An AI-based education system

Abstract

This invention relates to the field of educational technology, specifically to an artificial intelligence-based educational system. The system includes: identifying knowledge points for a student's course; randomly selecting two knowledge points as knowledge point pairs until all knowledge points are paired; identifying the student's historical grades to determine their mastery of each knowledge point; obtaining the student's accuracy rate on questions related to each knowledge point pair; performing a correlation strength analysis on the knowledge point pairs and calculating a recommendation weight for each pair by incorporating the mastery level; sorting the knowledge point pairs according to their recommendation weights and pushing all knowledge point pairs in order, selecting corresponding questions as content when pushing each pair; this invention gradually forms a network-like knowledge graph during the question training process, i.e., connecting different knowledge points separately; by pairing all knowledge points, it avoids knowledge omissions and ensures that each knowledge point is learned.

Description

Technical Field

[0001] This invention relates to the field of educational technology, and more specifically to an artificial intelligence-based educational system. Background Technology

[0002] Just-in-Time Teaching (JITT) integrates multiple mechanisms and features to promote effective interaction between teachers and students. First, this model employs real-time interactive teaching, emphasizing timely communication and feedback between teachers and students, making the teaching process more flexible and adaptable. Second, through the use of mind maps, curriculum design, and question setting, the JITT model constructs a clear knowledge framework in teaching, promoting knowledge co-construction between teachers and students and enhancing students' understanding and application of the learned content.

[0003] Practice has shown that timely instruction is essentially still a "diagnosis-remediation" model. The system diagnoses a student's knowledge weakness (e.g., not knowing point A), then pushes a large number of exercises related to point A, without connecting it with other weaknesses, lacking proactive knowledge construction and profound cognitive experience. Furthermore, existing systems focus on mastering the knowledge points themselves, but rarely pay attention to how these knowledge points are organized and connected in the student's mind. Modern education curricula are systematic, with knowledge points within the same course forming a network, interwoven with each other. That is, one knowledge point is interconnected with several or even dozens of other knowledge points. However, current mind mapping technology relies on a deeper understanding of knowledge points (knowledge points are related upstream and downstream), making it difficult to recommend knowledge points from different lines. The degree of interweaving is insufficient, making it difficult for students to form holistic thinking. Therefore, an AI-based educational system is needed that can change the existing knowledge point recommendation method, enabling all knowledge points to form an interwoven network. Summary of the Invention

[0004] The technical problem to be solved by this invention is to provide an artificial intelligence-based education system that can change the existing knowledge point recommendation method so that all knowledge points are intertwined.

[0005] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:

[0006] An artificial intelligence-based education system includes:

[0007] The course identification module identifies knowledge points for a student's course, obtaining n knowledge points (n > 2). It then randomly selects two knowledge points, i and j, as a knowledge point pair (i, j) until all knowledge points form knowledge point pairs. Finally, it identifies the structural correlation between knowledge points i and j. ;

[0008] The grades module identifies students' historical grades and their level of mastery of each knowledge point. , =β× +γ× +(1-β-γ)× ;

[0009] Where β and γ are weighting coefficients, The student's accuracy rate on the nth knowledge point; This represents the number of times a student attempts a question on the nth knowledge point. This refers to the time allotted for the student to complete a problem on the nth knowledge point. This represents the total number of times the course questions were answered. The total time a student spends on completing exercises in this course;

[0010] Simultaneously obtain the student's accuracy rate on questions that simultaneously involve knowledge point i and knowledge point j. ;

[0011] The analysis and recommendation module performs correlation analysis on knowledge point pairs (i,j). Intensity analysis, while incorporating the level of mastery Calculation obtained , ;according to , , The result yields the recommendation weight for each knowledge point pair (i,j). ;

[0012] =α× +(1-α)× ; = ×(1- ); where α is the weighting coefficient;

[0013] According to the recommendation weight The size of the knowledge point pairs (i,j) is sorted, and all knowledge point pairs (i,j) are pushed in order. When pushing knowledge point pairs (i,j), the questions containing the corresponding knowledge points i and j are selected as the content to be presented.

[0014] Preferably, three knowledge points i, j, and k are randomly selected as a knowledge point group (i,j,k) until all knowledge points form a knowledge point group; the structural correlation between knowledge points i, j, and k is then identified. ;

[0015] The performance module simultaneously retrieves the student's accuracy rate on questions that involve knowledge points i, j, and k. ;

[0016] The analysis and recommendation module performs correlation analysis on knowledge point groups (i,j,k). Intensity analysis, while incorporating the level of mastery Calculation obtained , , ;according to , , , The result yields the recommendation weight for each knowledge point group (i,j,k). ;

[0017] =α× +(1-α)× ; = ×(1- );

[0018] According to the recommendation weight The size of the knowledge point group (i,j,k) is used to sort the knowledge point groups (i,j,k) in order, and all knowledge point groups (i,j,k) are pushed out in order. When pushing knowledge point groups (i,j,k), the questions containing the corresponding knowledge points i, j, and k are selected as the content to be presented.

[0019] Preferably, when n>10, select 2-10 knowledge points to form a knowledge point group.

[0020] Preferably, the structural correlation between knowledge point i and knowledge point j is identified. The standardization is carried out by the teacher.

[0021] Preferably, the structural correlation between knowledge point i, knowledge point j, and knowledge point k is identified. Further includes:

[0022] Obtain the structural correlation between knowledge point i and knowledge point j. Obtain the structural correlation between knowledge point i and knowledge point k. Obtain the structural correlation between knowledge point j and knowledge point k. ;

[0023] = .

[0024] Preferably, the education system further includes a teacher module;

[0025] The teacher module obtains the recommendation weights of all students' sorted knowledge point pairs (i,j). And assign recommendation weights to the same knowledge points for different students on pairs (i,j). The weights are summed to obtain the teaching recommendation weights and then sorted. Teaching is then carried out on the knowledge points (i,j) corresponding to the sorted teaching recommendation weights.

[0026] Preferably, after students have received instruction and then completed the exercises based on the pushed questions, the grade module is updated, and the analysis and recommendation module re-pushes the information based on the updated grade module.

[0027] Preferably, the analysis and recommendation module re-obtains the recommendation weights of all students' sorted knowledge point pairs (i,j) based on the updated grade module. And assign recommendation weights to the same knowledge points for different students on pairs (i,j). The teaching recommendation weights are summed and then ranked to verify the teaching outcomes. The recommendation weights for the same knowledge point pair (i,j) are calculated for different students. If the sum of the values ​​results in a decrease in the teaching recommendation weight, it proves that the teaching is effective; if the value of the teaching recommendation weight increases, the teaching should be adjusted.

[0028] Preferably, after students complete the questions pushed to them, the grade module is updated, and the analysis and recommendation module re-pushes questions based on the updated grade module.

[0029] The beneficial effects of this invention are: by following the recommended weights The system prioritizes and recommends knowledge point pairs (i,j) based on their size, selecting two highly related knowledge points as interventions. These pairs are often areas where students have insufficient mastery of the knowledge points. Conversely, knowledge points with higher mastery are less recommended based on their weight. By recommending questions containing knowledge point pairs (i,j), students must simultaneously apply multiple knowledge points to solve problems. Furthermore, all knowledge points within the same course are grouped together, gradually forming a network of knowledge graphs during practice. This connects different knowledge points, fundamentally addressing the "island effect" and "fragmented learning paths" of knowledge points, and enhancing students' holistic thinking. Grouping all knowledge points together prevents omissions, ensuring that every knowledge point is learned and providing some degree of anti-forgetting. Starting with knowledge point pairs where mastery is insufficient strengthens less-known knowledge points, aligning with the original intention of JiTT teaching and representing an improvement upon it while retaining the personalized recommendations for each student. Attached Figure Description

[0030] Figure 1 This is a schematic diagram of an artificial intelligence-based education system according to a specific embodiment of the present invention. Detailed Implementation

[0031] To explain the technical content, objectives, and effects of the present invention in detail, the following description is provided in conjunction with the embodiments and accompanying drawings.

[0032] Please refer to Figure 1 An artificial intelligence-based education system includes:

[0033] The course identification module identifies knowledge points for a student's course, obtaining n knowledge points (n > 2). It then randomly selects two knowledge points, i and j, as a knowledge point pair (i, j) until all knowledge points form knowledge point pairs. Finally, it identifies the structural correlation between knowledge points i and j. ;

[0034] The grades module identifies students' historical grades and their level of mastery of each knowledge point. , =β× +γ× +(1-β-γ)× ;

[0035] Where β and γ are weighting coefficients, The student's accuracy rate on the nth knowledge point; This represents the number of times a student attempts a question on the nth knowledge point. This refers to the time allotted for the student to complete a problem on the nth knowledge point. This represents the total number of times the course questions were answered. The total time a student spends on completing exercises in this course;

[0036] Simultaneously obtain the student's accuracy rate on questions that simultaneously involve knowledge point i and knowledge point j. ;

[0037] The analysis and recommendation module performs correlation analysis on knowledge point pairs (i,j). Intensity analysis, while incorporating the level of mastery Calculation obtained , ;according to , , The result yields the recommendation weight for each knowledge point pair (i,j). ;

[0038] =α× +(1-α)× ; = ×(1- ); where α is the weighting coefficient;

[0039] According to the recommendation weight The size of the knowledge point pairs (i,j) is sorted, and all knowledge point pairs (i,j) are pushed in order. When pushing knowledge point pairs (i,j), the questions containing the corresponding knowledge points i and j are selected as the content to be presented.

[0040] As can be seen from the above description, by following the recommendation weight... The system prioritizes and recommends knowledge point pairs (i,j) based on their size, selecting two highly related knowledge points as interventions. These pairs are often areas where students have insufficient mastery of the knowledge points. Conversely, knowledge points with higher mastery are less recommended based on their weight. By recommending questions containing knowledge point pairs (i,j), students must simultaneously apply multiple knowledge points to solve problems. Furthermore, all knowledge points within the same course are grouped together, gradually forming a network of knowledge graphs during practice. This connects different knowledge points, fundamentally addressing the "island effect" and "fragmented learning paths" of knowledge points, and enhancing students' holistic thinking. Grouping all knowledge points together prevents omissions, ensuring that every knowledge point is learned and providing some degree of anti-forgetting. Starting with knowledge point pairs where mastery is insufficient strengthens less-known knowledge points, aligning with the original intention of JiTT teaching and representing an improvement upon it while retaining the personalized recommendations for each student.

[0041] Furthermore, when three knowledge points i, j, and k are randomly selected as a knowledge point group (i,j,k) until all knowledge points form a knowledge point group, the structural correlation between knowledge points i, j, and k is identified. ;

[0042] The performance module simultaneously retrieves the student's accuracy rate on questions that involve knowledge points i, j, and k. ;

[0043] The analysis and recommendation module performs correlation analysis on knowledge point groups (i,j,k). Intensity analysis, while incorporating the level of mastery Calculation obtained , , ;according to , , , The result yields the recommendation weight for each knowledge point group (i,j,k). ;

[0044] =α× +(1-α)× ; = ×(1- );

[0045] According to the recommendation weight The size of the knowledge point group (i,j,k) is used to sort the knowledge point groups (i,j,k) in order, and all knowledge point groups (i,j,k) are pushed out in order. When pushing knowledge point groups (i,j,k), the questions containing the corresponding knowledge points i, j, and k are selected as the content to be presented.

[0046] As can be seen from the above description, forming knowledge point groups can make the questions generated by the knowledge point groups more complex, generally more difficult, more challenging for students, and can quickly filter out students with stronger abilities.

[0047] Furthermore, when n > 10, select 2 to 10 knowledge points to form a knowledge point group.

[0048] Furthermore, identify the structural correlation between knowledge point i and knowledge point j. The standardization is carried out by the teacher.

[0049] As can be seen from the above description, because teachers have been engaged in teaching a single course for a long time, they have a deeper understanding of the relationship between the knowledge points contained in the course, and can improve the accuracy by directly labeling.

[0050] Furthermore, identify the structural correlation between knowledge point i, knowledge point j, and knowledge point k. Further includes:

[0051] Obtain the structural correlation between knowledge point i and knowledge point j. Obtain the structural correlation between knowledge point i and knowledge point k. Obtain the structural correlation between knowledge point j and knowledge point k. ;

[0052] = .

[0053] Furthermore, the education system also includes a teacher module;

[0054] The teacher module obtains the recommendation weights of all students' sorted knowledge point pairs (i,j). And assign recommendation weights to the same knowledge points for different students on pairs (i,j). The weights are summed to obtain the teaching recommendation weights and then sorted. Teaching is then carried out on the knowledge points (i,j) corresponding to the sorted teaching recommendation weights.

[0055] As described above, the recommendation weights for the same knowledge point pairs (i,j) from different students are used to determine the optimal pairing. By summing the results to obtain teaching recommendation weights and sorting them, we can obtain the overall mastery level of knowledge points in the class, select the knowledge points with the weakest mastery level, and intervene in teaching to further improve students' network thinking.

[0056] Furthermore, after students receive instruction and complete the questions provided, the grade module is updated, and the analysis and recommendation module re-submits information based on the updated grade module.

[0057] Furthermore, the analysis and recommendation module re-obtains the recommendation weights for all students' sorted knowledge point pairs (i,j) based on the updated grade module. And assign recommendation weights to the same knowledge points for different students on pairs (i,j). The teaching recommendation weights are summed and then ranked to verify the teaching outcomes. The recommendation weights for the same knowledge point pair (i,j) are calculated for different students. If the sum of the values ​​results in a decrease in the teaching recommendation weight, it proves that the teaching is effective; if the value of the teaching recommendation weight increases, the teaching should be adjusted.

[0058] As can be seen from the above description, by re-verifying teaching outcomes, the teaching effectiveness can be promptly fed back to teachers, thereby enhancing their enthusiasm.

[0059] Furthermore, after students complete the questions pushed to them, the grade module is updated, and the analysis and recommendation module re-pushes questions based on the updated grade module.

[0060] As can be seen from the above description, by updating and re-recommending questions, we can always find students' weaknesses and ensure teaching effectiveness.

[0061] Example 1

[0062] An artificial intelligence-based education system includes:

[0063] For a student's course, identify the knowledge points and obtain n knowledge points, where n > 2; randomly select two knowledge points i and j as knowledge point pairs (i, j) until all knowledge points form knowledge point pairs; identify the structural correlation between knowledge points i and j. Identify the structural correlation between knowledge point i and knowledge point j. The standardization is carried out by the teacher.

[0064] Simultaneously obtain the student's accuracy rate on questions that simultaneously involve knowledge point i and knowledge point j. ;

[0065] Assess the correlation between knowledge point pairs (i,j) Strength analysis, =α× +(1-α)× Where α is the weighting coefficient; Quantify the strength of the connections between knowledge points;

[0066] Identify students' historical performance to determine their level of mastery of each knowledge point. , =β× +γ× +(1-β-γ)× ;

[0067] Where β and γ are weighting coefficients (β+γ≤1). The student's accuracy rate on the nth knowledge point; This represents the number of times a student attempts a question on the nth knowledge point. This refers to the time allotted for the student to complete a problem on the nth knowledge point. This represents the total number of times the course questions were answered. The total time a student spends on completing exercises in this course;

[0068] To obtain the degree of mastery of knowledge point i ;

[0069] =β× +γ× +(1-β-γ)× ;

[0070] To obtain the degree of mastery of knowledge point j ;

[0071] =β× +γ× +(1-β-γ)× ;

[0072] according to , , The result yields the recommendation weight for each knowledge point pair (i,j). ;

[0073] = ×(1- );

[0074] According to the recommendation weight The size of the knowledge point pairs (i,j) is sorted, and all knowledge point pairs (i,j) are pushed in order. When pushing knowledge point pairs (i,j), the questions containing the corresponding knowledge points i and j are selected as the content to be presented.

[0075] After students complete the questions pushed to them, the grade module is updated, and the analysis and recommendation module pushes new questions based on the updated grade module.

[0076] Example 2

[0077] An artificial intelligence-based education system, which shares similarities with Embodiment 1 (and will not be repeated here), further includes:

[0078] Obtain the recommendation weights of all student knowledge point pairs (i,j) after sorting. And assign recommendation weights to the same knowledge points for different students on pairs (i,j). The weights are summed to obtain the teaching recommendation weights and then sorted. Teaching is then carried out on the knowledge points (i,j) corresponding to the sorted teaching recommendation weights.

[0079] After students receive instruction and complete the exercises provided, their scores are updated, and new exercises are then sent based on these updated scores.

[0080] Based on the updated scores, re-obtain the recommendation weights for all students' sorted knowledge point pairs (i,j). And assign recommendation weights to the same knowledge points for different students on pairs (i,j). The teaching recommendation weights are summed and then ranked to verify the teaching outcomes. The recommendation weights for the same knowledge point pair (i,j) are calculated for different students. If the sum of the values ​​results in a decrease in the teaching recommendation weight, it proves that the teaching is effective; if the value of the teaching recommendation weight increases, the teaching should be adjusted.

[0081] Example 3

[0082] An artificial intelligence-based education system, similar to Embodiment 1, will not be repeated here, except that:

[0083] For a student's course, identify knowledge points, obtaining n knowledge points (n > 3). Randomly select three knowledge points i, j, and k as a knowledge point group (i, j, k) until all knowledge points form a knowledge point group. Identify the structural correlation between knowledge points i, j, and k. ;

[0084] = .

[0085] Simultaneously, obtain the student's accuracy rate on questions that simultaneously involve knowledge points i, j, and k. ;

[0086] The analysis and recommendation module performs correlation analysis on knowledge point groups (i,j,k). Intensity analysis, while incorporating the level of mastery Calculation obtained , , ;according to , , , The result yields the recommendation weight for each knowledge point group (i,j,k). ;

[0087] =α× +(1-α)× ; = ×(1- );

[0088] According to the recommendation weight The size of the knowledge point group (i,j,k) is used to sort the knowledge point groups (i,j,k) in order, and all knowledge point groups (i,j,k) are pushed out in order. When pushing knowledge point groups (i,j,k), the questions containing the corresponding knowledge points i, j, and k are selected as the content to be presented.

[0089] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent modifications made based on the content of the present invention specification and drawings, or direct or indirect applications in related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. An artificial intelligence-based education system, characterized in that, include: The course identification module identifies knowledge points for a student's course, obtaining n knowledge points where n > 2. Randomly select two knowledge points i and j as knowledge point pairs (i,j) until all knowledge points form knowledge point pairs; identify the structural correlation between knowledge points i and j. Certification is conducted by teachers; The grades module identifies students' historical grades and their level of mastery of each knowledge point. , =β× +γ× +(1-β-γ)× ; Where β and γ are weighting coefficients, The student's accuracy rate on the nth knowledge point; This represents the number of times a student attempts a question on the nth knowledge point. This refers to the time allotted for the student to complete a problem on the nth knowledge point. This represents the total number of times the course questions were answered. The total time a student spends on completing the exercises in this course; Simultaneously obtain the student's accuracy rate on questions that simultaneously involve knowledge point i and knowledge point j. ; The analysis and recommendation module performs correlation analysis on knowledge point pairs (i,j). Intensity analysis, while incorporating the level of mastery Calculation obtained , ;according to , , The result yields the recommendation weight for each knowledge point pair (i,j). ; =α× +(1-α)× ; = ×(1- ); where α is the weighting coefficient; According to the recommendation weight The size of the knowledge point pairs (i,j) is sorted, and all knowledge point pairs (i,j) are pushed in order. When pushing knowledge point pairs (i,j), the questions containing the corresponding knowledge point i and knowledge point j are selected as the content to be presented. When three knowledge points i, j, and k are randomly selected as a knowledge point group (i, j, k) until all knowledge points form a knowledge point group, the structural correlation between knowledge points i, j, and k is identified. ; The performance module simultaneously retrieves the student's accuracy rate on questions that involve knowledge points i, j, and k. ; The analysis and recommendation module performs correlation analysis on knowledge point groups (i,j,k). Intensity analysis, while incorporating the level of mastery Calculation obtained , , ;according to , , , The result yields the recommendation weight for each knowledge point group (i,j,k). ; =α× +（1-α）× ； = ×（1- ）； According to the recommendation weight The size of the knowledge point group (i,j,k) is used to sort the knowledge point groups (i,j,k) in order, and all knowledge point groups (i,j,k) are pushed out in order. When pushing knowledge point groups (i,j,k), the questions containing the corresponding knowledge points i, j, and k are selected as the content to be presented.

2. The artificial intelligence-based education system according to claim 1, characterized in that, When n > 10, select 2 to 10 knowledge points to form a knowledge point group.

3. The artificial intelligence-based education system according to claim 1, characterized in that, Identify the structural correlation between knowledge point i, knowledge point j, and knowledge point k. Further includes: Obtain the structural correlation between knowledge point i and knowledge point j. Obtain the structural correlation between knowledge point i and knowledge point k. Obtain the structural correlation between knowledge point j and knowledge point k. ; = 。 4. The artificial intelligence-based education system according to claim 1, characterized in that, The education system also includes a teacher module; The teacher module obtains the recommendation weights of all students' sorted knowledge point pairs (i,j). And assign recommendation weights to the same knowledge points for different students on pairs (i,j). The weights are summed to obtain the teaching recommendation weights and then sorted. Teaching is then carried out on the knowledge points (i,j) corresponding to the sorted teaching recommendation weights.

5. The artificial intelligence-based education system according to claim 4, characterized in that, After students receive instruction and complete the questions provided, the grade module updates, and the analysis and recommendation module re-submits recommendations based on the updated grade module.

6. The artificial intelligence-based education system according to claim 5, characterized in that, The analysis and recommendation module re-obtains the recommendation weights for all students' sorted knowledge point pairs (i,j) based on the updated grade module. And assign recommendation weights to the same knowledge points for different students on pairs (i,j). The teaching recommendation weights are obtained by summing the weights and then ranked to verify the teaching outcomes. The recommendation weights for the same knowledge point pair (i,j) are calculated for different students. If the sum of the values ​​results in a decrease in the teaching recommendation weight, it proves that the teaching is effective; if the value of the teaching recommendation weight increases, the teaching should be adjusted.

7. The artificial intelligence-based education system according to claim 1, characterized in that, After students complete the questions pushed to them, the grade module is updated, and the analysis and recommendation module pushes new questions based on the updated grade module.

Images