Classroom attention traction method and system based on large model dynamic content generation

Abstract

The invention discloses a classroom attention traction method based on large model dynamic content generation, and relates to the technical field of intelligent education, and the method comprises the following steps: S1, monitoring the classroom attention state of a student in real time, and obtaining the real-time state information of the student; s2, current teaching content is acquired in real time, and key knowledge points are extracted; s3, when the attention state of the student is detected to be distraction, combining the student archive information, the key knowledge points and the cue word template, and constructing a situational generation instruction; and S4, inputting the generation instruction into the language model, and outputting the personalized attention traction content for the student. By generating the personalized attention traction content which accords with the interest of the student and is related to the knowledge points, the blocking intervention is changed into interesting guidance, the psychological resistance of the student is remarkably reduced, the teaching experience is improved, and the teaching experience is improved. Moreover, the whole intervention process is completed silently through the student terminal, the teaching of the teacher and the listening of other students are not interrupted, and the overall continuity of the classroom is maintained.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of intelligent education technology, and in particular to a classroom attention traction method and system based on large model dynamic content generation. BACKGROUND

[0002] In today's education field, classroom teaching as the core link of knowledge transmission and student ability cultivation, its quality and effect are directly related to the students' academic achievement and comprehensive quality development, however, in the actual classroom teaching process, the problem of students' attention dispersion is widespread, which seriously affects the teaching efficiency and learning effect. In order to effectively solve this problem and improve the attractiveness and effectiveness of classroom teaching, various classroom attention management technologies have emerged.

[0003] At present, the attention management technology in the classroom teaching environment has the following problems: (1) Observation intervention mechanism based on teacher experience: this method relies on teacher's subjective observation and experience judgment, and identifies student's abnormal behavior through visual scanning; but in the large class teaching scene, the teacher's attention resources are limited, it is difficult to achieve full coverage, and oral reminders will damage the teaching continuity, and even have negative impact on the teacher-student relationship; (2) Automatic warning system based on classroom behavior recognition detection: computer vision, sensor and other technical means are used to realize the automatic recognition of student behavior, and a standardized intervention mode is carried out; the standardized intervention mode ignores the complexity of the teaching scene, and the unified reminder content cannot distinguish individual differences, and the mechanical intervention is easy to cause psychological resistance, and the long-term effect is significantly attenuated; and the reminder content usually uses a general template, which lacks semantic association with the knowledge points being taught. This disconnection leads to intervention only interrupting the behavior level, and cannot guide students to refocus their thinking on learning content, making it difficult to achieve real cognitive reconstruction; this "one-size-fits-all" implementation strategy does not fully consider the individual differences of students, and the system cannot adjust the intervention strategy according to the students' knowledge base, learning style, interest preference and other personalized factors, affecting the intervention effect; (3) Programmed interaction mode based on preset content: to improve classroom participation, teachers need to preset interactive links such as questions and answers, voting and other interactive links and embed them in specific teaching nodes; however, this fixed interaction time point cannot adapt to the real-time fluctuations of classroom attention, and the static rule base is difficult to cover the diversified teaching scenarios in teaching, and this rigid structure leads to poor performance of the system in the face of complex and variable actual teaching situations; In summary, the existing classroom attention management technology has obvious deficiencies in intervention strategy adaptability, teaching content relevance, personalized support capability and system response mechanism, and it is difficult to meet the needs of modern classroom teaching. Therefore, a classroom attention traction method and system based on large model dynamic content generation is proposed. SUMMARY

[0004] The purpose of this invention is to solve the problems in the prior art by proposing a classroom attention-driving method and system based on dynamic content generation of a large model.

[0005] A classroom attention-driving method based on dynamic content generation using a large model includes the following steps: S1. Monitor students' attention in class in real time and obtain real-time information about students' status; S2. Real-time acquisition of current teaching content and extraction of key knowledge points; S3. When a student's attention state is detected as "distracted", combine the student's file information, key knowledge points and prompt word templates to construct a contextualized generation instruction. S4. Input the generated instructions into the language model and output personalized attention-driving content for the student. S5. Push the personalized attention-driving content to the target student's terminal.

[0006] Preferably, in step S1, the student's real-time status information includes an anonymous ID and its real-time status tag, and the real-time status tag includes "focused" and "distracted".

[0007] Preferably, in step S2, acquiring the current teaching content in real time and extracting key knowledge points specifically includes: acquiring the teacher's lecture content in real time through speech recognition technology, and extracting key knowledge points from the teacher's lecture content through text analysis technology.

[0008] Preferably, in step S3, the student profile information includes at least one of the following: learning foundation, interests and preferences, and historical interaction records.

[0009] Preferably, in step S4, the language model is a large-scale language model that has been fine-tuned in the field of education, and the personalized attention-driving content is at least one of the following: fun puzzles, life analogies, or guiding questions related to the current knowledge point.

[0010] Preferably, in step S4, the language model is a large-scale language model that has been fine-tuned in the field of education, and the personalized attention-driving content is at least one of the following: fun puzzles, life analogies, or guiding questions related to the current knowledge point.

[0011] Preferably, it further includes: S6. Record students' response behavior to the personalized attention-grabbing content, and continuously optimize the content generation strategy based on the response behavior data.

[0012] This invention also proposes a classroom attention-guiding system based on dynamic content generation from a large model, used to implement the above method, including: Student Status Monitoring Module: This module is used to detect and determine the student's attention status in real time. It outputs the student's anonymous ID and their real-time status label. Teaching context acquisition module: used to capture the current teaching content in real time and extract key knowledge points from the teacher's lecture content within 1-3 minutes through text analysis technology; Large Model Content Generation Engine Module: This module combines student profile information, key knowledge points, and prompt word templates to construct contextualized generation instructions, generate personalized attention-driving content, and perform relevance scoring and suitability filtering on the personalized attention-driving content. Personalized push module: Used to accurately push generated personalized content to the target students' terminal devices.

[0013] Preferably, the student status monitoring module is connected to a classroom behavior sensor or a student terminal device, the teaching context acquisition module is connected to a teacher microphone and teaching courseware system, the large model content generation engine module is connected to the student status monitoring module and the teaching context acquisition module, and the personalized push module is connected to the large model content generation engine module and all student terminals.

[0014] Preferably, the large model content generation engine module has a built-in large language model that has been fine-tuned in the education field and is configured with a multi-category prompt word template library to support personalized content generation strategies.

[0015] Compared with existing technologies, the advantages of this invention are: 1. The attention-guiding method of the present invention generates personalized attention-guiding content that matches students' interests and is related to knowledge points, transforming blocking intervention into fun guidance, significantly reducing students' psychological resistance, improving the teaching experience, and the entire intervention process is completed silently through the student terminal without interrupting the teacher's lecture and other students' listening, thus maintaining the overall continuity of the classroom.

[0016] 2. This invention utilizes AI's dynamic generation capabilities to apply to real-time teaching intervention, establishing a deep semantic connection between intervention content and teaching knowledge points, helping students better understand, memorize, and apply knowledge, and achieving a leap from behavior correction to cognitive guidance. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the classroom attention-guiding system in this invention.

[0018] Figure 2 This is a flowchart of the classroom attention-guiding method in this invention. Detailed Implementation

[0019] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0020] Reference Figure 1 As shown, a classroom attention-driving system based on dynamic content generation using a large model includes: Student Status Monitoring Module: Connects to classroom behavior sensors (cameras deployed in the classroom) or student terminals (such as tablets), and uses computer vision algorithms to analyze students' facial orientation, eye state, body movements, etc. in real time, calculate attention indicators, and determine whether the status is "focused" or "distracted", while outputting anonymized student IDs and status labels.

[0021] Teaching context acquisition module: Connects to the teacher's microphone (on the lapel or on the podium) and the teaching courseware system such as the PPT and electronic whiteboard that are being played. It converts the teacher's speech into text through real-time speech recognition (ASR) and uses natural language processing technology to analyze the lecture text of the most recent 1-3 minutes to extract the core knowledge points that are currently being explained.

[0022] Large Model Content Generation Engine Module: This is the intelligent hub of the system. It receives data streams from the student status monitoring module and the teaching context acquisition module in real time and is connected to the student database (which stores student profile information). When it receives a student's "distracted" status tag, the module starts immediately. It retrieves the student's personalized information from the student profile, combines it with the current knowledge points provided by the teaching context acquisition module, selects a suitable template from the pre-set prompt word template library, fills in the specific information, constructs a highly contextualized generation instruction, generates personalized attention-driving content, and performs relevance scoring and suitability filtering on the personalized attention-driving content.

[0023] Personalized push module: Connected to the large model content generation engine module and all student terminals, it receives personalized attention-grabbing content (such as a text or graphic message) generated by the large model content generation engine module, and selects the most appropriate push method (such as pop-up window, sidebar prompt, vibration, etc.) based on the type of student terminal and the current interface to accurately send the content to the student's terminal.

[0024] A classroom attention-driving method based on dynamic content generation using a large model includes the following steps: S1. Monitor students' attention in class in real time and obtain real-time information about students' status; The student's real-time status information includes an anonymous ID and its real-time status tag, which includes "focused" and "distracted".

[0025] S2. Real-time acquisition of current teaching content and extraction of key knowledge points; Real-time acquisition of current teaching content and extraction of key knowledge points specifically includes: acquiring the teacher's lecture content in real time through speech recognition technology, and extracting key knowledge points from the teacher's lecture content through text analysis technology.

[0026] S3. When a student's attention state is detected as "distracted", combine the student's file information, key knowledge points and prompt word templates to construct a contextualized generation instruction. The student profile information includes at least one of the following: learning foundation, interests and preferences, and historical interaction records. The prompt template is such as "Create a relevant content format for a student with a learning foundation who likes [interests] but has difficulty with [knowledge points]".

[0027] S4. Input the generated instructions into the language model and output personalized attention-driving content for the student. The personalized attention-driving content is at least one of the following: fun puzzles, relatable analogies, or guiding questions related to the current knowledge point.

[0028] S5. Push the personalized attention-driving content to the target student's terminal; S6. Record students' response behavior to the personalized attention-grabbing content, and continuously optimize the content generation strategy based on the response behavior data.

[0029] Example Suppose that in a C programming class, the system detects that student A (profile: computer science student, enthusiast of online games) is distracted. At this moment, the teacher is explaining the key knowledge point of "the relationship between pointers and arrays".

[0030] After the system is triggered, the generated instruction might be: "Create a fun puzzle for a computer science student who enjoys online games but has difficulty understanding the relationship between pointers and arrays." The language model might then generate content 1 (fun puzzle): "If the game inventory is an array, then isn't a pointer a 'treasure compass' that can find any of the artifacts within it? Click to see how to use this compass."

[0031] The constructor's generation instruction is: "Create a relevant, everyday analogy for a computer science student who enjoys online games but has difficulty understanding the relationship between pointers and arrays." The language model might generate content 2 (a relatable analogy): "Imagine the pointer as a deliveryman, the array as a dormitory building, and the pointer '++' as the deliveryman going to the next dormitory. This is much faster than remembering all the dormitory numbers (array indices)! Curious?"

[0032] The content is pushed to Student A's device screen via a pop-up window or a slight vibration notification. Student A is drawn in by the game or analogy that interests them, clicking to view detailed explanations or engage in simple interactions, thus refocusing their attention on the current knowledge point without disrupting the class.

[0033] In summary, the attention-driving method of this invention transforms rigid "command interruption" into natural "fun guidance" by generating "interest hooks" that match students' interests and are related to knowledge points. This significantly reduces students' psychological resistance, enhances the teaching experience, and the entire intervention process is completed silently through the student's terminal without interrupting the teacher's lecture or other students' listening, thus maintaining the overall continuity of the classroom.

[0034] As is known from common technical knowledge, this invention can be implemented through other embodiments that do not depart from its spirit or essential characteristics. Therefore, the disclosed embodiments described above are merely illustrative in all respects and are not the only ones. All modifications within the scope of this invention or its equivalents are included in this invention.

Claims

1. A classroom attention-driving method based on dynamic content generation from a large model, characterized by: Includes the following steps: S1. Monitor students' attention in class in real time and obtain real-time information about students' status; S2. Real-time acquisition of current teaching content and extraction of key knowledge points; S3. When a student's attention state is detected as "distracted", combine the student's file information, key knowledge points and prompt word templates to construct a contextualized generation instruction. S4. Input the generated instructions into the language model and output personalized attention-driving content for the student. S5. Push the personalized attention-driving content to the target student's terminal.

2. The classroom attention-driving method based on large-scale dynamic content generation according to claim 1, characterized in that: In step S1, the student's real-time status information includes an anonymous ID and its real-time status label, which includes "focused" and "distracted".

3. The classroom attention-driving method based on dynamic content generation of a large model according to claim 1, characterized in that: In step S2, acquiring the current teaching content in real time and extracting key knowledge points specifically includes: acquiring the teacher's lecture content in real time through speech recognition technology, and extracting key knowledge points from the teacher's lecture content through text analysis technology.

4. The classroom attention-driving method based on dynamic content generation of a large model according to claim 1, characterized in that: In step S3, the student profile information includes at least one of the following: learning foundation, interests and preferences, and historical interaction records.

5. The classroom attention-driving method based on large-scale dynamic content generation according to claim 1, characterized in that: In step S4, the language model is a large-scale language model that has been fine-tuned in the field of education, and the personalized attention-driving content is at least one of the following: fun puzzles, life analogies, or guiding questions related to the current knowledge point.

6. The classroom attention-driving method based on dynamic content generation of a large model according to claim 1, characterized in that: Also includes: S6. Record students' response behavior to the personalized attention-grabbing content, and continuously optimize the content generation strategy based on the response behavior data.

7. A classroom attention-guiding system based on dynamic content generation from a large model, used to implement the method as described in any one of claims 1-6, characterized in that: include: Student Status Monitoring Module: This module is used to detect and determine the student's attention status in real time. It outputs the student's anonymous ID and their real-time status label. Teaching context acquisition module: used to capture the current teaching content in real time and extract key knowledge points from the teacher's lecture content within 1-3 minutes through text analysis technology; Large Model Content Generation Engine Module: This module combines student profile information, key knowledge points, and prompt word templates to construct contextualized generation instructions, generate personalized attention-driving content, and perform relevance scoring and suitability filtering on the personalized attention-driving content. Personalized push module: Used to accurately push generated personalized content to the target students' terminal devices.

8. A classroom attention-driving system based on large-scale dynamic content generation according to claim 7, characterized in that: The student status monitoring module is connected to classroom behavior sensors or student terminal devices; the teaching context acquisition module is connected to the teacher's microphone and teaching courseware system; the large model content generation engine module is connected to the student status monitoring module and the teaching context acquisition module; and the personalized push module is connected to the large model content generation engine module and all student terminals.

9. A classroom attention-guiding system based on dynamic content generation of a large model according to claim 7, characterized in that: The large-scale content generation engine module has a built-in large-scale language model that has been fine-tuned for the education field, and is equipped with a multi-category prompt word template library to support personalized content generation strategies.

Images