AR teaching material rapid generation method and related equipment

Abstract

The invention belongs to the technical field of augmented reality teaching, and discloses an AR teaching material rapid generation method and related equipment, basic AR teaching materials and extended AR teaching materials are prepared in advance, data volumes and learning gain scores of the basic AR teaching materials and the extended AR teaching materials are quantified, and according to natural language feedback of students, the mastery degree of the students on knowledge points is evaluated, so that the learning efficiency of the students is improved. The method comprises the following steps: acquiring the mastering degree of students, differentially generating a plurality of teaching material generation schemes according to the mastering degree, evaluating the performance of each teaching material generation scheme from two dimensions of bandwidth saving and learning gain improvement, finally selecting the teaching material generation scheme with the best performance, and generating and distributing AR teaching materials for each student according to the teaching material generation scheme; therefore, the teaching materials can be dynamically adjusted according to the feedback of the students, while personalized teaching is ensured, bandwidth resource utilization is optimized, and rapid, efficient and high-quality AR teaching material generation and distribution are realized.

Description

Technical Field

[0001] The present application relates to the field of augmented reality teaching technology, and more specifically, to a method for quickly generating AR teaching materials and related equipment. Background Art

[0002] Augmented reality (AR) teaching technology is gradually becoming an important innovative force in the field of education. In teaching scenarios with a large number of students, the use of augmented reality glasses to assist teaching has been seen as a key means to improve the interactivity and personalization of teaching. Advances in natural language processing (NLP) technology have made it possible to quickly generate customized AR teaching materials. However, in teaching scenarios with a large number of students, the teaching environment is complex, and the learning progress and comprehension abilities of different students vary significantly. Teachers need to quickly generate and push differentiated augmented reality teaching content based on real-time feedback from students to meet the learning needs of different students.

[0003] Existing AR teaching material generation and distribution technologies are difficult to simultaneously take into account rapid content generation, personalized customization, and efficient distribution in large-scale teaching scenarios. Especially when content customization is required based on students' natural language feedback, how to achieve a balance between the efficiency of teaching content distribution and the quality of personalized teaching under limited network bandwidth and the computing power of augmented reality glasses is still a technical problem that needs to be solved urgently.

[0004] In view of the above problems, the existing technology needs to be improved urgently. Summary of the invention

[0005] The purpose of this application is to provide a method for quickly generating AR teaching materials and related equipment, which can dynamically generate AR teaching materials based on student feedback in large-scale teaching scenarios, taking into account the rapid generation, personalized customization and efficient distribution of teaching materials.

[0006] In a first aspect, the present application provides a method for quickly generating AR teaching materials, which is used to dynamically generate AR teaching materials according to student feedback, and the method comprises the following steps: S1. Retrieve the basic AR teaching materials and extended AR teaching materials pre-entered by the teacher according to the target knowledge points from the teaching database, and determine the amount of data and learning gain score of the basic AR teaching materials and the extended AR teaching materials; S2. receiving natural language feedback on the target knowledge point submitted by the student group through the augmented reality glasses, analyzing the feedback content using natural language processing technology, estimating the mastery of the target knowledge point by each student, and generating multiple teaching material generation schemes based on the mastery; the teaching material generation schemes include a composition method of AR teaching materials corresponding to each student, and the composition method includes only basic AR teaching materials or basic AR teaching materials and extended AR teaching materials; S3. Calculate the bandwidth saving rate of each teaching material generation scheme according to each teaching material generation scheme, the amount of data of the basic AR teaching material and the amount of data of the extended AR teaching material; S4. Calculate the learning gain improvement rate of each teaching material generation scheme according to each teaching material generation scheme, the learning gain score of the basic AR teaching material and the learning gain score of the extended AR teaching material; S5. Calculate the comprehensive effectiveness score of each teaching material generation scheme according to the bandwidth saving rate and the learning gain improvement rate; S6. Generate AR teaching materials for each student based on the teaching material generation scheme with the highest comprehensive effectiveness score, and send them to the augmented reality glasses of each student accordingly.

[0007] This method dynamically generates AR teaching materials based on student feedback, taking into account bandwidth savings and learning gains. It can dynamically generate AR teaching materials based on student feedback in large-scale teaching scenarios, taking into account rapid generation, personalized customization and efficient distribution of teaching materials.

[0008] Preferably, step S1 comprises: S101. Retrieve basic AR teaching materials and extended AR teaching materials pre-entered by the teacher according to the target knowledge points from the teaching database; S102. Reading storage information of basic AR teaching materials and extended AR teaching materials from the teaching database, parsing the storage information to determine the file size of each teaching material as the data volume; S103. According to the preset learning gain scoring criteria, combined with the content depth of the teaching materials to be scored, the average cognitive level of the target audience, and the amount of data, the basic AR teaching materials and the extended AR teaching materials are scored to obtain a learning gain score.

[0009] Through the synergistic effect of the above steps, the data volume and learning gain scores of basic AR teaching materials and extended AR teaching materials are effectively determined, which provides a guarantee for the reliability and effectiveness of the entire AR teaching material rapid generation method.

[0010] Preferably, step S2 comprises: S201. Receive voice feedback on the target knowledge point submitted by the student group through the augmented reality glasses, and use a voice activity detection algorithm to filter out background noise in the voice feedback to obtain a pure voice signal; S202. Performing speech recognition on the clean speech signal, converting the clean speech signal into text feedback, and performing word segmentation processing on the text feedback using a sliding window technology to obtain multiple word fragments; S203. For each student, based on the word fragments, the frequency of occurrence of keywords associated with the target knowledge point in the text feedback is counted, and the student's mastery of the target knowledge point is calculated in combination with the knowledge point mastery evaluation rules; S204. Based on the mastery level scores, the students are divided into two categories of high mastery level and low mastery level according to a plurality of different division criteria, and a plurality of division results are obtained; S205. According to each division result, determine the composition of AR teaching materials corresponding to students with different mastery levels, and obtain the corresponding teaching material generation plan; the composition of AR teaching materials corresponding to students with high mastery levels only includes basic AR teaching materials, and the composition of AR teaching materials corresponding to students with low mastery levels includes basic AR teaching materials and extended AR teaching materials.

[0011] Through the above steps, it is possible to quickly generate personalized AR teaching materials based on student feedback, and achieve a balance between the efficiency of teaching content distribution and the quality of personalized teaching.

[0012] Preferably, step S203 includes: Analyze the semantic distance between each word fragment and the target knowledge point in the preset knowledge graph, and select keywords with a semantic distance less than a preset threshold as keywords associated with the target knowledge point; For each student, the frequency of keywords associated with the target knowledge point in the text feedback is counted, and the frequencies of different keywords are weighted according to the preset keyword weight table to obtain the weighted keyword frequency; The students' mastery scores of target knowledge points are calculated based on the weighted keyword frequency and the knowledge point mastery evaluation rules. The knowledge point mastery evaluation rules include: calculating the students' mastery scores based on the deviation between the weighted keyword frequency and the preset benchmark value, and the degree of deviation is positively correlated with the mastery score.

[0013] Preferably, step S204 includes: Extract each student's historical learning data; historical learning data includes historical question answering accuracy, learning time, and knowledge point mastering preference; According to the historical learning data, a clustering algorithm matching the characteristics of the student group is selected; the clustering algorithm includes a K-means clustering algorithm, a hierarchical clustering algorithm and a density clustering algorithm; Using the selected clustering algorithm, according to the mastery level scores and the preset clustering number range, the students are clustered and analyzed to obtain multiple clustering results; each clustering result represents a student grouping scheme, each grouping scheme contains multiple student subsets, and each student subset represents a group of students with similar mastery levels; For each clustering result, the average mastery score of each student subset is calculated, and the average mastery score of each student subset is compared with the preset mastery score threshold. If the average mastery score is higher than the mastery score threshold, the corresponding student subset is judged to have a high mastery level; otherwise, the corresponding student subset is judged to have a low mastery level, thereby obtaining multiple division results.

[0014] Preferably, step S3 comprises: S301. According to the number of students corresponding to the two AR teaching material composition methods in each teaching material generation scheme, the data volume of the basic AR teaching material and the data volume of the extended AR teaching material, the total data volume of the teaching materials corresponding to each teaching material generation scheme is calculated, recorded as the first data volume; S302. Calculate the total amount of teaching materials data when all students' AR teaching materials include basic AR teaching materials and extended AR teaching materials, recorded as the second amount of data; S303. Calculate the bandwidth saving rate of each teaching material generation scheme according to the following formula: J=1-M1 / M2, wherein J is the bandwidth saving rate, M1 is the first data volume, and M2 is the second data volume.

[0015] Preferably, step S4 comprises: S401. Taking the learning gain score of the basic AR teaching material as the first learning gain score, and taking the weighted average of the learning gain score of the basic AR teaching material and the learning gain score of the extended AR teaching material as the second learning gain score; S402. Calculate the total learning gain score of each teaching material generation scheme according to the number of students corresponding to the two AR teaching material composition methods in each teaching material generation scheme, the first learning gain score and the second learning gain score, and record it as the first total learning gain score; S403. According to the second learning gain score, calculate the total learning gain score when the AR teaching material composition of all students includes basic AR teaching materials and extended AR teaching materials, and record it as the second total learning gain score; S404. Calculate the learning gain improvement rate of each teaching material generation plan according to the following formula: R=(N1-N2) / N2, where R is the learning gain improvement rate, N1 is the first total learning gain score, and N2 is the second total learning gain score.

[0016] Preferably, step S5 comprises: S501. Calculate the distribution of bandwidth saving rate and learning gain improvement rate in historical teaching data respectively, and normalize the current bandwidth saving rate and the current learning gain improvement rate by using a standardization method; S502. Obtaining bandwidth resource information of the AR teaching system to determine bandwidth saving weights and learning gain weights; S503. Based on the normalized bandwidth saving rate and learning gain improvement rate, as well as the determined bandwidth saving weight and learning gain weight, the comprehensive effectiveness score of each teaching material generation scheme is calculated by weighted summation.

[0017] In a second aspect, the present application provides an electronic device, including a processor and a memory, wherein the memory stores a computer program executable by the processor, and when the processor executes the computer program, it runs the steps in the method for quickly generating AR teaching materials as described above.

[0018] In a third aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps in the method for quickly generating AR teaching materials as described above are executed.

[0019] Beneficial effects: The present application provides a method for quickly generating AR teaching materials and related equipment, which dynamically generates AR teaching materials through student feedback, taking into account bandwidth saving and learning gain. It can dynamically generate AR teaching materials based on student feedback in large-scale teaching scenarios, taking into account rapid generation, personalized customization and efficient distribution of teaching materials. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 Flow chart of the method for quickly generating AR teaching materials provided in an embodiment of the present application.

[0021] Figure 2 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

[0022] Description of reference numerals: 301, processor; 302, memory; 303, communication bus. DETAILED DESCRIPTION

[0023] The technical solutions in the present application will be clearly and completely described below in conjunction with the drawings in the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. The components of the present application generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the present application for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present application.

[0024] It should be noted that similar reference numerals and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. At the same time, in the description of this application, the terms "first", "second", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance.

[0025] refer to Figure 1 , this application proposes a method for quickly generating AR teaching materials, which is used to dynamically generate AR teaching materials based on student feedback. The method includes the following steps: S1. retrieve the basic AR teaching materials and extended AR teaching materials pre-entered by the teacher according to the target knowledge points from the teaching database, and determine the data volume and learning gain scores of the basic AR teaching materials and extended AR teaching materials; S2. Receive natural language feedback on target knowledge points submitted by the student group through augmented reality glasses, analyze the feedback content using natural language processing technology, estimate the mastery of the target knowledge points by each student, and generate multiple teaching material generation plans based on the mastery level; the teaching material generation plan includes the composition of AR teaching materials corresponding to each student, which may include only basic AR teaching materials or basic AR teaching materials and extended AR teaching materials; S3. Calculate the bandwidth saving rate of each teaching material generation scheme according to the amount of data of each teaching material generation scheme, the basic AR teaching material, and the amount of data of the extended AR teaching material; S4. Calculate the learning gain improvement rate of each teaching material generation scheme according to each teaching material generation scheme, the learning gain score of the basic AR teaching material, and the learning gain score of the extended AR teaching material; S5. Calculate the comprehensive effectiveness score of each teaching material generation scheme based on the bandwidth saving rate and the learning gain improvement rate; S6. Generate AR teaching materials for each student based on the teaching material generation plan with the highest comprehensive effectiveness score, and send them to the augmented reality glasses of each student accordingly.

[0026] Among them, in step S1, basic AR teaching materials and extended AR teaching materials are pre-entered to provide content reserves for the dynamic generation of subsequent teaching materials. The data volume of basic AR teaching materials and extended AR teaching materials is determined to provide data support for the subsequent calculation of bandwidth saving rate. The determination of learning gain scores is intended to quantify the teaching effect of teaching materials and provide a basis for the calculation of subsequent learning gain improvement rate and comprehensive effectiveness score. Among them, basic AR teaching materials may include at least one of the basic concepts, core principles, typical examples, etc. of knowledge points, and extended AR teaching materials may include at least one of the in-depth expansion of knowledge points, application cases, cutting-edge progress, etc.

[0027] Among them, in step S2, natural language feedback submitted by students through augmented reality glasses is received, and the student feedback content is analyzed using natural language processing technology, with the aim of evaluating the students' mastery of the target knowledge points. Multiple teaching material generation schemes are generated based on the students' mastery level, providing selection space for the screening and optimization of subsequent schemes. The teaching material generation scheme is designed to include a composition method of AR teaching materials corresponding to each student, and the composition method is set to include only basic AR teaching materials or basic AR teaching materials and extended AR teaching materials to achieve differentiation and personalization of teaching content.

[0028] Among them, in step S3, according to each teaching material generation scheme, the data volume of the basic AR teaching material and the data volume of the extended AR teaching material, the bandwidth saving rate of each teaching material generation scheme is calculated to evaluate the efficiency of different schemes in bandwidth resource utilization.

[0029] Among them, in step S4, the learning gain improvement rate of each teaching material generation scheme is calculated according to each teaching material generation scheme, the learning gain score of the basic AR teaching material and the learning gain score of the extended AR teaching material, so as to evaluate the potential of different schemes in improving teaching effects.

[0030] Among them, in step S5, the comprehensive effectiveness score of each teaching material generation scheme is calculated based on the bandwidth saving rate and the learning gain improvement rate, aiming to comprehensively consider the teaching efficiency and teaching quality and provide a comprehensive evaluation index for the final selection of the scheme.

[0031] Among them, in step S6, AR teaching materials for each student are generated according to the teaching material generation plan with the highest comprehensive effectiveness score, and are sent to the augmented reality glasses of each student accordingly to achieve the final distribution and application of the teaching materials.

[0032] Specifically, this method aims to solve the problem that it is difficult to balance the efficiency of AR teaching material generation and distribution and the quality of personalized teaching in large-scale teaching scenarios. First, basic AR teaching materials and extended AR teaching materials are prepared in advance, and their data volume and learning gain scores are quantified to provide a data basis for subsequent steps. Subsequently, students' natural language feedback is received through augmented reality glasses, and natural language processing technology is used to analyze student feedback, evaluate students' mastery of knowledge points, and generate multiple teaching material generation schemes based on their mastery. The core of these schemes is to configure different teaching material combinations for students with different mastery levels, that is, students with high mastery levels only receive basic AR teaching materials, and students with low mastery levels receive basic AR teaching materials and extended AR teaching materials. After that, the performance of each teaching material generation scheme is evaluated from two dimensions: bandwidth saving and learning gain improvement. The bandwidth saving rate is calculated to quantify the efficiency of each scheme in data transmission. The learning gain improvement rate is calculated to evaluate the potential of each scheme in improving teaching effectiveness. Taking into account the bandwidth saving rate and the learning gain improvement rate, the comprehensive effectiveness score of each scheme is calculated to balance teaching efficiency and teaching quality. Finally, the teaching material generation scheme with the highest comprehensive effectiveness score is selected, and AR teaching materials are generated and distributed to each student accordingly. Through the above steps, this method can dynamically adjust teaching materials according to student feedback, optimize bandwidth resource utilization while ensuring personalized teaching, and achieve fast, efficient and high-quality AR teaching material generation and distribution.

[0033] In some specific embodiments, in step S1, the teacher enters basic AR teaching materials and extended AR teaching materials for the target knowledge points in advance. For example, for the knowledge point of "Newton's First Law", the basic AR teaching material can be a three-dimensional model animation containing basic concepts and simple examples, and the extended AR teaching material can be an interactive simulation experiment containing more in-depth principle analysis and complex application scenarios. In step S2, the student submits feedback on "Newton's First Law" in the form of voice through augmented reality glasses, such as "I don't quite understand what inertia is" or "I feel I have basically mastered it." Natural language processing technology analyzes these feedbacks to evaluate the students' mastery of "Newton's First Law." The mastery level assessment can adopt a combination of keyword matching and semantic analysis. For example, the appearance of negative keywords such as "don't understand" and "don't master", as well as deviations in the understanding of core concepts such as "inertia", will be identified by the system as a low level of mastery. On the contrary, if the student feedback contains positive keywords such as "understood" and "mastered", and can correctly apply the relevant concepts, it will be assessed as a high level of mastery. Based on the mastery level assessment results, the system generates multiple teaching material generation schemes. For example, in scheme 1, students with high mastery receive only basic AR teaching materials, and students with low mastery receive basic AR teaching materials and extended AR teaching materials; in scheme 2, students with slightly low mastery receive only basic AR teaching materials, and students with severely low mastery receive basic AR teaching materials and extended AR teaching materials. In step S3, it is assumed that the data volume of the basic AR teaching materials is 10MB, and the data volume of the extended AR teaching materials is 20MB. For scheme 1, if 30 students out of 100 have low mastery, the total data volume of the teaching materials is (70*10MB)+(30*(10MB+20MB))=1600MB. If all students receive basic AR teaching materials and extended AR teaching materials, the total data volume of the teaching materials is 100*(10MB+20MB)=3000MB. The bandwidth saving rate is 1-(1600MB / 3000MB)=46.7%. In step S4, it is assumed that the learning gain score of the basic AR teaching materials is 0.7, and the learning gain score of the extended AR teaching materials is 0.9. For scheme one, the total learning gain score is (70*0.7)+(30*0.9)=76. If all students receive basic AR teaching materials and extended AR teaching materials, the total learning gain score is 100*0.9=90. The learning gain improvement rate is (76-90) / 90=-15.6%. This is a negative improvement rate. In actual applications, it needs to be adjusted according to the specific scoring model and parameters. The goal is to improve learning gain. In step S5, assume that the bandwidth saving weight is 0.6 and the learning gain weight is 0.4. The comprehensive effectiveness score of scheme one is (0.6*normalized bandwidth saving rate)+(0.4*normalized learning gain improvement rate).By comparing the comprehensive effectiveness scores of different schemes, the scheme with the highest score is selected as the final teaching material generation scheme. In step S6, according to the selected scheme, the system automatically generates and sends teaching content containing only basic AR teaching materials to students with a high level of mastery, and sends teaching content containing basic AR teaching materials and extended AR teaching materials to students with a low level of mastery, and distributes the teaching materials to students' augmented reality glasses through the network, so as to realize the rapid generation and efficient distribution of personalized teaching content. Through the above embodiments, the application of the AR teaching material rapid generation method in actual teaching scenarios is demonstrated, and its effectiveness in solving the problem of balancing teaching efficiency and personalized teaching quality is verified.

[0034] In some embodiments, step S1 comprises: S101. Retrieve basic AR teaching materials and extended AR teaching materials pre-entered by the teacher according to the target knowledge points from the teaching database; S102. Reading storage information of basic AR teaching materials and extended AR teaching materials from the teaching database, parsing the storage information to determine the file size of each teaching material as the data volume; S103. According to the preset learning gain scoring criteria, combined with the content depth of the teaching materials to be scored, the average cognitive level of the target audience, and the amount of data, the basic AR teaching materials and the extended AR teaching materials are scored to obtain a learning gain score.

[0035] Among them, in step S101, the teaching database is configured to store pre-entered teaching materials. The teacher prepares basic AR teaching materials and extended AR teaching materials in advance according to the teaching target knowledge points, and enters these materials into the teaching database. The entry process may include file uploading, data annotation and information registration, etc., to ensure that each teaching material is associated with a specific target knowledge point and can be retrieved and called by the system.

[0036] Among them, in step S102, the storage information parsing process is specifically that the system reads the metadata of the teaching materials from the teaching database. The metadata includes information such as the file storage path, file type, and file size. The system parses the file size information and directly uses the file size value as the data volume of the teaching materials. The file size is measured in bytes, KB, MB, etc.

[0037] Among them, in step S103, the preset learning gain scoring standard is a rule for quantifying the effectiveness of teaching materials. The standard may include multiple dimensions such as content depth, average cognitive level of the target audience, data volume, etc. For different content types (such as video, three-dimensional model, text), the learning gain scoring standard may be different. The content depth can be pre-set by the teacher or evaluated by the system through a content analysis algorithm. The average cognitive level of the target audience can be estimated based on the students' historical learning data or test results. The data volume is the file size determined in step S102. The scoring process can adopt weighted average, expert scoring, model prediction and other methods to obtain the final learning gain score by integrating information from each dimension.

[0038] Specifically, through the above method, the method for determining the data volume and learning gain score of the basic AR teaching materials and the extended AR teaching materials is clarified. Step S101 ensures that the source of the teaching materials is controllable and prepared in advance, providing a data basis for subsequent steps; Step S102 provides a direct and easy-to-implement method for determining the data volume, which obtains the file size by parsing the file storage information, thereby quantifying the data resource occupancy of the teaching materials; Step S103 constructs a comprehensive learning gain scoring mechanism to evaluate the learning effect of the teaching materials from multiple angles such as content quality, teaching objects and resource consumption. This multi-dimensional evaluation method makes the learning gain score more objective and reasonable, and provides an important basis for the optimization of the subsequent teaching material generation plan. Through the synergistic effect of steps S101, S102 and S103, the data volume and learning gain score of the basic AR teaching materials and the extended AR teaching materials are effectively determined, which provides a guarantee for the reliability and effectiveness of the entire AR teaching material rapid generation method.

[0039] In some specific implementations, the teaching database is constructed using a relational database or a NoSQL database to store teaching materials and related metadata. The teacher uploads the AR teaching materials through the teaching content management platform, and the platform automatically extracts and stores the file name, file type, file size, storage path and other information of the teaching materials as storage information. In the formulation of the learning gain scoring standard, for video teaching materials, indicators such as content depth level (such as: superficial, medium, in-depth), average cognitive level of the target audience (such as: elementary, intermediate, advanced) and data volume range (such as: <10MB, 10-50MB, >50MB) can be set, and a corresponding scoring value is set for each indicator level. The learning gain score of the video teaching material is obtained by table lookup or weighted calculation. For example, a video teaching material with medium content depth, intermediate average cognitive level of the target audience and 30MB of data can have a learning gain score calculated as 75 points.

[0040] In some preferred implementations, step S103 includes: Identify the content type of instructional materials to be graded; content types include video, 3D model, and text; The corresponding learning gain scoring model is called according to the content type; among them, the learning gain scoring model of video teaching materials takes the weighted average of video duration, information density and interactivity indicators, as well as content depth, average cognitive level of the target audience and data volume as input, and outputs the video learning gain scoring sub-items; the learning gain scoring model of three-dimensional model teaching materials takes the weighted average of calculation model complexity, interaction mode and visualization effect indicators, as well as content depth, average cognitive level of the target audience and data volume as input, and outputs the three-dimensional model learning gain scoring sub-items; the learning gain scoring model of text teaching materials takes the weighted average of text length, information organization structure and degree of text and picture combination indicators, as well as content depth, average cognitive level of the target audience and data volume as input, and outputs the text learning gain scoring sub-items; Adjust the weight of each indicator in the learning gain scoring model according to the average cognitive level of the target audience; The learning gain scoring model with adjusted weights is used to calculate the learning gain scoring items for each type of content in the teaching materials to be scored, and the learning gain scores of the teaching materials to be scored are calculated by combining these learning gain scoring items.

[0041] Among them, identifying the content type of the teaching material to be graded refers to determining which of the teaching materials belongs to, video, three-dimensional model or text. Specifically, the identification of the content type can be achieved by file extension analysis, content format recognition or user manual marking. For example, for uploaded teaching material files, the system can first check the file extension, such as ".mp4", ".avi" is identified as a video type, ".obj", ".stl" is identified as a three-dimensional model type, and ".txt", ".pdf" is identified as a text type. Furthermore, the content format recognition technology can be combined, such as by analyzing specific bytes or metadata in the file header, to more accurately determine the file type. As a supplement, the user is allowed to manually specify the content type when uploading teaching materials to handle errors or special circumstances that may occur in automatic recognition.

[0042] Among them, calling the corresponding learning gain scoring model according to the content type means that the system maintains a learning gain scoring model library, which stores learning gain scoring models for different content types. After identifying the content type of the teaching material to be scored, the system will automatically select a model that matches the content type from the model library. For example, if the teaching material to be scored is identified as a video type, the system calls the learning gain scoring model for video teaching materials. The model library can be constructed and managed in the form of a database or configuration file to facilitate the addition, update and maintenance of the model. Among them, each learning gain scoring model can be a calculation formula model or a deep learning model based on expert experience training.

[0043] Among them, the learning gain scoring model for video teaching materials takes the weighted average of video duration, information density and interactivity indicators, as well as content depth, average cognitive level of the target audience and data volume as input, and outputs the video learning gain scoring sub-items, which means that this model is specifically used to evaluate the learning gain of video teaching materials. The input parameters of the model include the characteristic indicators of the video itself (video duration, information density, interactivity) and the general evaluation indicators of teaching materials (content depth, average cognitive level of the target audience, data volume). Video duration can be directly extracted from the metadata of the video file. Information density can use natural language processing technology to analyze the subtitles or voice content of the video to calculate the amount of information contained in a unit of time. Interactive indicators can be evaluated by counting the number and type of interactive elements in the video, such as questions, barrages, chapter quizzes, etc. The model performs weighted average of these input parameters to obtain the video learning gain scoring sub-items. The weight value can be pre-set based on teaching practice data or expert experience, and can be adjusted later.

[0044] Among them, the learning gain scoring model for 3D model teaching materials takes the weighted average of the model complexity, interaction mode and visualization effect indicators as well as the content depth, the average cognitive level of the target audience and the amount of data as input, and outputs the 3D model learning gain scoring sub-items, which means that the model is used to evaluate the learning gain of 3D model teaching materials. The input parameters of the model include the characteristic indicators of the 3D model itself (model complexity, interaction mode, visualization effect) and the general evaluation indicators of teaching materials (content depth, average cognitive level of the target audience, data volume). The model complexity can be calculated from parameters such as the file size, number of polygons or number of vertices of the model. The interaction mode can be evaluated based on the type and richness of interactive operations supported by the model, such as rotation, scaling, translation, sectioning, animation demonstration, etc. The visualization effect can use image processing technology to analyze the rendering quality, texture details and lighting effects of the model. The model performs weighted average of these input parameters to obtain the 3D model learning gain scoring sub-items. The weight value can also be set and adjusted based on teaching practice data or expert experience.

[0045] Among them, the learning gain scoring model for text-based teaching materials takes the weighted average of the text length, information organization structure, and the degree of text and pictures as well as the content depth, the average cognitive level of the target audience, and the amount of data as input, and the output is the text learning gain scoring sub-item. This model is used to evaluate the learning gain of text-based teaching materials. The input parameters of the model include the characteristic indicators of the text itself (text length, information organization structure, and the degree of text and pictures) and the general evaluation indicators of teaching materials (content depth, average cognitive level of the target audience, and data volume). The text length can be obtained by directly counting the number of words or paragraphs in the text. The information organization structure can be analyzed by natural language processing technology to analyze the chapter structure, paragraph logic, and keyword distribution of the text. The degree of text and pictures can be evaluated by detecting the number and proportion of non-text elements such as pictures, charts, and formulas in the text. The model performs a weighted average of these input parameters to obtain the text learning gain scoring sub-items. The weight value can be set and adjusted based on teaching practice data or expert experience.

[0046] Among them, adjusting the weights of each indicator in the called learning gain scoring model according to the average cognitive level of the target audience means that, considering that students with different cognitive levels have different focuses on different types of teaching materials, the system will dynamically adjust the weights of each indicator in the scoring model according to the average cognitive level of the target audience. The average cognitive level of the target audience can be obtained based on students' historical learning data, test scores, or teacher evaluations. For example, for students with a higher average cognitive level, the weights of indicators such as information density and model complexity can be appropriately increased, and the weights of indicators such as video length and text length can be reduced; for students with a lower average cognitive level, the opposite weight adjustment can be made. The adjustment of weights can adopt a preset weight adjustment strategy or an adaptive adjustment algorithm based on machine learning.

[0047] Among them, the learning gain scoring model after weight adjustment is used to calculate the learning gain scoring items of various contents in the teaching materials to be scored, and the learning gain score of the teaching materials to be scored is calculated by combining these learning gain scoring items. If a teaching material contains multiple content types, such as both videos and texts, the system will first calculate the learning gain scoring items for each content type using the corresponding learning gain scoring model after adjusting the weights. Then, the system will combine these items to obtain the overall learning gain score of the teaching material. The comprehensive method can use weighted average, summation or other appropriate mathematical models. The weight can be determined according to the proportion or importance of different content types in the teaching materials. The final learning gain score will serve as an important basis for evaluating the quality of teaching materials and selecting teaching materials.

[0048] Specifically, when the application scheme scores the learning gain of teaching materials, it first identifies the content type of the teaching materials, such as video, three-dimensional model or text. After identifying the content type, the system will call the learning gain scoring model corresponding to the content type. For video teaching materials, the scoring model will consider video-specific indicators such as video length, information density and interactivity, as well as general indicators such as content depth, average cognitive level of the target audience and data volume, and calculate the weighted average as the video learning gain scoring item. For three-dimensional model teaching materials, the scoring model focuses on three-dimensional model-specific indicators such as calculation model complexity, interaction mode and visualization effect, and also combines general indicators for weighted average calculation to obtain three-dimensional model learning gain scoring items. For text teaching materials, the scoring model will focus on text-specific indicators such as text length, information organization structure and degree of text and pictures, and combine general indicators for weighted average calculation to obtain text learning gain scoring items. During the scoring process, the system will also dynamically adjust the weights of each indicator according to the average cognitive level of the target audience to make the scoring model more in line with actual teaching needs. If the teaching materials contain multiple content types, the learning gain score items of various content types will be combined to obtain the overall learning gain score of the teaching materials. Therefore, the present application scheme can adopt different scoring models and indicator weights for teaching materials of different content types, thereby achieving a more refined and accurate learning gain evaluation, overcoming the limitations of the unified scoring method in the prior art, and providing more effective support for the optimization and personalized push of teaching materials.

[0049] As a preferred embodiment, the solution of the present application is specifically implemented as follows: For an AR teaching material containing two types of content, video and text, first, the video content and text content are identified by analyzing the file extension and content format. Then, the system calls the video learning gain scoring model and the text learning gain scoring model respectively. In the video learning gain scoring model, the weight of video duration is set to 0.3, the weight of information density is set to 0.4, the weight of interactivity is set to 0.3, and the weights of content depth, average cognitive level of target audience and data volume are set to 0.1 respectively. In the text learning gain scoring model, the weight of text length is set to 0.2, the weight of information organization structure is set to 0.4, the weight of the degree of text and pictures is set to 0.4, and the weights of content depth, average cognitive level of target audience and data volume are set to 0.1 respectively. Assuming that the average cognitive level of the target audience is high, the weights are further adjusted. For example, in the video model, the weight of information density is increased to 0.5 and the weight of video duration is reduced to 0.2; in the text model, the weight of information organization structure is increased to 0.5 and the weight of text length is reduced to 0.1. Using the adjusted weights, the learning gain scoring items of video and text are calculated respectively. Assuming that the video learning gain score is 85 points, the text learning gain score is 90 points, the video content accounts for 60% of the teaching materials, and the text content accounts for 40%, the final teaching material learning gain score is 85*0.6+90*0.4=87 points.

[0050] Through the above technical solution, the present application can use a customized learning gain scoring model to evaluate teaching materials of different content types, and can dynamically adjust the model parameters according to the average cognitive level of the target audience, thereby achieving a more accurate evaluation of the learning gain of teaching materials, and providing a more effective technical means for quality control and personalized recommendation of teaching materials.

[0051] In some embodiments, step S2 comprises: S201. Receive voice feedback on target knowledge points submitted by the student group through augmented reality glasses, and use a voice activity detection algorithm to filter out background noise in the voice feedback to obtain a pure voice signal; S202. Perform speech recognition on the clean speech signal, convert the clean speech signal into text feedback, and use the sliding window technology to perform word segmentation on the text feedback to obtain multiple word fragments; S203. For each student, based on the word fragments, the frequency of occurrence of keywords associated with the target knowledge point in the text feedback is counted, and the student's mastery of the target knowledge point is calculated in combination with the knowledge point mastery evaluation rules; S204. Based on the mastery level scores, the students are divided into two categories of high mastery level and low mastery level according to a plurality of different division criteria, and a plurality of division results are obtained; S205. According to each division result, determine the composition of AR teaching materials corresponding to students with different mastery levels, and obtain the corresponding teaching material generation plan; the composition of AR teaching materials corresponding to students with high mastery levels only includes basic AR teaching materials, and the composition of AR teaching materials corresponding to students with low mastery levels includes basic AR teaching materials and extended AR teaching materials.

[0052] Among them, step S2 refers to using natural language processing technology to analyze students' natural language feedback to evaluate the degree of mastery. Specifically, deep learning models such as recurrent neural networks and Transformer models can be used to implement natural language processing analysis, and traditional methods such as bag-of-words models and TF-IDF can be used to implement natural language processing analysis.

[0053] Among them, step S201 refers to receiving voice feedback on target knowledge points submitted by the student group through the augmented reality glasses, filtering out background noise in the voice feedback using a voice activity detection algorithm, and obtaining a pure voice signal. Specifically, voice activity detection algorithms such as WebRTC VAD and Speex VAD can be used to filter out background noise in voice feedback, hardware devices such as directional microphones and bone conduction microphones can be used to reduce the interference of background noise, and wireless communication technologies such as Wi-Fi and Bluetooth can be used to realize voice data transmission between the augmented reality glasses and the data processing center.

[0054] Among them, step S202 refers to performing speech recognition on the clean speech signal, converting the clean speech signal into text feedback, and performing word segmentation processing on the text feedback using the sliding window technology to obtain multiple word fragments. Specifically, speech recognition engines such as Kaldi and Whisper can be used to realize speech-to-text conversion, and word segmentation tools such as jieba and spaCy can be used for word segmentation processing. The size of the sliding window can be set to 3-5 words, and the sliding step can be set to 1 word.

[0055] Among them, step S203 refers to, for each student, according to the word fragments, counting the frequency of keywords associated with the target knowledge point in the text feedback, combining the knowledge point mastery evaluation rules, and calculating the student's mastery score of the target knowledge point. Specifically, a knowledge graph can be pre-constructed, the knowledge graph includes the target knowledge point and keywords related to the target knowledge point, the keywords can include concept words, attribute words, instance words, etc., the mastery evaluation rules can be linear models, nonlinear models, rule bases, etc., and the mastery score can be a percentage score, a grade score, etc.

[0056] Among them, step S204 refers to dividing students into two categories of high mastery and low mastery based on the mastery score and using multiple different division criteria to obtain multiple division results. Specifically, methods such as threshold division and cluster division can be used. The division criteria can include the absolute value of the mastery score, the relative ranking of the mastery score, the distribution of the mastery score, etc. The clustering algorithm can use K-means clustering, hierarchical clustering, DBSCAN clustering, etc., and the number of clusters can be set to 2-5.

[0057] Among them, step S205 refers to determining the composition of AR teaching materials corresponding to students with different mastery levels according to each division result, and obtaining the corresponding teaching material generation plan; the composition of AR teaching materials corresponding to students with high mastery levels only includes basic AR teaching materials, and the composition of AR teaching materials corresponding to students with low mastery levels includes basic AR teaching materials and extended AR teaching materials.

[0058] Specifically, aiming at evaluating students' mastery of knowledge points by analyzing their natural language feedback, and generating personalized teaching materials generation schemes based on this, this application aims to achieve this goal more accurately and effectively through a series of detailed steps. First, the feedback form is limited to voice feedback, and a voice activity detection algorithm is added to filter out noise. This is because voice interaction is more natural and convenient in AR teaching scenarios, but environmental noise may affect the accuracy of voice recognition. By filtering out noise in advance, the accuracy of subsequent voice recognition can be improved, laying the foundation for subsequent text analysis. Then, the pure voice signal is converted into text feedback, and the sliding window technology is used for word segmentation. Speech recognition is a basic step in natural language processing, and word segmentation decomposes the continuous text stream into discrete word fragments, which is convenient for subsequent keyword extraction and semantic analysis. For each student, the keyword frequency is counted and the mastery score is calculated in combination with the mastery evaluation rules. The keyword frequency reflects the degree of correlation between the student feedback content and the target knowledge point, and the mastery evaluation rules quantify this degree of correlation into comparable scores, providing a basis for subsequent student grouping and teaching material generation scheme selection. Based on the mastery level score, students are grouped according to multiple different division criteria. Different division criteria can produce different student group division schemes, providing more options for the subsequent selection of the optimal teaching material generation scheme. Finally, the composition of the teaching materials is determined based on the grouping results. Students with a high level of mastery only include basic materials, and students with a low level of mastery include basic and extended materials. This differentiated configuration aims to ensure teaching effectiveness while saving bandwidth resources as much as possible and meeting the different learning needs of students with different levels of mastery. In this way, a more accurate assessment of students' mastery level and a more effective generation of teaching materials are achieved, providing strong support for the realization of fast, personalized, and efficient AR teaching material generation. It also makes the natural language processing process more controllable and the evaluation results more reliable, laying the foundation for the personalized generation of subsequent teaching materials.

[0059] Through the above technical solution, this application can more accurately evaluate students' mastery of target knowledge points, and based on this, quickly generate differentiated AR teaching materials for students with different mastery levels. This enables the AR teaching system to dynamically adjust teaching content based on students' real-time feedback, taking into account both teaching efficiency and personalization, and ensuring rapid generation, personalization, and efficient distribution of teaching content in teaching scenarios with a large number of students. Under limited network bandwidth and the computing power of augmented reality glasses, a balance is achieved between the efficiency of teaching content distribution and the quality of personalized teaching.

[0060] In some preferred implementations, step S203 includes: Analyze the semantic distance between each word fragment and the target knowledge point in the preset knowledge graph, and select keywords with a semantic distance less than a preset threshold as keywords associated with the target knowledge point; For each student, the frequency of keywords associated with the target knowledge point in the text feedback is counted, and the frequencies of different keywords are weighted according to the preset keyword weight table to obtain the weighted keyword frequency; The students' mastery scores of target knowledge points are calculated based on the weighted keyword frequency and the knowledge point mastery evaluation rules. The knowledge point mastery evaluation rules include: calculating the students' mastery scores based on the deviation between the weighted keyword frequency and the preset benchmark value, and the degree of deviation is positively correlated with the mastery score.

[0061] Among them, analyzing the semantic distance between each word fragment and the target knowledge point in the preset knowledge graph, and selecting keywords with a semantic distance less than a preset threshold as keywords associated with the target knowledge point means not directly counting the frequency of all keywords, but first analyzing the semantic distance between each word fragment and the target knowledge point in the preset knowledge graph. Specifically, the following method can be used to achieve this: first, construct a preset knowledge graph, which contains the target knowledge point and word fragments related to the target knowledge point; then, use word vector technology, such as Word2Vec or GloVe, to map each word fragment and the target knowledge point to the vector space; then, calculate the semantic distance between the vector of each word fragment and the vector of the target knowledge point, and the semantic distance can be calculated using metrics such as cosine similarity or Euclidean distance; finally, set a preset threshold, and select word fragments with a semantic distance less than the preset threshold as keywords associated with the target knowledge point. Through the analysis of semantic distance, keywords with stronger relevance to the target knowledge point can be effectively selected, and those words that appear in the text but have weak relevance to the current knowledge point can be excluded, ensuring that the keywords used for subsequent mastery assessment are truly around the target knowledge point.

[0062] Among them, for each student, the frequency of keywords associated with the target knowledge point in the text feedback is counted, and the frequencies of different keywords are weighted according to the preset keyword weight table to obtain the weighted keyword frequency, which means that after determining the associated keywords, it is not simply to count the number of times the keywords appear, but to introduce the concept of keyword weight table to weight the frequencies of different keywords. Specifically, the following method can be used to achieve this: first, establish a keyword weight table, which records each keyword and the corresponding weight value. The weight value can be set according to the importance of the keyword or the degree of association with the target knowledge point; then, for each student, the frequency of each associated keyword in the text feedback is counted; then, according to the keyword weight table, the weight value corresponding to each associated keyword is found; finally, the frequency of each associated keyword is multiplied by its corresponding weight value to obtain the weighted keyword frequency. This means that different keywords have different importance in reflecting the students' mastery level. By presetting the weight table, those keywords that better represent the students' understanding level can be highlighted, further optimizing the calculation of the mastery level score, so that the score result can more truly reflect the students' knowledge mastery level.

[0063] Among them, according to the weighted keyword frequency, combined with the knowledge point mastery evaluation rules, the student's mastery score of the target knowledge point is calculated; the knowledge point mastery evaluation rules include: calculating the student's mastery score according to the deviation between the weighted keyword frequency and the preset benchmark value, and the degree of deviation is positively correlated with the mastery score, which means that the mastery evaluation rule is based on the deviation between the weighted keyword frequency and the preset benchmark value to calculate the mastery score, and the degree of deviation is positively correlated with the mastery score. Specifically, the following method can be used to achieve this: first, determine the preset benchmark value, which can be a fixed value set according to historical data or expert experience, or a value dynamically adjusted according to the average level of the student group; then, calculate the deviation between the weighted keyword frequency of each student and the preset benchmark value, and the deviation can be an absolute deviation or a relative deviation; then, according to the preset mastery score function, map the deviation value to the mastery score, and the mastery score function needs to ensure that the degree of deviation is positively correlated with the mastery score. For example, a linear function or a nonlinear function can be used to achieve this. This evaluation rule compares the student's feedback with a preset standard, and quantifies the student's mastery by the size of the deviation, so that the evaluation result is more objective and quantifiable, which is convenient for the system to make subsequent teaching material generation and push decisions.

[0064] The above steps ensure the relevance of keywords to target knowledge points, achieve a more accurate and effective assessment of students' mastery, and provide a more reliable basis for the generation of subsequent teaching materials.

[0065] Through the above technical solution, the present application can more accurately and effectively evaluate students' mastery of target knowledge points, solve the problem of the prior art of ignoring the correlation between keywords and target knowledge points by simply counting keyword frequencies, improve the accuracy and pertinence of the evaluation, and provide more reliable data support for the subsequent generation of AR teaching materials based on students' mastery, thereby helping to improve the personalization and effectiveness of AR teaching.

[0066] In some preferred implementations, step S204 includes: Extract each student's historical learning data; historical learning data includes historical question answering accuracy, learning time, and knowledge point mastering preference; According to historical learning data, select a clustering algorithm that matches the characteristics of the student group; clustering algorithms include K-means clustering algorithm, hierarchical clustering algorithm and density clustering algorithm; Using the selected clustering algorithm, according to the mastery level scores and the preset clustering number range, the students are clustered and analyzed to obtain multiple clustering results; each clustering result represents a student grouping scheme, each grouping scheme contains multiple student subsets, and each student subset represents a group of students with similar mastery levels; For each clustering result, the average mastery score of each student subset is calculated, and the average mastery score of each student subset is compared with the preset mastery score threshold. If the average mastery score is higher than the mastery score threshold, the corresponding student subset is judged to have a high mastery level; otherwise, the corresponding student subset is judged to have a low mastery level, thereby obtaining multiple division results.

[0067] Among them, historical learning data refers to a data set that can reflect students' past learning situation. It can be achieved by students' historical answer records on the learning platform, learning time statistics, and knowledge point mastery preference records. By collecting and organizing these data, we can have a more comprehensive understanding of students' learning characteristics and habits.

[0068] Among them, the accuracy rate of history questions refers to the proportion of the number of questions answered correctly by students in the process of answering history questions to the total number of questions. It can be obtained by counting the answer records of students in previous exercises and tests. The higher the accuracy rate of questions, the higher the students' mastery of knowledge points. Among them, the learning time refers to the length of time students spend in the process of learning history. It can be obtained by recording the length of time students spend on learning activities on the learning platform. The longer the learning time, the more energy students put into learning. Among them, the preference for mastering knowledge points refers to the degree of preference students show for different knowledge points in the process of learning history. It can be obtained by analyzing data such as the length of time students spend studying on different knowledge points, the accuracy rate of questions answered, etc. For example, if students spend a long time studying on a certain knowledge point and have a high accuracy rate of questions answered, it means that students may have a preference for mastering this knowledge point.

[0069] Among them, the clustering algorithm refers to an algorithm used to divide a data set into several non-overlapping subsets, so that the data similarity within the subsets is high, while the data similarity between subsets is low. Specifically, it can be implemented by using K-means clustering algorithm, hierarchical clustering algorithm or density clustering algorithm. These algorithms are mature and commonly used clustering methods that can effectively divide student groups.

[0070] The cluster number range refers to the pre-set value range of the number of cluster results. For example, the cluster number range can be set to 2-5, which means that the cluster analysis will try to generate 2, 3, 4 or 5 cluster results to select the best grouping scheme. The setting of the cluster number range can be adjusted according to the actual teaching scenario and needs. Different cluster results can correspond to different numbers of cluster centers in part or in whole.

[0071] Among them, cluster analysis refers to the process of analyzing data using clustering algorithms to discover the clustering structure contained in the data set. In this solution, cluster analysis refers to using the selected clustering algorithm to cluster students according to their mastery scores and historical learning data to obtain multiple clustering results, each of which represents a student grouping plan.

[0072] Among them, the mastery level scoring threshold refers to the pre-set critical value used to divide students' mastery level, which can be adjusted according to specific scoring standards and teaching requirements.

[0073] The division result refers to the final student grouping result obtained through cluster analysis and threshold comparison. The division result includes multiple student subsets, each of which represents a group of students with similar mastery levels, and each student subset is judged to have a high or low mastery level.

[0074] Specifically, in order to group students more accurately, in step S204, first, the system extracts each student's historical learning data, which includes the accuracy of historical answers, learning time and knowledge point mastering preferences, to comprehensively evaluate the student's learning situation. Then, the system selects a clustering algorithm that matches the characteristics of the student group based on these historical learning data. For example, if the number of student groups is large and the characteristic distribution is relatively uniform, the K-means clustering algorithm can be selected; if the characteristics of the student group present an obvious hierarchical structure, the hierarchical clustering algorithm can be selected; if there is uneven density in the student group, the density clustering algorithm can be selected. Then, the system uses the selected clustering algorithm, combined with the previously obtained mastery score and the preset cluster number range, to perform cluster analysis on the students, and obtain multiple clustering results, each of which represents a possible student grouping scheme. Finally, for each clustering result, the system calculates the average mastery score of each student subset, and compares the average mastery score of each student subset with the preset mastery score threshold, thereby determining whether the corresponding student subset belongs to a high mastery level or a low mastery level, and obtaining the final division result. Through the above steps, students can be divided into groups with different mastery levels in a more scientific and reasonable manner, overcoming the one-sidedness of grouping based solely on current mastery level scores and improving the accuracy and reliability of student grouping.

[0075] Through the above technical solution, this application can more accurately and comprehensively evaluate students' mastery level, overcome the limitation of grouping students based solely on the current mastery level scores, improve the accuracy and scientificity of student grouping, and lay the foundation for the personalized generation and distribution of subsequent AR teaching materials, thereby improving the personalization and effectiveness of AR teaching.

[0076] In some preferred implementations, after step S203 and before step S204, the following steps are further included: S206. Calculate the average delay time of student feedback according to the feedback time of each student's natural language feedback, and modify the mastery score according to the average delay time of student feedback; the longer the average delay time of student feedback, the lower the mastery score; Among them, step S206 first completes the calculation of the average delay time of student feedback, and the average delay time reflects the overall feedback speed of the student group. Then, the average delay time is used to correct the mastery score of each student. The correction method is that the longer the average delay time of the student's feedback, the lower the student's mastery score is.

[0077] Specifically, in the teaching system, after the student submits natural language feedback through the augmented reality glasses, the system records the feedback submission timestamp of each student. Step S206 is first used to calculate the average feedback delay time of the student group. The calculation method can be to set a reference time point (for example, the time point when the teacher asks the question), and then calculate the time difference between the time point when each student submits feedback and the reference time point to obtain the feedback delay time of each student. After that, the average value of the feedback delay time of all students is calculated to obtain the average student feedback delay time. After obtaining the average student feedback delay time, step S206 is used to correct the preliminary mastery score calculated in step S203 according to this average delay time. The correction process can be adopted in many ways. As an example, a delay time threshold and a mastery score adjustment coefficient can be preset. If the student's average feedback delay time exceeds the delay time threshold, the student's mastery score is reduced according to the adjustment coefficient. The size of the adjustment coefficient can be adjusted according to the actual teaching effect to ensure the accuracy of the mastery score. The specific values ​​of the delay time threshold and the adjustment coefficient can be optimized according to teaching practice and data analysis.

[0078] By correcting the mastery score through feedback delay time, the assessment of students' mastery level can be more refined, providing a more reliable basis for the personalized generation of subsequent teaching materials, thereby improving teaching effectiveness.

[0079] Specifically, step S3 includes: S301. According to the number of students corresponding to the two AR teaching material composition modes in each teaching material generation scheme, the amount of data of the basic AR teaching materials and the amount of data of the extended AR teaching materials, the total amount of teaching material data corresponding to each teaching material generation scheme is calculated, recorded as the first amount of data; S302. Calculate the total amount of teaching materials data when all students' AR teaching materials include basic AR teaching materials and extended AR teaching materials, recorded as the second amount of data; S303. Calculate the bandwidth saving rate of each teaching material generation scheme according to the following formula: J=1-M1 / M2, where J is the bandwidth saving rate, M1 is the first data volume, and M2 is the second data volume.

[0080] Among them, in step S301, for each teaching material generation scheme, the number of students assigned with only basic AR teaching materials and the number of students assigned with basic AR teaching materials and extended AR teaching materials are counted. Then, the number of students with only basic AR teaching materials is multiplied by the data volume of basic AR teaching materials to obtain the first part of data volume; the number of students with basic AR teaching materials and extended AR teaching materials is multiplied by the sum of the data volume of basic AR teaching materials and extended AR teaching materials to obtain the second part of data volume; the first part of data volume is added to the second part of data volume to calculate the total teaching material data volume corresponding to each teaching material generation scheme, that is, the first data volume.

[0081] In step S302, it is assumed that all students are assigned a composition method including basic AR teaching materials and extended AR teaching materials, and the total number of students is multiplied by the sum of the data volume of the basic AR teaching materials and the extended AR teaching materials to calculate the second data volume, which represents the total data volume under the maximum bandwidth consumption.

[0082] In step S303, the bandwidth saving rate is calculated using the formula J=1-M1 / M2. M1 represents the first data volume of the current teaching material generation scheme, and M2 represents the second data volume under the maximum bandwidth consumption condition. Through this formula, the bandwidth saving degree of each teaching material generation scheme relative to the maximum bandwidth consumption condition can be quantitatively evaluated.

[0083] Therefore, through the above bandwidth saving rate calculation method, the bandwidth efficiency of different teaching material generation schemes can be effectively quantified and compared.

[0084] Further, step S4 includes: S401. Taking the learning gain score of the basic AR teaching material as the first learning gain score, and taking the weighted average of the learning gain score of the basic AR teaching material and the learning gain score of the extended AR teaching material as the second learning gain score; S402. Calculate the total learning gain score of each teaching material generation scheme according to the number of students corresponding to the two AR teaching material composition methods in each teaching material generation scheme, the first learning gain score and the second learning gain score, and record it as the first total learning gain score; S403. According to the second learning gain score, calculate the total learning gain score when the AR teaching material composition of all students includes basic AR teaching materials and extended AR teaching materials, and record it as the second total learning gain score; S404. Calculate the learning gain improvement rate of each teaching material generation scheme according to the following formula: R = (N1-N2) / N2, where R is the learning gain improvement rate, N1 is the first total learning gain score, and N2 is the second total learning gain score Among them, in step S401, the first learning gain score is set as the learning gain score of the basic AR teaching material, which represents the learning gain level when only the basic teaching material is used. The second learning gain score is set as the weighted average of the learning gain scores of the basic AR teaching material and the extended AR teaching material, reflecting the comprehensive learning gain effect when the two teaching materials are used at the same time. The weight distribution can be adjusted according to the actual teaching scenario and the characteristics of the teaching materials. For example, if the extended AR teaching materials play a greater role in improving the depth of students' understanding, a higher weight can be given to the extended AR teaching materials.

[0085] Among them, in step S402, the number of students whose AR teaching materials composition only includes basic AR teaching materials is multiplied by the first learning gain score to obtain the total learning gain score of the first part, and the number of students whose AR teaching materials composition includes basic AR teaching materials and extended AR teaching materials is multiplied by the second learning gain score to obtain the total learning gain score of the second part, and then the sum of the total learning gain score of the first part and the total learning gain score of the second part is calculated to obtain the first total learning gain score.

[0086] In step S403, the second total learning gain score is calculated as a benchmark value, which represents the total learning gain score when all students receive a combination of basic AR teaching materials and extended AR teaching materials. The establishment of this benchmark value provides a comparison basis for the subsequent calculation of the learning gain improvement rate.

[0087] Among them, in step S404, the learning gain improvement rate is calculated by the formula R=(N1-N2) / N2, where N1 represents the first total learning gain score and N2 represents the second total learning gain score. Thus, the learning gain improvement rate is quantified as the learning gain improvement ratio of different teaching material generation schemes relative to the benchmark scheme, realizing the quantifiable evaluation and comparison of the learning gain effect.

[0088] Specifically, in the process of calculating the learning gain improvement rate, the learning gain scores of the basic AR teaching materials and the extended AR teaching materials are first determined. Then, for each teaching material generation plan, the total learning gain score of the plan is calculated according to the distribution of different types of teaching materials received by students in the plan, as the first total learning gain score. At the same time, the total learning gain score when all students receive a combination of basic AR teaching materials and extended AR teaching materials is calculated as the second total learning gain score. Therefore, by comparing the first total learning gain score and the second total learning gain score, and using the formula to calculate the learning gain improvement rate, the performance of each teaching material generation plan in terms of learning gain can be quantitatively evaluated. The higher the learning gain improvement rate, the greater the potential of the plan to improve the overall learning effect. Through the calculation of the learning gain improvement rate, an important reference for the learning gain dimension can be provided for selecting the teaching material generation plan with the highest comprehensive effectiveness score, thereby optimizing the generation and distribution strategy of teaching materials.

[0089] Specifically, step S5 includes: S501. Calculate the distribution of bandwidth saving rate and learning gain improvement rate in historical teaching data, and normalize the current bandwidth saving rate and the current learning gain improvement rate using a standardized method; S502. Obtaining bandwidth resource information of the AR teaching system to determine bandwidth saving weights and learning gain weights; S503. Based on the normalized bandwidth saving rate and learning gain improvement rate, as well as the determined bandwidth saving weight and learning gain weight, the comprehensive effectiveness score of each teaching material generation scheme is calculated by weighted summation.

[0090] Among them, step S501 performs normalization processing, the purpose of which is to eliminate the inequality between the bandwidth saving rate and the learning gain improvement rate due to different dimensions or numerical ranges. The normalization method can adopt minimum-maximum value normalization or Z-score normalization. For example, when the minimum-maximum value normalization is adopted, the normalization formula can be expressed as: normalized value = (original value-minimum value) / (maximum value-minimum value). Among them, the minimum value and the maximum value are obtained by analyzing the distribution of the bandwidth saving rate and the learning gain improvement rate in the historical teaching data.

[0091] Among them, in step S502, bandwidth resource information can be obtained from the network monitoring module of the AR teaching system, such as the current bandwidth utilization. The bandwidth saving weight and the learning gain weight can be dynamically determined based on the bandwidth resource information. Specifically, when the system bandwidth resources are tight, the bandwidth saving weight is increased and the learning gain weight is reduced; when the system bandwidth resources are sufficient, the learning gain weight is increased and the bandwidth saving weight is reduced. The weight determination method can be linear mapping or nonlinear function mapping.

[0092] Among them, step S503 comprehensively considers the normalized bandwidth saving rate and learning gain improvement rate through weighted summation. The comprehensive effectiveness score calculation formula can be expressed as: comprehensive effectiveness score = (normalized bandwidth saving rate × bandwidth saving weight) + (normalized learning gain improvement rate × learning gain weight).

[0093] Specifically, in order to improve the accuracy of the calculation results of the comprehensive effectiveness score, first, in step S501, by normalizing the bandwidth saving rate and the learning gain improvement rate, the influence caused by the difference in dimension and numerical range is eliminated, so that the two can be compared and integrated on the same scale. Then, in step S502, the bandwidth resource information of the AR teaching system is introduced as the basis for determining the bandwidth saving weight and the learning gain weight, and the dynamic adjustment of the weight parameters is realized, so that the calculation of the comprehensive effectiveness score can adapt to the actual operation state of the teaching system. Finally, in step S503, the normalized bandwidth saving rate and the learning gain improvement rate are effectively integrated through weighted summation, and the comprehensive effectiveness score is obtained according to the dynamically adjusted weight parameters. In this way, a scoring result that can more accurately and comprehensively reflect the comprehensive effectiveness of the teaching material generation scheme can be obtained, providing a basis for the subsequent selection of the optimal scheme.

[0094] Please refer to Figure 2 , Figure 2This is a structural diagram of an electronic device provided in an embodiment of the present application. The present application provides an electronic device, including: a processor 301 and a memory 302. The processor 301 and the memory 302 are interconnected and communicate with each other through a communication bus 303 and / or other forms of connection mechanisms (not shown). The memory 302 stores a computer program executable by the processor 301. When the electronic device is running, the processor 301 executes the computer program to execute the AR teaching material rapid generation method in any optional implementation of the above embodiments to achieve the following functions: retrieve the basic AR teaching materials and extended AR teaching materials pre-entered by the teacher according to the target knowledge points from the teaching database, determine the data volume and learning gain scores of the basic AR teaching materials and extended AR teaching materials; receive natural language feedback on the target knowledge points submitted by the student group through the augmented reality glasses, analyze the feedback content using natural language processing technology, and evaluate the learning gain scores of the basic AR teaching materials and extended AR teaching materials. The mastery degree of each student on the target knowledge point is calculated, and multiple teaching material generation plans are generated based on the mastery degree; the teaching material generation plan includes the composition of the AR teaching materials corresponding to each student, and the composition is only basic AR teaching materials or basic AR teaching materials and extended AR teaching materials; according to each teaching material generation plan, the data volume of the basic AR teaching materials and the data volume of the extended AR teaching materials, the bandwidth saving rate of each teaching material generation plan is calculated; according to each teaching material generation plan, the learning gain score of the basic AR teaching materials and the learning gain score of the extended AR teaching materials, the learning gain improvement rate of each teaching material generation plan is calculated; according to the bandwidth saving rate and the learning gain improvement rate, the comprehensive effectiveness score of each teaching material generation plan is calculated; according to the teaching material generation plan with the highest comprehensive effectiveness score, the AR teaching materials of each student are generated, and sent to the augmented reality glasses of each student accordingly.

[0095] The embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for quickly generating AR teaching materials in any optional implementation of the above embodiment is executed to achieve the following functions: retrieve the basic AR teaching materials and extended AR teaching materials pre-entered by the teacher according to the target knowledge points from the teaching database, determine the data volume and learning gain score of the basic AR teaching materials and extended AR teaching materials; receive natural language feedback on the target knowledge points submitted by the student group through augmented reality glasses, analyze the feedback content using natural language processing technology, estimate the mastery of the target knowledge points by each student, and generate multiple teaching material generation plans based on the mastery; the teaching material generation plan includes a corresponding A composition method of AR teaching materials, which composition method includes only basic AR teaching materials or basic AR teaching materials and extended AR teaching materials; calculating the bandwidth saving rate of each teaching material generation scheme according to each teaching material generation scheme, the data volume of the basic AR teaching materials and the data volume of the extended AR teaching materials; calculating the learning gain improvement rate of each teaching material generation scheme according to each teaching material generation scheme, the learning gain score of the basic AR teaching materials and the learning gain score of the extended AR teaching materials; calculating the comprehensive effectiveness score of each teaching material generation scheme according to the bandwidth saving rate and the learning gain improvement rate; generating AR teaching materials for each student according to the teaching material generation scheme with the highest comprehensive effectiveness score, and sending them to the augmented reality glasses of each student accordingly.

[0096] Among them, the computer-readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (Static Random Access Memory, referred to as SRAM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, referred to as EEPROM), erasable programmable read-only memory (Erasable Programmable Read-Only Memory, referred to as EPROM), programmable read-only memory (Programmable Red-Only Memory, referred to as PROM), read-only memory (Read-Only Memory, referred to as ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0097] In the embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0098] In addition, the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0099] Furthermore, the functional modules in the various embodiments of the present application may be integrated together to form an independent part, or each module may exist separately, or two or more modules may be integrated to form an independent part.

[0100] In this document, relational terms such as first and second, etc. are used merely to distinguish one entity or operation from another entity or operation, but do not necessarily require or imply any such actual relationship or order between these entities or operations.

[0101] The above description is only an embodiment of the present application and is not intended to limit the protection scope of the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for quickly generating AR teaching materials, which is used to dynamically generate AR teaching materials based on student feedback, characterized in that: The method comprises the following steps: S1. Retrieve the basic AR teaching materials and extended AR teaching materials pre-entered by the teacher according to the target knowledge points from the teaching database, and determine the amount of data and learning gain score of the basic AR teaching materials and the extended AR teaching materials; S2. receiving natural language feedback on the target knowledge point submitted by the student group through the augmented reality glasses, analyzing the feedback content using natural language processing technology, estimating the mastery of the target knowledge point by each student, and generating multiple teaching material generation schemes based on the mastery; the teaching material generation schemes include a composition method of AR teaching materials corresponding to each student, and the composition method includes only basic AR teaching materials or basic AR teaching materials and extended AR teaching materials; S3. Calculate the bandwidth saving rate of each teaching material generation scheme according to each teaching material generation scheme, the amount of data of the basic AR teaching material and the amount of data of the extended AR teaching material; S4. Calculate the learning gain improvement rate of each teaching material generation scheme according to each teaching material generation scheme, the learning gain score of the basic AR teaching material and the learning gain score of the extended AR teaching material; S5. Calculate the comprehensive effectiveness score of each teaching material generation scheme according to the bandwidth saving rate and the learning gain improvement rate; S6. Generate AR teaching materials for each student based on the teaching material generation scheme with the highest comprehensive effectiveness score, and send them to the augmented reality glasses of each student accordingly.

2. The method for quickly generating AR teaching materials according to claim 1, characterized in that: Step S1 includes: S101. Retrieve basic AR teaching materials and extended AR teaching materials pre-entered by the teacher according to the target knowledge points from the teaching database; S102. Reading storage information of basic AR teaching materials and extended AR teaching materials from the teaching database, parsing the storage information to determine the file size of each teaching material as the data volume; S103. According to the preset learning gain scoring criteria, combined with the content depth of the teaching materials to be scored, the average cognitive level of the target audience, and the amount of data, the basic AR teaching materials and the extended AR teaching materials are scored to obtain a learning gain score.

3. The method for rapidly generating AR teaching materials according to claim 1, characterized in that: Step S2 includes: S201. Receive voice feedback on the target knowledge point submitted by the student group through the augmented reality glasses, and use a voice activity detection algorithm to filter out background noise in the voice feedback to obtain a pure voice signal; S202. Performing speech recognition on the clean speech signal, converting the clean speech signal into text feedback, and performing word segmentation processing on the text feedback using a sliding window technology to obtain multiple word fragments; S203. For each student, based on the word fragments, the frequency of occurrence of keywords associated with the target knowledge point in the text feedback is counted, and the student's mastery of the target knowledge point is calculated in combination with the knowledge point mastery evaluation rules; S204. Based on the mastery level scores, the students are divided into two categories of high mastery level and low mastery level according to a plurality of different division criteria, and a plurality of division results are obtained; S205. According to each division result, determine the composition of AR teaching materials corresponding to students with different mastery levels, and obtain the corresponding teaching material generation plan; the composition of AR teaching materials corresponding to students with high mastery levels only includes basic AR teaching materials, and the composition of AR teaching materials corresponding to students with low mastery levels includes basic AR teaching materials and extended AR teaching materials.

4. The method for quickly generating AR teaching materials according to claim 3, characterized in that: Step S203 includes: Analyze the semantic distance between each word fragment and the target knowledge point in the preset knowledge graph, and select keywords with a semantic distance less than a preset threshold as keywords associated with the target knowledge point; For each student, the frequency of keywords associated with the target knowledge point in the text feedback is counted, and the frequencies of different keywords are weighted according to the preset keyword weight table to obtain the weighted keyword frequency; The students' mastery scores of target knowledge points are calculated based on the weighted keyword frequency and the knowledge point mastery evaluation rules. The knowledge point mastery evaluation rules include: calculating the students' mastery scores based on the deviation between the weighted keyword frequency and the preset benchmark value, and the degree of deviation is positively correlated with the mastery score.

5. The method for quickly generating AR teaching materials according to claim 3, characterized in that: Step S204 includes: Extract each student's historical learning data; historical learning data includes historical question answering accuracy, learning time, and knowledge point mastering preference; According to the historical learning data, a clustering algorithm matching the characteristics of the student group is selected; the clustering algorithm includes a K-means clustering algorithm, a hierarchical clustering algorithm and a density clustering algorithm; Using the selected clustering algorithm, according to the mastery level scores and the preset clustering number range, the students are clustered and analyzed to obtain multiple clustering results; each clustering result represents a student grouping scheme, each grouping scheme contains multiple student subsets, and each student subset represents a group of students with similar mastery levels; For each clustering result, the average mastery score of each student subset is calculated, and the average mastery score of each student subset is compared with the preset mastery score threshold. If the average mastery score is higher than the mastery score threshold, the corresponding student subset is judged to have a high mastery level; otherwise, the corresponding student subset is judged to have a low mastery level, thereby obtaining multiple division results.

6. The method for rapidly generating AR teaching materials according to claim 1, characterized in that: Step S3 includes: S301. According to the number of students corresponding to the two AR teaching material composition methods in each teaching material generation scheme, the data volume of the basic AR teaching material and the data volume of the extended AR teaching material, the total data volume of the teaching materials corresponding to each teaching material generation scheme is calculated, recorded as the first data volume; S302. Calculate the total amount of teaching materials data when all students' AR teaching materials include basic AR teaching materials and extended AR teaching materials, recorded as the second amount of data; S303. Calculate the bandwidth saving rate of each teaching material generation scheme according to the following formula: J=1-M1 / M2, wherein J is the bandwidth saving rate, M1 is the first data volume, and M2 is the second data volume.

7. The method for rapidly generating AR teaching materials according to claim 1, characterized in that: Step S4 includes: S401. Taking the learning gain score of the basic AR teaching material as the first learning gain score, and taking the weighted average of the learning gain score of the basic AR teaching material and the learning gain score of the extended AR teaching material as the second learning gain score; S402. Calculate the total learning gain score of each teaching material generation scheme according to the number of students corresponding to the two AR teaching material composition methods in each teaching material generation scheme, the first learning gain score and the second learning gain score, and record it as the first total learning gain score; S403. According to the second learning gain score, calculate the total learning gain score when the AR teaching material composition of all students includes basic AR teaching materials and extended AR teaching materials, and record it as the second total learning gain score; S404. Calculate the learning gain improvement rate of each teaching material generation plan according to the following formula: R=(N1-N2) / N2, where R is the learning gain improvement rate, N1 is the first total learning gain score, and N2 is the second total learning gain score.

8. The method for rapidly generating AR teaching materials according to claim 1, characterized in that: Step S5 includes: S501. Calculate the distribution of bandwidth saving rate and learning gain improvement rate in historical teaching data respectively, and normalize the current bandwidth saving rate and the current learning gain improvement rate by using a standardization method; S502. Obtaining bandwidth resource information of the AR teaching system to determine bandwidth saving weights and learning gain weights; S503. Based on the normalized bandwidth saving rate and learning gain improvement rate, as well as the determined bandwidth saving weight and learning gain weight, the comprehensive effectiveness score of each teaching material generation scheme is calculated by weighted summation.

9. An electronic device, characterized in that: The method comprises a processor and a memory, wherein the memory stores a computer program executable by the processor, and when the processor executes the computer program, the steps in the method for quickly generating AR teaching materials as described in any one of claims 1 to 8 are executed.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps in the method for quickly generating AR teaching materials as described in any one of claims 1 to 8 are executed.

Images