construct the evaluation template, such as Figure 4 .

[0057] After creating the lesson evaluation template, you can use the template to carry out automated lesson evaluation, such as Figure 5 ,This method can select the evaluation method and evaluate the class according to different ,classroom multimodal video stream information. After adding and selecting various ,basic information, you can click Submit and wait for about 30 minutes to get the ,evaluation result.

[0058] Click the management button in the operation bar on the right side of the course evaluation record to view the results. This method provides two visual results, including course portrait information and course evaluation report. The report is in pdf format, which is convenient for users to download and view. Figure 6, click the "Download Report" button in the upper right corner to download. Figure 7 The course portrait information is shown. On this page, we can also click the teaching reflection button at the bottom right to reflect and summarize the suggestions and analysis of the course evaluation results.

[0059] In addition to the above automatic evaluation functions, this method also supports teaching and research experts to manually evaluate and score the uploaded classroom content, such as Figure 8 During the expert's classroom observation, you can click the new record button below the video on the left to record the process for the timestamp you are currently watching. The right side also supports teaching and research experts to score various indicators. Finally, both the expert's manual scoring and the method's automated scoring will be presented in the lesson evaluation report and results.

[0060] From the above, we can know that:

[0061] In the present invention, a multimodal large model is introduced, and class evaluation analysis is performed by integrating multiple modal data such as text, audio, and images. Compared with the traditional class evaluation system that only processes a single modality (such as text or audio), the introduction of multimodal data makes the class evaluation results more comprehensive and accurate. It can understand classroom teaching activities from multiple perspectives, such as identifying multidimensional information such as teacher explanations, student behaviors, and classroom interactions, which significantly improves the objectivity and diversity of class evaluation. Through the data processing and analysis capabilities of the multimodal large model, it can accurately identify and interpret various nuances in classroom teaching, and generate differentiated class evaluation results. Experimental data show that the system can significantly improve the accuracy of class evaluation, especially in scientific experiment classes. The class evaluation results generated by the system show high accuracy and consistency in multiple dimensions such as experimental design, teaching methods, and student participation.

[0062] In the present invention, the CTS2 differentiated course evaluation model is the core innovation of the present invention, which emphasizes the personalization and diversification of course evaluation standards. By comprehensively considering factors such as cultural background, subject, stage of study, and stage of teacher development, the model can provide a highly targeted course evaluation plan that conforms to the actual teaching situation. This differentiated course evaluation breaks through the traditional "one-size-fits-all" fixed standard model, adapts to different teaching environments and needs, and effectively supports the professional development of teachers. In a real classroom environment, the course evaluation results of this system are highly consistent with the manual course evaluation results of multiple experts, proving its reliability and effectiveness in practical applications. In addition, the system's multi-dimensional analysis function can provide teachers with more specific and targeted improvement suggestions, further improving the quality of teaching.

[0063] In the present invention, an intelligent class evaluation system based on a multimodal large model is constructed to improve the automation of class evaluation. By automatically collecting classroom data, preprocessing audio and video information, and combining the intelligent analysis capabilities of the multimodal model, the system can automatically generate class evaluation results, reduce manual intervention, improve class evaluation efficiency and accuracy, and provide detailed evidence-based analysis and personalized improvement suggestions. The automated class evaluation function greatly reduces the subjective bias of manual class evaluation and improves class evaluation efficiency. Through real-time data collection and analysis, the system can instantly generate teaching feedback, help teachers quickly understand the classroom effect, and adjust teaching strategies in a timely manner.

[0064] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the term "comprises" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0065] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A CTS2 differentiated intelligent course evaluation method based on a multimodal large model, characterized by: The method comprises the following steps: S1: First, collect classroom data and obtain classroom audio and video materials from the classroom recording and broadcasting system; S2: Then pre-process the classroom audio and video materials, convert the audio into text, and form classroom dialogue text materials; S3: intercepting frames from the classroom video at fixed time intervals to form classroom image materials; storing the classroom data before and after processing in objects; S4: Conduct CTS2 analysis and construct an indicator system. The teaching and research staff will collect differentiated evaluation information on six core elements of differentiated evaluation, namely cultural background, classroom type, teaching topic, teacher development stage, subject, and academic stage, and design differentiated evaluation indicators and indicator systems based on this. S5: The pre-processed text and pictures of the classroom data, as well as all the evaluation index details under the differentiated evaluation system, are input into the multimodal large model through prompt words to generate evaluation index scores, result interpretation, and evidence-based time-stamped evaluation results; S6: Finally, the above results are visualized through the system front end; in addition, the accumulation of lesson evaluation data can also serve as the data basis for tuning multimodal large models.

2. The CTS2 differentiated intelligent course evaluation method based on a multimodal large model as claimed in claim 1, characterized in that: In step S1, the classroom data also includes the teacher's lesson plan document.

3. The CTS2 differentiated intelligent course evaluation method based on a multimodal large model as claimed in claim 1, characterized in that: In step S4, the course evaluation indicator pool may store common course evaluation indicators that can be shared among different course evaluation indicator systems.

4. The CTS2 differentiated intelligent course evaluation method based on a multimodal large model as claimed in claim 1, characterized in that: In step S4, each indicator can allow the multimodal large model to combine with CTS2 information, including the teacher and teaching topic information of each class, to dynamically generate more detailed evaluation criteria to improve the differentiated class evaluation effect.

5. The CTS2 differentiated intelligent course evaluation method based on a multimodal large model as claimed in claim 1, characterized in that: In step S4, the cultural background includes national policies at the macro level and school-based culture at the micro level.

6. The CTS2 differentiated intelligent course evaluation method based on a multimodal large model as claimed in claim 1, characterized in that: In step S4, in terms of subjects, different subjects have their own characteristics and requirements, which have a significant impact on the design, implementation and evaluation of classroom teaching.

7. The CTS2 differentiated intelligent course evaluation method based on a multimodal large model as claimed in claim 1, characterized in that: In step S4, in terms of academic stage, students in different academic stages have different characteristics in terms of cognitive development level and learning needs.

8. The CTS2 differentiated intelligent course evaluation method based on a multimodal large model as claimed in claim 1, characterized in that: In step S4, in terms of teacher development stage, "the professional growth of teachers is an ongoing process that requires continuous learning, reflection and practice;" existing studies have found that the classroom teaching behaviors of groups of teachers at different stages of teacher professional development will show group differences.

9. The CTS2 differentiated intelligent course evaluation method based on a multimodal large model as claimed in claim 1, characterized in that: In step S4, the teaching topic refers to the topic of the specific teaching content; teachers need to adopt different teaching methods and teaching techniques to achieve teaching objectives according to the specific teaching topic; the evaluation of specific teaching content ensures that the teaching activities meet the established educational standards and quality requirements, which reflects the teacher's ability to design and integrate teaching content.

10. The CTS2 differentiated intelligent course evaluation method based on a multimodal large model as claimed in claim 1, characterized in that: In step S4, in terms of class types, the types of classroom teaching are divided according to different teaching objectives, content, methods and organizational forms. For example, lectures focus on knowledge transfer, discussion classes emphasize student thinking and communication, and experimental classes focus on the cultivation of practical operation capabilities. Different class types have different emphases on teaching content. Therefore, it is necessary to pay attention to the evaluation differences of different class types in class evaluation.

Images