Internet-based learning resource sharing management method, device and electronic device

Abstract

The present application relates to the field of intelligent management of learning resource sharing, and specifically discloses a learning resource sharing management method, device and electronic device based on the Internet, which uses multiple learning resource videos of a subject that the students to be analyzed have watched in the past, the knowledge point evaluation data of the subject by the students to be analyzed, and multiple learning resource videos to be recommended related to the subject as input data, and convolutionally encodes these input data respectively through deep learning technology. That is, the input data is processed by deep learning technology. In this way, a more accurate feature representation can be obtained. These feature representations are then analyzed to obtain a classification result for indicating whether the text description of each learning resource to be recommended is suitable for the student. In this way, according to the classification result, the matching of educational resources can be adjusted more accurately, so as to share learning resources suitable for their current learning stage with students, so that the learning effect is better.

Description

Technical Field

[0001] The present application relates to the field of intelligent management of learning resource sharing, and more specifically, to an Internet-based learning resource sharing management method, device and electronic device. Background Art

[0002] There are a series of problems in my country's traditional education model, such as difficulty in obtaining high-quality educational resources, outdated teaching methods, and a single curriculum structure. These problems have led to uneven distribution of educational resources between urban and rural areas and regions, and it is difficult for students to choose appropriate educational resources according to their own learning stage, which affects the learning effect. However, the innovative application of information technology in the field of education provides opportunities to solve these problems. Through the Internet, high-quality educational resources can be delivered to learners at a very low cost, realizing the universal sharing and diversified acquisition of educational resources. This educational resource sharing model gives students more rights to learn independently, meets the public's demand for high-quality and diversified educational resources, and helps promote educational equity and balanced regional development.

[0003] However, the existing technology still has many shortcomings. Educational resource sharing refers to resource sharing based on the Internet. The producers of educational resources can upload various educational resources to the Internet, and people can obtain various educational resources through Internet searches. However, today's students cannot choose educational resources suitable for themselves according to their current learning stage, so as to make better use of these educational resources, resulting in unsatisfactory learning results and a lot of time wasted.

[0004] Therefore, we look forward to an Internet-based learning resource sharing management method, device and electronic equipment that can adjust the matching of educational resources according to the category of educational resources, the school's teaching progress and the students' evaluation results, so as to share with students the learning resources suitable for their current learning stage and achieve better learning results. Summary of the invention

[0005] In order to solve the above-mentioned technical problems, the present application is proposed. The embodiments of the present application provide a kind of learning resource sharing management method, device and electronic device based on the Internet, which uses multiple learning resource videos of a subject that the students to be analyzed have watched in the past in the resource sharing platform, the knowledge point evaluation data of the subject of the students to be analyzed, and multiple learning resource videos to be recommended related to the subject as input data, and convolutionally encodes these input data respectively through deep learning technology. That is, the input data is processed by deep learning technology. In this way, a more accurate feature representation can be obtained. These feature representations are then analyzed to obtain a classification result for indicating whether the text description of each learning resource to be recommended is suitable for the student. In this way, according to the classification result, the matching of educational resources can be adjusted more accurately, so as to share learning resources suitable for their current learning stage with students, so that the learning effect is better.

[0006] According to a first aspect of the present application, there is provided an Internet-based learning resource sharing management method, which comprises:

[0007] Collect multiple learning resource videos of a subject that the students to be analyzed have watched in the past, knowledge point evaluation data of the students to be analyzed on the subject, and multiple learning resource videos to be recommended related to the subject on the resource sharing platform;

[0008] Extract and analyze the features of the multiple learning resource videos of a subject that the student to be analyzed has watched in the past, the knowledge point evaluation data of the student to be analyzed on the subject, and the multiple learning resource videos to be recommended related to the subject to obtain a student learning demand feature vector and multiple learning resource feature vectors;

[0009] Using a joint encoder, the student learning demand feature vector and each learning resource feature vector in the multiple learning resource feature vectors are combined to obtain multiple learning resource matching feature matrices;

[0010] A classification result is obtained based on the matching feature matrix of the multiple learning resources, and the classification result is used to indicate whether the text description of each learning resource to be recommended is suitable for the student.

[0011] In combination with the first aspect of the present application, in an Internet-based learning resource sharing and management method of the first aspect of the present application, feature extraction and analysis are performed on multiple learning resource videos of a subject that the student to be analyzed has watched in the past, the knowledge point evaluation data of the student to be analyzed on the subject, and the multiple learning resource videos to be recommended related to the subject to obtain a student learning need feature vector and multiple learning resource feature vectors, including: feature encoding the multiple learning resource videos of a subject that the student to be analyzed has watched in the past to obtain a global feature vector of learning progress; feature encoding the knowledge point evaluation data of the student to be analyzed on the subject to obtain a student evaluation feature vector; feature fusion of the global feature vector of learning progress and the student evaluation feature vector to obtain the student learning need feature vector; feature encoding the multiple learning resource videos to be recommended related to the subject to obtain the multiple learning resource feature vectors.

[0012] According to a second aspect of the present application, there is provided an Internet-based learning resource sharing management device, comprising:

[0013] The learning subject data collection module is used to collect multiple learning resource videos of a subject that the students to be analyzed have watched in the past, the knowledge point evaluation data of the students to be analyzed on the subject, and multiple learning resource videos to be recommended related to the subject in the resource sharing platform;

[0014] A learning subject feature extraction module is used to extract and analyze features of a plurality of learning resource videos of a subject that the student to be analyzed has watched in the past, the knowledge point evaluation data of the student to be analyzed on the subject, and the plurality of learning resource videos to be recommended related to the subject to obtain a student learning demand feature vector and a plurality of learning resource feature vectors;

[0015] A learning subject feature fusion module, used for combining the student learning demand feature vector with each of the multiple learning resource feature vectors using a joint encoder to obtain multiple learning resource matching feature matrices;

[0016] The learning subject feature classification result generating module is used to obtain a classification result based on the matching feature matrix of the multiple learning resources, and the classification result is used to indicate whether the text description of each learning resource to be recommended is suitable for the student.

[0017] In combination with the second aspect of the present application, in an Internet-based learning resource sharing and management device of the second aspect of the present application, the learning subject feature classification result generation module includes: passing the multiple learning resource matching feature matrices through a learning resource matching feature extraction module based on a convolutional neural network model to obtain multiple learning resource matching classification feature matrices; performing expected activation of feature representation on each of the multiple learning resource matching classification feature matrices to obtain multiple optimized learning resource matching classification feature matrices; passing the multiple optimized learning resource matching classification feature matrices through a classifier respectively to obtain the classification result; and based on the classification result, visually displaying the learning resources suitable for the student.

[0018] According to the third aspect of the present application, an electronic device is provided, comprising: a processor; and a memory, in which computer program instructions are stored, and when the computer program instructions are executed by the processor, the processor executes the Internet-based learning resource sharing and management method as described above.

[0019] Compared with the prior art, the present application provides an Internet-based learning resource sharing management method, device and electronic device, which uses multiple learning resource videos of a subject that the students to be analyzed have watched in the past, the knowledge point evaluation data of the subject by the students to be analyzed, and multiple learning resource videos to be recommended related to the subject as input data, and convolutionally encodes these input data respectively through deep learning technology. That is, the input data is processed by deep learning technology. In this way, a more accurate feature representation can be obtained. These feature representations are then analyzed to obtain a classification result for indicating whether the text description of each learning resource to be recommended is suitable for the student. In this way, according to the classification results, the matching of educational resources can be adjusted more accurately, so as to share learning resources suitable for their current learning stage with students, so that the learning effect is better. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0021] Figure 1 The figure illustrates a flow chart of an Internet-based learning resource sharing and management method according to an embodiment of the present application.

[0022] Figure 2The figure illustrates a flow chart of step S2 in the Internet-based learning resource sharing and management method according to an embodiment of the present application.

[0023] Figure 3 The flowchart of step S21 in step S2 of the Internet-based learning resource sharing and management method according to an embodiment of the present application is illustrated.

[0024] Figure 4 The flowchart of step S22 in step S2 of the Internet-based learning resource sharing and management method according to an embodiment of the present application is illustrated.

[0025] Figure 5 The flowchart of step S24 in step S2 of the Internet-based learning resource sharing and management method according to an embodiment of the present application is illustrated.

[0026] Figure 6 The figure illustrates a flow chart of step S4 in the Internet-based learning resource sharing and management method according to an embodiment of the present application.

[0027] Figure 7 The figure illustrates a schematic block diagram of an Internet-based learning resource sharing and management device according to an embodiment of the present application.

[0028] Figure 8 A block diagram of an electronic device according to an embodiment of the present application is illustrated.

[0029] Fig. 9 A block diagram of a non-transitory computer-readable storage medium according to an embodiment of the present application is illustrated. DETAILED DESCRIPTION

[0030] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0031] Exemplary Methods

[0032] Figure 1 FIG. 1 is a flowchart of a method for sharing and managing learning resources based on the Internet according to an embodiment of the present application. Figure 1As shown, according to the Internet-based learning resource sharing management method of the embodiment of the present application, the method includes: S1, collecting multiple learning resource videos of a subject that the student to be analyzed has watched in the past, the knowledge point evaluation data of the student to be analyzed on the subject, and multiple learning resource videos related to the subject to be recommended in the resource sharing platform; S2, respectively extracting and analyzing the multiple learning resource videos of a subject that the student to be analyzed has watched in the past, the knowledge point evaluation data of the student to be analyzed on the subject, and the multiple learning resource videos related to the subject to be recommended to obtain a student learning demand feature vector and multiple learning resource feature vectors; S3, using a joint encoder to combine the student learning demand feature vector and each learning resource feature vector in the multiple learning resource feature vectors to obtain multiple learning resource matching feature matrices; S4, obtaining a classification result based on the multiple learning resource matching feature matrices, and the classification result is used to indicate whether the text description of each learning resource to be recommended is suitable for the student.

[0033] The traditional education model has many problems, including difficulty in obtaining high-quality educational resources, outdated teaching methods, and a single curriculum structure. These difficulties have led to an uneven distribution of educational resources between different regions and between urban and rural areas, and have also made it difficult for students to choose appropriate resources based on their own learning needs, thus affecting their learning outcomes. However, the innovative application of information technology in the field of education has provided new opportunities to solve these problems. Through the Internet, high-quality educational resources can be delivered to learners at a very low cost, achieving the popularization and diversified acquisition of educational resources. This educational resource sharing model gives students more opportunities for independent learning, meets the public's demand for high-quality and diversified educational resources, and helps promote fairness in education and balanced regional development.

[0034] However, the existing technology still has many deficiencies. Educational resource sharing refers to resource sharing based on the Internet. The producers of educational resources can upload various educational resources to the Internet, and people can obtain various educational resources through Internet searches. However, students nowadays cannot choose educational resources suitable for themselves according to their current learning stage, so as to make better use of these educational resources, resulting in unsatisfactory learning effects and a lot of time wasted. Therefore, a learning resource sharing management method, device and electronic device based on the Internet are expected to adjust the matching of educational resources according to the category of educational resources, the school's teaching progress and the evaluation results of students, so as to share learning resources suitable for their current learning stage with students, so as to achieve better learning effects.

[0035] In step S1, multiple learning resource videos of a subject that the students to be analyzed have watched in the past, the knowledge point evaluation data of the subject that the students to be analyzed, and multiple learning resource videos to be recommended related to the subject are collected in the resource sharing platform. It should be understood that obtaining the learning resources to be recommended for the subject in the resource sharing platform can provide candidate resources for the system to meet the needs of students at a specific learning stage, which is the basis for personalized recommendation for students. Therefore, first collect multiple learning resource videos to be recommended related to the subject as input data. Further, it should be understood that by understanding the teaching progress of a subject in the school where the students are located, the system can adjust the recommended learning resources according to the school teaching progress so that students can keep up with the course progress. In addition, considering that each student has a different degree of mastery of knowledge points, analyzing the evaluation results of the students' knowledge points of the subject can help the system understand the students' learning level and needs, thereby recommending resources suitable for their personal abilities and learning stages. Therefore, it is also necessary to collect multiple learning resource videos of a subject that the students to be analyzed have watched in the past, and the knowledge point evaluation data of the subject that the students to be analyzed are collected in the resource sharing platform.

[0036] In step S2, feature extraction and analysis are performed on multiple learning resource videos of a subject that the student to be analyzed has watched in the past, the knowledge point evaluation data of the subject by the student to be analyzed, and the multiple learning resource videos to be recommended related to the subject to obtain a student learning demand feature vector and multiple learning resource feature vectors. It should be understood that after obtaining these data, the next step is to perform feature extraction and analysis on these input data.

[0037] Specifically, Figure 2 FIG. 1 is a flowchart of step S2 in the Internet-based learning resource sharing management method according to an embodiment of the present application. Figure 2 As shown, step S2, respectively extracts and analyzes features of multiple learning resource videos of a subject that the student to be analyzed has watched in the past, the knowledge point evaluation data of the student to be analyzed on the subject, and the multiple learning resource videos to be recommended related to the subject to obtain a student learning demand feature vector and multiple learning resource feature vectors, including: S21, feature encoding the multiple learning resource videos of a subject that the student to be analyzed has watched in the past to obtain a global feature vector of learning progress; S22, feature encoding the knowledge point evaluation data of the student to be analyzed on the subject to obtain a student evaluation feature vector; S23, feature fusion of the global feature vector of learning progress and the student evaluation feature vector to obtain the student learning demand feature vector; S24, feature encoding the multiple learning resource videos to be recommended related to the subject to obtain the multiple learning resource feature vectors.

[0038] First, feature extraction and analysis are performed on multiple learning resource videos of a certain subject that the student to be analyzed has watched in the past. Specifically, Figure 3 FIG. 1 is a flowchart of step S21 in step S2 of the Internet-based learning resource sharing management method according to an embodiment of the present application. Figure 3 As shown, step S21 includes: step S211, obtaining key frames of each learning resource video in multiple learning resource videos of a certain subject that the student to be analyzed has watched in the past to obtain multiple key frames of each video; step S212, respectively passing the multiple key frames of each video through a learning progress feature extractor based on a deep and shallow feature fusion module to obtain multiple learning progress feature matrices of each video; step S213, arranging the multiple learning progress feature matrices of each video according to the time dimension as a learning progress timing input tensor of each video, and then passing through a learning progress timing feature extraction module including a three-dimensional convolution kernel to obtain a learning progress timing feature graph of each video; step S214, cascading the learning progress timing feature graphs of each video to obtain a learning progress global feature graph; step S215, performing a pooling operation on the learning progress global feature graph to obtain the learning progress global feature vector.

[0039] It should be understood that a video is composed of a series of continuous image frames, and directly extracting features from a video will result in a large amount of data processing and storage requirements. In contrast, extracting the key frames of a video can reduce the amount of data and simplify the subsequent feature extraction and processing process. Therefore, in order to improve the efficiency of feature extraction, first obtain the key frames of each learning resource video in a plurality of learning resource videos of a subject that the student to be analyzed has watched in the past to obtain multiple key frames of each video.

[0040] Next, considering that key frames contain not only shallow features such as color, texture, and edge, but also more complex semantic information and abstract features. Based on this, in order to more comprehensively capture the high-dimensional features in key frames, a learning progress feature extractor based on a deep and shallow feature fusion module is used to perform feature encoding on multiple key frames of each video to obtain multiple learning progress feature matrices of each video. Among them, the deep and shallow feature fusion module can combine the advantages of shallow features and deep features, comprehensively consider the low-level and high-level feature information of the image, and thus more comprehensively describe the video content.

[0041] Specifically, step S212 is used to: extract a shallow feature matrix from the i-th layer of the learning progress feature extractor, the i-th layer being the first to sixth layers of the learning progress feature extractor; extract a deep feature matrix from the j-th layer of the learning progress feature extractor, the ratio between the j-th layer and the i-th layer being greater than or equal to 5; and use the deep and shallow feature fusion module to fuse the shallow feature matrix and the deep feature matrix to obtain each of the multiple learning progress feature matrices of the videos.

[0042] Next, it should be understood that video data has a temporal nature. In order to better preserve these temporal features, the multiple learning progress feature matrices of each video are first arranged according to the time dimension as the learning progress temporal input tensors of each video. Then, a learning progress temporal feature extraction module including a three-dimensional convolution kernel is used to perform feature encoding on the learning progress temporal input tensors of each video to obtain the learning progress temporal feature graph of each video. Among them, the three-dimensional convolution kernel can simultaneously consider the features in the time and space dimensions, so as to better capture the temporal features and spatial features of the video content.

[0043] Specifically, step S213 is used to: use each layer of the learning progress temporal feature extraction module including the three-dimensional convolution kernel to perform the following on the input data in the forward transfer of the layer: perform convolution processing based on the three-dimensional convolution kernel on the input data to obtain a convolution feature map; perform pooling processing based on the local feature matrix on the convolution feature map to obtain a pooled feature map; perform nonlinear activation on the pooled feature map to obtain an activation feature map; wherein the input of the first layer of the learning progress temporal feature extraction module including the three-dimensional convolution kernel is the learning progress temporal input tensor of each video, and the output of the last layer of the learning progress temporal feature extraction module including the three-dimensional convolution kernel is the learning progress temporal feature map of each video.

[0044] Furthermore, considering that different videos may cover different knowledge points, learning contents and teaching styles. Cascading their learning progress time series feature graphs can integrate the information of each video, making the final learning progress global feature graph more representative and comprehensive. Therefore, the learning progress time series feature graphs of each video are further cascaded to obtain a learning progress global feature graph. In this way, the learning progress of the student to be analyzed can be obtained more comprehensively.

[0045] Then, considering that the feature map may contain a large number of dimensions, if the pooling operation is not performed, the feature dimension will be too high, increasing the computational complexity and memory consumption. Therefore, in order to improve the efficiency of data processing, the learning progress global feature map is further pooled to obtain the learning progress global feature vector. In this way, pooling the feature map helps to extract key features, reduce dimensions, improve the robustness and generalization ability of the model, and improve the computational efficiency of the system.

[0046] Then, feature extraction and analysis are performed on the knowledge point assessment data of the subject of the student to be analyzed. Specifically, Figure 4 FIG. 1 is a flowchart of step S22 in step S2 of the Internet-based learning resource sharing management method according to an embodiment of the present application. Figure 4 As shown, step S22 includes: step S221, embedding and encoding the knowledge point evaluation data of the subject to be analyzed by the student to obtain a student evaluation embedding vector; step S222, passing the student evaluation embedding vector through a student evaluation feature extraction module based on a one-dimensional convolutional neural network model to obtain the student evaluation feature vector.

[0047] It should be understood that the knowledge point evaluation data of the subject by the students to be analyzed refers to the students' answers to each knowledge point in the test or quiz, usually including whether the answer is correct, the answer time, etc. These text data are discrete, and it is difficult for the machine to directly recognize these discrete data, which may affect the accuracy of feature extraction. Therefore, the knowledge point evaluation data of the subject by the students to be analyzed is first embedded and encoded to obtain a vector representation that can be recognized by the machine. Among them, the embedded coding converts high-dimensional text data into a low-dimensional dense vector representation, which helps to reduce the dimension of the data, improve computing efficiency, and better capture the semantic information of the text data.

[0048] Next, considering that the one-dimensional convolutional neural network adopts parameter sharing in the convolution process, this means that the same convolution kernel shares weights on the entire input sequence, reducing the number of model parameters and the risk of overfitting. Therefore, in order to better capture the characteristic information of the knowledge point evaluation data of the subject to be analyzed by the student, a student evaluation feature extraction module based on the one-dimensional convolutional neural network model is further used to extract features from the student evaluation embedding vector to obtain the student evaluation feature vector.

[0049] Specifically, step S222 is used to: use each layer of the one-dimensional convolutional neural network model to perform the following on the input data in the forward pass of the layer: perform convolution processing based on the one-dimensional convolution kernel on the input data to obtain a convolution feature map; perform pooling processing based on the feature matrix on the convolution feature map to obtain a pooled feature map; perform nonlinear activation on the pooled feature map to obtain an activation feature map; wherein the input of the first layer of the one-dimensional convolutional neural network model is the student evaluation embedding vector, and the output of the last layer of the one-dimensional convolutional neural network model is the student evaluation feature vector.

[0050] Then, the learning progress global feature vector and the student evaluation feature vector are subjected to feature fusion operation. It should be understood that the learning progress global feature vector and the student evaluation feature vector represent information about the student's current learning stage and learning performance, respectively. Through feature fusion, these two aspects of information can be combined to provide a more comprehensive description of the student, which helps the system better understand the student's current learning status and mastery. Therefore, in order to better understand the student's learning status and mastery of the subject, the learning progress global feature vector and the student evaluation feature vector are subjected to feature fusion to obtain the student's learning needs feature vector.

[0051] Specifically, step S23 is used to: fuse the learning progress global feature vector and the student assessment feature vector using the following cascade formula to obtain the student learning demand feature vector; wherein the cascade formula is: ,in, represents the global feature vector of the learning progress, represents the student evaluation feature vector, represents the student learning demand feature vector, Represents a cascade function.

[0052] Finally, feature extraction and analysis are performed on the multiple recommended learning resource videos related to the subject. Specifically, Figure 5 FIG. 1 is a flowchart of step S24 in step S2 of the Internet-based learning resource sharing management method according to an embodiment of the present application. Figure 5 As shown, step S24 includes: step S241, obtaining the key frames of each of the multiple learning resource videos to be recommended related to the subject to obtain multiple learning resource key frames of each video; step S242, arranging the multiple learning resource key frames of each video into learning resource timing input tensors of each video according to the time dimension, and then passing them through a learning resource timing feature extraction module based on the spatial attention mechanism to obtain multiple learning resource timing feature graphs; step S243, performing pooling operations on the multiple learning resource timing feature graphs respectively to obtain the multiple learning resource feature vectors.

[0053] It should be understood that video data contains a large amount of information, and directly extracting features from video data requires a large amount of storage space and computing power. Therefore, in order to improve the efficiency of data processing, firstly, the key frames of each of the multiple learning resource videos to be recommended related to the subject are obtained, thereby obtaining multiple learning resource key frames of each video.

[0054] Next, considering that video data has temporal characteristics, in order to retain these temporal features, the multiple learning resource key frames of each video are arranged into learning resource temporal input tensors of each video according to the time dimension.

[0055] Then, the learning resource time series feature extraction module based on the spatial attention mechanism is used to extract and analyze the learning resource time series input tensors of each video to obtain the multiple learning resource time series feature graphs. Among them, the spatial attention mechanism can adaptively learn weights according to the characteristics of different learning resource time series, and better adapt to the feature distribution and time series structure of different learning resources. This flexibility and adaptability make the spatial attention mechanism more superior when processing different types of learning resource time series.

[0056] Specifically, step S242 is used to: use the convolution coding part of the learning resource timing feature extraction module to perform deep convolution coding on each of the learning resource timing input tensors of each video to obtain an initial convolution feature map; input the initial convolution feature map into the spatial attention part of the learning resource timing feature extraction module to obtain a spatial attention map; pass the spatial attention map through a Softmax activation function to obtain a spatial attention feature map; calculate the point-by-point multiplication of the spatial attention feature map and the initial convolution feature map to obtain each learning resource timing feature map in the multiple learning resource timing feature maps.

[0057] Further, it should be understood that the feature map captured using the spatial attention mechanism may increase the risk of overfitting due to excessive focus on local details, causing the model to perform poorly on unseen data. Considering that the pooling operation can help extract the main features in the feature map, retain key information and discard unimportant information. This helps to reduce the risk of overfitting and improve the generalization ability of the model. Therefore, the multiple learning resource time series feature maps are further subjected to a global pooling operation based on the feature matrix to obtain the multiple learning resource feature vectors.

[0058] In step S3, a joint encoder is used to perform feature fusion on the student learning demand feature vector and each learning resource feature vector in the multiple learning resource feature vectors to obtain multiple learning resource matching feature matrices. It should be understood that feature fusion can help to effectively integrate student learning demand features and learning resource features, so as to better achieve the matching between students and learning resources. Therefore, a joint encoder is used to perform feature fusion on the student learning demand feature vector and each learning resource feature vector in the multiple learning resource feature vectors to obtain multiple learning resource matching feature matrices. Among them, the joint encoder can promote the interaction and fusion between student learning demand features and learning resource features, help to extract more informative feature representations, and provide better input for subsequent matching recommendations.

[0059] Specifically, step S3 is used to: use the joint encoder to jointly encode the student learning demand feature vector and each learning resource feature vector in the multiple learning resource feature vectors using the following formula to generate each learning resource matching feature matrix in the multiple learning resource matching feature matrices; wherein the formula is: ,in represents vector multiplication, represents the learning resource matching feature matrix, represents the student learning demand feature vector, represents the learning resource feature vector, Represents the transpose of the learning resource feature vector.

[0060] In step S4, a classification result is obtained based on the matching feature matrix of the plurality of learning resources, and the classification result is used to indicate whether the text description of each learning resource to be recommended is suitable for the student. Specifically, Figure 6 FIG. 4 is a flowchart of step S4 in the Internet-based learning resource sharing management method according to an embodiment of the present application. Figure 6 As shown, step S4 includes: step S41, passing the multiple learning resource matching feature matrices through a learning resource matching feature extraction module based on a convolutional neural network model to obtain multiple learning resource matching classification feature matrices; step S42, performing expected activation of feature representation on each of the multiple learning resource matching classification feature matrices to obtain multiple optimized learning resource matching classification feature matrices; step S43, passing the multiple optimized learning resource matching classification feature matrices through a classifier respectively to obtain the classification result; step S44, based on the classification result, visually displaying the learning resources suitable for the student.

[0061] It should be understood that feature fusion alone may not be able to fully explore the complex relationships and patterns between learning resource matching features, resulting in information loss or fuzzification, and reducing matching accuracy. Therefore, in order to better capture the deeper abstract features in the fused features, a learning resource matching feature extraction module based on a convolutional neural network model is used to perform deep convolution encoding on the multiple learning resource matching feature matrices to obtain multiple learning resource matching classification feature matrices. In this way, by using the convolutional neural network model, the high-level representation of the learning resource matching features can be better extracted, thereby improving matching accuracy, generalization ability and model performance, and achieving better learning effects.

[0062] In particular, in the technical solution of the present application, considering that different students have different learning behaviors and habits, this may lead to differences in the distribution of learning progress and assessment data, thereby generating local abnormal distribution in the learning resource matching classification feature matrix. There may be significant differences in the degree of mastery of different knowledge points by students, and this imbalance may be reflected in the learning resource matching classification feature matrix, resulting in local abnormal distribution. The quality of the learning resource videos to be recommended may be uneven, and some videos may contain less key information or the information expression is unclear, which may affect the accuracy of feature extraction and lead to abnormal distribution. In the process of feature extraction and fusion, noise or errors may be introduced, and these factors may produce abnormal distribution in the learning resource matching classification feature matrix. Local abnormal distribution may cause the recommendation system to be unable to accurately judge the degree of matching between learning resources and student needs, thereby recommending learning resources that are not suitable for students. If the recommended learning resources do not match the actual learning needs of students, it may affect the learning effect of students and even lead to a decrease in learning efficiency. In order to reduce the impact of the local abnormal distribution of the learning resource matching classification feature matrix, in the technical solution of the present application, the expected activation of the feature representation of the learning resource matching classification feature matrix is ​​activated to obtain an optimized learning resource matching classification feature matrix.

[0063] Specifically, step S42 includes: calculating an exponential function value raised to the power of the eigenvalue at the predetermined position of the learning resource matching classification feature matrix and then subtracting one to obtain an exponentially adjusted eigenvalue; calculating the exponentially adjusted eigenvalue divided by the eigenvalue at the predetermined position of the learning resource matching classification feature matrix, and calculating the absolute value of the result to obtain a normalized exponentially adjusted eigenvalue ratio; activating the normalized exponentially adjusted eigenvalue ratio through a softmax function and then multiplying it by the eigenvalue at the predetermined position of the learning resource matching classification feature matrix to obtain the eigenvalue at the predetermined position of the optimized learning resource matching classification feature matrix.

[0064] Here, in view of the local abnormal distribution in the learning resource matching classification feature matrix, in the technical solution of the present application, the expected activation of the feature representation of the learning resource matching classification feature matrix is ​​activated, and each element in the learning resource matching classification feature matrix is ​​analyzed to determine its contribution to the final classification result. In this way, those features that contribute more to the classification result can be strengthened, while those features that contribute less or are unstable can be suppressed. In this way, not only can the overall consistency of the learning resource matching classification feature matrix be improved, but also the sensitivity and recognition ability of the model for important features in the data can be enhanced.

[0065] Next, considering that the classifier can classify learning resources according to characteristics such as the student's educational resource category, school teaching progress and evaluation results, personalized learning resource recommendations can be achieved. Different students have different needs for learning resources at different learning stages, and the classifier can help the system more accurately match resources suitable for the student's current learning stage. Therefore, the classifier is used to perform feature classification on the multiple optimized learning resource matching classification feature matrices to obtain a classification result, and the classification result is used to indicate whether the text description of each learning resource to be recommended is suitable for the student.

[0066] Specifically, step S43 is used to: use the classifier to process each of the multiple optimized learning resource matching classification feature matrices according to the following formula to obtain the classification result;

[0067] Wherein, the formula is: ,in represents the projection of the optimized learning resource matching classification feature matrix into a vector, is the weight matrix, represents the bias vector, represents the normalized exponential function, Indicates the classification result.

[0068] Then, according to the classification result, the learning resources suitable for the student are visualized. In this way, by visually displaying the learning resources suitable for the student, the student can intuitively understand the content of the resources recommended by the system, thereby improving the student's understanding and acceptance of the learning resources.

[0069] In summary, the Internet-based learning resource sharing management method according to the embodiment of the present application is explained, which uses multiple learning resource videos of a subject that the students to be analyzed have watched in the past, the knowledge point evaluation data of the subject to be analyzed by the students, and multiple learning resource videos to be recommended related to the subject as input data, and convolutionally encodes these input data respectively through deep learning technology. That is, the input data is processed by deep learning technology. In this way, a more accurate feature representation can be obtained. These feature representations are then analyzed to obtain a classification result for indicating whether the text description of each learning resource to be recommended is suitable for the student. In this way, according to the classification result, the matching of educational resources can be adjusted more accurately, so as to share learning resources suitable for their current learning stage with students, so that the learning effect is better.

[0070] Exemplary Devices

[0071] Figure 7 FIG. 1 is a schematic block diagram of an Internet-based learning resource sharing and management device according to an embodiment of the present application. Figure 7 As shown, according to the embodiment of the present application, the Internet-based learning resource sharing management device 100 includes: a learning subject data collection module 110, which is used to collect multiple learning resource videos of a subject that the student to be analyzed has watched in the past, the knowledge point evaluation data of the subject by the student to be analyzed, and a plurality of learning resource videos related to the subject to be recommended in the resource sharing platform; a learning subject feature extraction module 120, which is used to extract and analyze the multiple learning resource videos of a subject that the student to be analyzed has watched in the past, the knowledge point evaluation data of the subject by the student to be analyzed, and the plurality of learning resource videos related to the subject to be recommended to obtain a student learning demand feature vector and a plurality of learning resource feature vectors; a learning subject feature fusion module 130, which is used to use a joint encoder to combine the student learning demand feature vector and each of the multiple learning resource feature vectors to obtain a plurality of learning resource matching feature matrices; a learning subject feature classification result generation module 140, which is used to obtain a classification result based on the multiple learning resource matching feature matrices, and the classification result is used to indicate whether the text description of each learning resource to be recommended is suitable for the student.

[0072] As described above, the Internet-based learning resource sharing management device 100 according to the embodiment of the present application can be implemented in various wireless terminals, such as a server for Internet-based learning resource sharing management. In one example, the Internet-based learning resource sharing management device 100 according to the embodiment of the present application can be integrated into the wireless terminal as a software module and / or a hardware module. For example, the Internet-based learning resource sharing management device 100 can be a software module in the operating system of the wireless terminal, or can be an application developed for the wireless terminal; of course, the Internet-based learning resource sharing management device 100 can also be one of the many hardware modules of the wireless terminal.

[0073] Alternatively, in another example, the Internet-based learning resource sharing management apparatus 100 and the wireless terminal may also be separate devices, and the Internet-based learning resource sharing management apparatus 100 may be connected to the wireless terminal via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

[0074] Here, those skilled in the art will appreciate that the specific functions and operations of each step in the above-mentioned Internet-based learning resource sharing management device have been described in the above reference. Figure 1 The description of the Internet-based learning resource sharing management method has been introduced in detail, and therefore, its repeated description will be omitted.

[0075] Further, Figure 8 FIG. 1 is a block diagram of an electronic device according to an embodiment of the present application. Figure 8 As shown, the electronic device includes a processor 801, a communication interface 802, a memory 803 and a communication bus 804. The processor 801, the communication interface 802, and the memory 803 communicate with each other through the communication bus 804. The processor 801, the communication interface 802, the memory 803 and other components can also communicate through a network connection. The present disclosure does not limit the type and function of the network. It should be noted that Figure 8 The components of the electronic device shown are merely exemplary and non-limiting. The electronic device may also have other components according to actual application requirements.

[0076] For example, the communication bus 804 may be a Peripheral Component Interconnect (PCI) bus or an Extended Industrial Standard Architecture (EISA) bus, etc. The communication bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0077] For example, the communication interface 802 is used to implement communication between the electronic device and other devices.

[0078] For example, the processor 801 and the memory 803 may be arranged on a server side (or a cloud side).

[0079] For example, the processor 801 can control other components in the electronic device to perform the desired functions. The processor 801 can be a device with data processing capabilities and / or program execution capabilities, such as a central processing unit (CPU), a network processor (NP), a tensor processing unit (TPU) or a graphics processing unit (GPU); it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The central processing unit (CPU) can be an X86 or ARM architecture, etc.

[0080] For example, the memory 803 may include any combination of one or more computer program products, and the computer program product may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and / or cache memory (cache), etc. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, erasable programmable read-only memory (EPROM), portable compact disk read-only memory (CD-ROM), USB memory, flash memory, etc. One or more computer-readable instructions may be stored on the computer-readable storage medium, and the processor 801 may execute the computer-readable instructions to implement various functions of the electronic device. Various applications and various data may also be stored in the storage medium.

[0081] For example, in some embodiments, the electronic device may further include an image acquisition component. The image acquisition component is used to obtain an input image of the object. The memory 803 is also used to store the input image.

[0082] For example, the image acquisition component may be a camera of a smartphone, a camera of a tablet computer, a camera of a personal computer, a lens of a digital camera, or even a webcam.

[0083] For example, the input image may be an original image directly acquired by the image acquisition component, or may be an image obtained after preprocessing the original image. Preprocessing may eliminate irrelevant information or noise information in the original image so as to better process the input image. Preprocessing may include, for example, image augmentation (DataAugment), image scaling, gamma correction, image enhancement, or noise reduction filtering on the original image.

[0084] Further, Fig. 9FIG. 1 is a block diagram of a non-transitory computer-readable storage medium according to an embodiment of the present application. Fig. 9 As shown, one or more computer readable instructions 901 may be non-transitorily stored on storage medium 900 .

[0085] For example, the storage medium 900 may be applied to the above-mentioned electronic device. For example, the storage medium 900 may include the memory 803 in the electronic device.

[0086] For example, the description of the storage medium 900 may refer to the description of the memory in the embodiment of the electronic device, and the repeated parts will not be repeated.

[0087] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0088] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, or it can be an electrical, mechanical or other form of connection.

[0089] The above is only a preferred embodiment of the present application and does not limit the present application in any form. Any technician familiar with the profession may use the above disclosed technical content to change or modify it into an equivalent embodiment with equivalent changes. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present application without departing from the content of the technical solution of the present application still falls within the protection scope of the technical solution of the present application.

Claims

1. A learning resource sharing management method based on the Internet, characterized in that: include: Collect multiple learning resource videos of a subject that the students to be analyzed have watched in the past, knowledge point evaluation data of the students to be analyzed on the subject, and multiple learning resource videos to be recommended related to the subject on the resource sharing platform; Extract and analyze the features of the multiple learning resource videos of a subject that the student to be analyzed has watched in the past, the knowledge point evaluation data of the student to be analyzed on the subject, and the multiple learning resource videos to be recommended related to the subject to obtain a student learning demand feature vector and multiple learning resource feature vectors; Using a joint encoder, the student learning demand feature vector and each learning resource feature vector in the multiple learning resource feature vectors are combined to obtain multiple learning resource matching feature matrices; Based on the matching feature matrix of the multiple learning resources, a classification result is obtained, wherein the classification result is used to indicate whether the text description of each learning resource to be recommended is suitable for the student; The classification result is obtained based on the matching feature matrix of the multiple learning resources, and the classification result is used to indicate whether the text description of each learning resource to be recommended is suitable for the student, including: Passing the multiple learning resource matching feature matrices through a learning resource matching feature extraction module based on a convolutional neural network model to obtain multiple learning resource matching classification feature matrices; Performing expected activation of feature representation on each of the plurality of learning resource matching classification feature matrices to obtain a plurality of optimized learning resource matching classification feature matrices; Passing the plurality of optimized learning resource matching classification feature matrices through classifiers respectively to obtain the classification results; Based on the classification results, visually display the learning resources suitable for the student; The method of performing expected activation of feature representation on each of the plurality of learning resource matching classification feature matrices to obtain a plurality of optimized learning resource matching classification feature matrices includes: Calculate an exponential function value raised to the power of the eigenvalue of the predetermined position of the learning resource matching classification feature matrix and then subtract one to obtain an exponentially adjusted eigenvalue; Calculating the index-adjusted eigenvalue divided by the eigenvalue of a predetermined position of the learning resource matching classification feature matrix, and calculating an absolute value of the result to obtain a normalized index-adjusted eigenvalue ratio; The normalized index adjusted eigenvalue ratio is activated by a softmax function and then multiplied by the eigenvalue of the predetermined position of the learning resource matching classification feature matrix to obtain the eigenvalue of the predetermined position of the optimized learning resource matching classification feature matrix.

2. The Internet-based learning resource sharing and management method according to claim 1, characterized in that: Feature extraction and analysis are performed on a plurality of learning resource videos of a subject that the student to be analyzed has watched in the past, the knowledge point evaluation data of the subject by the student to be analyzed, and the plurality of learning resource videos to be recommended related to the subject to obtain a student learning demand feature vector and a plurality of learning resource feature vectors, including: Performing feature coding on a plurality of learning resource videos of a certain subject that the student to be analyzed has watched in the past to obtain a global feature vector of learning progress; Performing feature coding on the knowledge point evaluation data of the subject of the student to be analyzed to obtain a student evaluation feature vector; Performing feature fusion on the learning progress global feature vector and the student assessment feature vector to obtain the student learning demand feature vector; Feature encoding is performed on the multiple learning resource videos to be recommended that are related to the subject to obtain the multiple learning resource feature vectors.

3. The Internet-based learning resource sharing and management method according to claim 2, characterized in that: Feature encoding is performed on multiple learning resource videos of a certain subject that the student to be analyzed has watched in the past to obtain a global feature vector of learning progress, including: Obtaining key frames of each learning resource video among multiple learning resource videos of a subject that the student to be analyzed has watched in the past to obtain multiple key frames of each video; Passing the multiple key frames of each video through a learning progress feature extractor based on a deep and shallow feature fusion module to obtain multiple learning progress feature matrices of each video; Arranging the multiple learning progress feature matrices of each video into learning progress time series input tensors of each video according to the time dimension respectively, and then passing them through a learning progress time series feature extraction module including a three-dimensional convolution kernel to obtain a learning progress time series feature graph of each video; Cascading the learning progress time series feature graphs of the respective videos to obtain a learning progress global feature graph; A pooling operation is performed on the learning progress global feature map to obtain the learning progress global feature vector.

4. The Internet-based learning resource sharing and management method according to claim 3 is characterized in that: The multiple key frames of each video are respectively passed through a learning progress feature extractor based on a deep and shallow feature fusion module to obtain multiple learning progress feature matrices of each video, which are used for: Extracting a shallow feature matrix from the i-th layer of the learning progress feature extractor, wherein the i-th layer is the first to sixth layers of the learning progress feature extractor; Extracting a deep feature matrix from the jth layer of the learning progress feature extractor, wherein a ratio between the jth layer and the ith layer is greater than or equal to 5; as well as The shallow feature matrix and the deep feature matrix are fused using the deep-shallow feature fusion module to obtain each learning progress feature matrix in the multiple learning progress feature matrices of each video.

5. The Internet-based learning resource sharing and management method according to claim 4 is characterized in that: The knowledge point evaluation data of the subject to be analyzed by the student is feature-encoded to obtain a student evaluation feature vector, including: Embedding the knowledge point evaluation data of the subject by the student to be analyzed to obtain a student evaluation embedding vector; The student evaluation embedding vector is passed through a student evaluation feature extraction module based on a one-dimensional convolutional neural network model to obtain the student evaluation feature vector.

6. The Internet-based learning resource sharing and management method according to claim 5, characterized in that: Performing feature coding on the plurality of learning resource videos to be recommended related to the subject to obtain the plurality of learning resource feature vectors includes: Acquire key frames of each of the multiple learning resource videos to be recommended related to the subject to obtain multiple learning resource key frames of each video; Arranging the multiple learning resource key frames of each video into learning resource time series input tensors of each video according to the time dimension respectively, and then passing through a learning resource time series feature extraction module based on a spatial attention mechanism to obtain multiple learning resource time series feature graphs; The multiple learning resource time series feature graphs are respectively subjected to pooling operations to obtain the multiple learning resource feature vectors.

7. A learning resource sharing management device based on the Internet, characterized in that ,include: The learning subject data collection module is used to collect multiple learning resource videos of a subject that the students to be analyzed have watched in the past, the knowledge point evaluation data of the students to be analyzed on the subject, and multiple learning resource videos to be recommended related to the subject in the resource sharing platform; A learning subject feature extraction module is used to extract and analyze features of a plurality of learning resource videos of a subject that the student to be analyzed has watched in the past, the knowledge point evaluation data of the student to be analyzed on the subject, and the plurality of learning resource videos to be recommended related to the subject to obtain a student learning demand feature vector and a plurality of learning resource feature vectors; A learning subject feature fusion module, used for combining the student learning demand feature vector with each of the multiple learning resource feature vectors using a joint encoder to obtain multiple learning resource matching feature matrices; A learning subject feature classification result generating module, used for matching the feature matrix based on the plurality of learning resources to obtain a classification result, wherein the classification result is used to indicate whether the text description of each learning resource to be recommended is suitable for the student; The classification result is obtained based on the matching feature matrix of the multiple learning resources, and the classification result is used to indicate whether the text description of each learning resource to be recommended is suitable for the student, including: Passing the multiple learning resource matching feature matrices through a learning resource matching feature extraction module based on a convolutional neural network model to obtain multiple learning resource matching classification feature matrices; Performing expected activation of feature representation on each of the plurality of learning resource matching classification feature matrices to obtain a plurality of optimized learning resource matching classification feature matrices; Passing the plurality of optimized learning resource matching classification feature matrices through classifiers respectively to obtain the classification results; Based on the classification results, visually display the learning resources suitable for the student; The method of performing expected activation of feature representation on each of the plurality of learning resource matching classification feature matrices to obtain a plurality of optimized learning resource matching classification feature matrices includes: Calculate an exponential function value raised to the power of the eigenvalue of the predetermined position of the learning resource matching classification feature matrix and then subtract one to obtain an exponentially adjusted eigenvalue; Calculating the index-adjusted eigenvalue divided by the eigenvalue of a predetermined position of the learning resource matching classification feature matrix, and calculating an absolute value of the result to obtain a normalized index-adjusted eigenvalue ratio; The normalized index adjusted eigenvalue ratio is activated by a softmax function and then multiplied by the eigenvalue of the predetermined position of the learning resource matching classification feature matrix to obtain the eigenvalue of the predetermined position of the optimized learning resource matching classification feature matrix.

8. An electronic device comprising: processor; as well as A memory, in which computer program instructions are stored, and when the computer program instructions are executed by the processor, the processor executes the Internet-based learning resource sharing and management method as described in any one of claims 1-6.

Images