A multimodal limited-time heuristic homework tutoring method and system

Abstract

The application discloses a multimodal limited-time heuristic homework tutoring method and system. The method comprises the following steps: receiving a tutoring request through multiple modes such as passive identification (behavior confusion), active instruction (voice / key), and contact triggering (finger stay); obtaining the remaining available times of the day, which are calculated from the preset daily threshold and the used times of the day; if the remaining times are greater than 0, executing a complete tutoring mode to generate hierarchical tutoring content including knowledge point prompts, problem solving ideas, step-by-step explanations, and similar example problems; if the remaining times are equal to 0, executing a simplified tutoring mode to only generate knowledge point prompts and problem solving idea frameworks; and recording the tutoring behavior and pushing a summary to a specified terminal. The application can help children solve problems while effectively avoiding over-reliance, cultivating independent thinking ability, and can be linked with the concentration supervision system submitted by the applicant on the same day to build a complete learning behavior closed loop.

Description

Technical Field

[0001] This invention relates to the field of intelligent education technology, and more specifically, to a multimodal limited-time heuristic homework tutoring method and system applied to intelligent learning devices. Background Technology

[0002] Currently, students often need parental guidance when encountering difficulties in their homework. However, parents may be overwhelmed due to busy work schedules, forgotten knowledge, or inappropriate explanation methods. While existing homework help software can provide answers, it often directly provides the results, which can easily lead to student dependency and a loss of independent thinking skills. Existing patent applications have proposed heuristic tutoring methods, but they lack effective management of the frequency of tutoring and fail to address the problem of students' over-reliance on tutoring. Therefore, there is an urgent need for an intelligent homework tutoring method and system that can help students solve difficult problems, cultivate independent thinking habits, and be flexibly controlled by parents or teachers. Summary of the Invention

[0003] The technical problem to be solved by this invention is to provide a multimodal limited-time heuristic homework tutoring method and system, which addresses the shortcomings of existing tutoring tools that easily lead to student dependence and lack a frequency management mechanism.

[0004] The technical solution adopted by this invention to solve its technical problem is: to construct a multimodal limited-time heuristic homework tutoring method, including the following steps:

[0005] S1: Multimodal trigger tutoring request, receiving the user's tutoring request through one or more of the following methods:

[0006] Passive triggering: Real-time acquisition of user images via camera, and use of behavior recognition model to identify whether the user has preset confused behavior characteristics, including at least one of scratching head, frowning, staring at the same area for more than a preset time threshold;

[0007] Active trigger: Receives help signals explicitly issued by the user via physical buttons or voice commands;

[0008] Contact triggering: The device detects the user's finger movement when it stays in the question area for more than a preset time threshold using a touch sensor or camera.

[0009] S2: Get the number of tutoring sessions available for the day, query the preset daily tutoring session threshold and the number of sessions already used for the day, and calculate the remaining number of sessions available for the day;

[0010] S3: Count determination and branching: If the remaining available counts for the day are greater than 0, then execute the full tutoring mode of step S4; if the remaining available counts for the day are equal to 0, then execute the simplified tutoring mode of step S5.

[0011] S4: Complete Tutoring Mode, which calls upon a large AI model to generate complete tutoring content containing multiple levels, and provides it to the user through speech synthesis and / or display; the multiple levels include at least three of the following: knowledge point prompts, problem-solving frameworks, step-by-step explanations, and similar examples; after execution, the number of times used that day is incremented by 1;

[0012] S5: Simplified tutoring mode, which calls the AI ​​big model to generate simplified tutoring content. The simplified tutoring content only includes knowledge point prompts and problem-solving frameworks, without complete solutions or step-by-step explanations, and is provided to users through voice synthesis and / or display.

[0013] S6: Recording and Feedback. Record the content of this tutoring request, the mode executed, and the key information generated into the user's learning profile, and generate a tutoring summary to be pushed to a preset designated terminal on a regular basis.

[0014] Furthermore, the daily tutoring frequency threshold is set by parents through a parental terminal (such as a WeChat mini-program or mobile app), with a range of 1-10 times and a default value of 3 times. Parents can adjust this flexibly according to their child's grade level and self-discipline in learning.

[0015] Furthermore, the "preset time threshold" in the passive trigger can be set according to the actual application scenario, preferably 10-15 seconds. For example, when a user stares at the same area for more than 10 seconds and is accompanied by other confused behavioral characteristics, the system can determine that it is a potential intention to seek help.

[0016] Furthermore, the AI ​​large-scale model can employ commercially available or open-source large-scale language models known in the field, ensuring the generated content is both inspiring and age-appropriate through carefully designed prompts. For example, in the complete tutoring mode, the prompts require the model to "explain this math problem in three steps using language that a third-grade student can understand, and finally provide a similar practice problem"; in the simplified tutoring mode, the prompts require the model to "only prompt the knowledge points and problem-solving strategies involved in this problem, without providing the specific answer."

[0017] Furthermore, this invention also includes a linkage mechanism with the "A Method and System for Monitoring Work Concentration Based on Random Triggering and Multimodal Verification" submitted by the applicant on the same day. The system receives concentration data generated by the monitoring method. If there are too many distractions on a given day, the threshold for the number of tutoring sessions the following day is automatically lowered to encourage the user to improve concentration. Conversely, if the concentration remains excellent, the number of tutoring sessions can be appropriately increased as a reward.

[0018] Compared with the prior art, the present invention has the following beneficial effects:

[0019] 1. Cultivate independent thinking ability: By using a daily limit mechanism, only hints are provided after the number of attempts has been used up, forcing students to think independently and effectively avoiding over-reliance on tutoring.

[0020] 2. Flexible and controllable for parents or teachers: Parents or teachers can set the number of tutoring sessions per day according to the child's actual situation and receive daily tutoring reports, enabling remote companionship and guidance.

[0021] 3. Multimodal triggering, intelligent and seamless: Tutoring is triggered through multiple methods such as behavior recognition, voice commands, and finger hovering, which is more in line with students' usage habits and improves the user experience.

[0022] 4. Heuristic teaching: The complete tutoring model adopts a tiered approach, from knowledge point prompts to thought frameworks and step-by-step explanations, which conforms to cognitive laws and helps students truly understand the questions.

[0023] 5. Collaborate with the monitoring system: Link with monitoring data to build an incentive mechanism where "focus affects the number of tutoring sessions", forming a closed loop of learning behavior.

[0024] 6. Dependency-resistant design: Only the idea is given after the number of attempts is exhausted, which is fundamentally different from the "never give an answer" or "give an answer an unlimited number of times" in the market, and has unique technological innovation and market competitiveness. Attached Figure Description

[0025] Figure 1 This is a flowchart of the multimodal limited-time heuristic homework tutoring method of the present invention.

[0026] Figure 2 This is a schematic diagram illustrating the interaction between the system of the present invention and a designated terminal.

[0027] Figure 3 This is a schematic diagram of the linkage mechanism between the present invention and the monitoring system. Detailed Implementation

[0028] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0029] Example 1: Basic Tutoring Process

[0030] Reference Figure 1 As shown in this embodiment, parents set the daily tutoring frequency threshold to 3 times via WeChat mini-program. One evening at 7 PM, the child encounters a math word problem and says to the desk lamp, "Xiao Le, Xiao Le, how do I do this problem?" (Actively triggered). The desk lamp's camera captures an image of the problem, and the system initiates the tutoring process.

[0031] The system query shows that 2 attempts have been used today, and 1 attempt remains (greater than 0). Therefore, the full tutoring mode is executed. The system calls the AI ​​large model, generates the following content, and broadcasts it via voice:

[0032] "This question tests your understanding of 'sum and multiple' problems. First, identify the 'sum' and 'multiple' in the problem. Second, draw a line segment diagram, treating the smaller number as one part. Third, divide the sum by the number of parts to find out how much is one part. For example, in this problem, Xiaoming and Xiaohong have a total of 30 books. Xiaoming has twice as many books as Xiaohong. Therefore, Xiaohong has 10 books, and Xiaoming has 20 books. Did you understand? Try doing the exercises next to you."

[0033] After the broadcast is finished, the system will update the number of times used that day to 3 and record the content of this tutoring session.

[0034] Ten minutes later, the child asked the question again. The system checked and found that there were 0 tutoring attempts remaining for the day, so it switched to simplified tutoring mode. The system announced: "Today's tutoring attempts have been used up. This question is still a 'sum and multiple problem.' Think about how you analyzed the previous question and try to draw a line segment diagram yourself." The child had to think for himself, thus cultivating his independent problem-solving ability.

[0035] Example 2: Linkage with the monitoring system

[0036] In this embodiment, the system is integrated into the same desk lamp with a homework focus monitoring system based on random triggering and multimodal verification. When the monitoring system detects that the child's attention has wandered six times in a day (exceeding a threshold), it sends the focus data to this system. Based on preset rules, this system automatically adjusts the tutoring frequency threshold for the next day from three to two times, and pushes a notification to the parent's terminal (e.g., a WeChat mini-program) at the beginning of the next day: "Yesterday's focus was insufficient; today's tutoring frequency has been automatically adjusted to two times. Please encourage your child to concentrate on their studies." When the child asks questions the next day, the system executes according to the new threshold. Conversely, if the child has excellent focus for three consecutive days, the system automatically increases the tutoring frequency threshold to four times as a reward.

[0037] Example 3: Parental Remote Control

[0038] A parent, traveling on business, checked their child's daily tutoring report through a parent terminal (WeChat mini-program) and noticed the child frequently requesting tutoring on Chinese composition. The parent felt the child needed more writing practice, so they increased the daily tutoring frequency from 3 to 5 times, adding a note saying, "Encourage him to think for himself first." The system received the setting and it took effect immediately.

[0039] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A multimodal, limited-repetition heuristic homework tutoring method, applied to intelligent learning devices, characterized in that, Includes the following steps: S1: Multimodal trigger tutoring request, receiving the user's tutoring request through one or more of the following methods: Passive triggering: Real-time acquisition of user images via camera, and use of behavior recognition model to identify whether the user has preset confused behavior characteristics, including at least one of scratching head, frowning, staring at the same area for more than a preset time threshold; Active trigger: Receives help signals explicitly issued by the user via physical buttons or voice commands; Contact triggering: The device detects the user's finger movement when it stays in the question area for more than a preset time threshold using a touch sensor or camera. S2: Get the number of tutoring sessions available for the day, query the preset daily tutoring session threshold and the number of sessions already used for the day, and calculate the remaining number of sessions available for the day; S3: Count determination and branching. If the remaining available attempts for the day are greater than 0, then proceed to step S4; If the remaining available attempts for the day are 0, then proceed to step S5; S4: Complete Tutoring Mode, which calls upon a large AI model to generate complete tutoring content containing multiple levels, and provides it to the user through speech synthesis and / or display; the multiple levels include at least three of the following: knowledge point prompts, problem-solving frameworks, step-by-step explanations, and similar examples; after execution, the number of times used that day is incremented by 1; S5: Simplified tutoring mode, which calls the AI ​​big model to generate simplified tutoring content. The simplified tutoring content only includes knowledge point prompts and problem-solving frameworks, without complete solutions or step-by-step explanations, and is provided to users through voice synthesis and / or display. S6: Recording and Feedback. Record the content of this tutoring request, the mode executed, and the key information generated into the user's learning profile, and generate a tutoring summary to be pushed to a preset designated terminal on a regular basis.

2. The multimodal limited-time heuristic homework tutoring method according to claim 1, characterized in that, In step S2, the daily tutoring frequency threshold is set by the parents through the parent terminal, and its value ranges from 1 to 10 times.

3. The multimodal limited-time heuristic homework tutoring method according to claim 1, characterized in that, In step S4, the complete tutoring content generated by the AI ​​model also includes at least one variation exercise related to the knowledge points of this question, for users to consolidate their practice.

4. The multimodal limited-time heuristic homework tutoring method according to claim 1, characterized in that, In step S5, the simplified tutoring mode further includes: based on the user's historical learning archive, pushing past tutoring records related to the knowledge points of the current question to guide the user to review them independently.

5. The multimodal limited-time heuristic homework tutoring method according to claim 1, characterized in that, Before step S4, the method further includes: outputting a prompt message, asking the user whether to enable the full tutoring mode, and executing step S4 after receiving the user's confirmation instruction.

6. The multimodal limited-time heuristic homework tutoring method according to claim 1, characterized in that, It also includes step S7: automatically resetting the number of uses for the day to zero at a preset time each day, and generating a tutoring summary report for the previous day and pushing it to the designated terminal.

7. The multimodal limited-time heuristic homework tutoring method according to claim 1, characterized in that, It also includes step S8: receiving focus data generated by another patent application filed on the same day as this application, entitled "A Method and System for Monitoring Work Focus Based on Random Triggering and Multimodal Verification", and dynamically adjusting the daily tutoring frequency threshold according to the focus data.

8. A multimodal intelligent learning tutoring system implementing the method of any one of claims 1-6, characterized in that, include: A multimodal input module, including a camera, a microphone, and a touch sensor, is used to implement the multiple triggering methods described in claim 1; The storage module is used to store the preset daily tutoring frequency threshold, the number of times the tutoring has been used that day, and the user's learning profile. The processing module, connected to the input module and the storage module, is used to determine the number of executions and call the AI ​​model to generate tutoring content. Output module, including speakers and / or a display screen, for outputting the generated tutoring content to the user; The communication module is used to interact with a designated terminal, receive thresholds set by the terminal, and push reports to the terminal.

9. The multimodal intelligent learning tutoring system according to claim 7, characterized in that, When the processing module calls the large AI model, it uses prompt word engineering technology to explicitly require the hierarchical structure of the generated content and the constraint of "not providing a complete answer" in the prompt words.

10. The multimodal intelligent learning tutoring system according to claim 7, characterized in that, The system is integrated into a smart study lamp and shares the same hardware platform with a task focus monitoring system based on random triggering and multimodal verification.

11. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the multimodal limited-time heuristic job tutoring method as described in any one of claims 1 to 6.

12. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the multimodal limited-time heuristic job tutoring method as described in any one of claims 1 to 6.

Images