AI-based automatic auxiliary learning system

Abstract

The invention relates to the technical field of AI learning, and discloses an AI-based automatic auxiliary learning system, which comprises a data transmission module, a learning identification module, an identification matching module, a learning prediction module, a learning push module and a personal learning library, the data transmission module is used for receiving the current time, the number of the learner and the learning content needing to be learned by the learner; the learning identification module is used for receiving information transmitted by the data transmission module and inputting learning content into the identification matching module; the identification matching module is used for identifying the learning content and associating the identified learning content with the content in the personal learning library; the learning prediction module is used for predicting the predicted training time of the current learning content; in the learning process of the learner, whether the learning content is learned or not can be judged, so that the learning time is adjusted in real time, the time is reasonably arranged, and the learner can learn efficiently.

Description

Technical Field

[0001] The present invention belongs to the technical field of AI-assisted learning, and in particular, relates to an AI-based automatic assisted learning system. Background Art

[0002] With the rapid development of artificial intelligence technology, the field of intelligent education has also ushered in changes. Traditional learning models usually rely on fixed teaching materials and a single learning path, which cannot meet personalized and diversified learning needs. Especially in the information society, students, working people and all kinds of learners face multiple challenges such as massive information, fragmented knowledge and time management in the learning process. Therefore, how to effectively improve learning efficiency and provide personalized and precise learning support has become a problem that needs to be solved urgently in the field of modern education.

[0003] At present, many traditional learning systems focus more on content presentation or simple test assessment, and have not yet formed a complete intelligent learning system. Most existing intelligent learning platforms and tools lack the ability to automatically generate learning plans, accurately match content, and provide personalized feedback. At the same time, with the development of big data and AI technology, how to accurately identify learners' needs in huge data and dynamically adjust according to learning progress has become the key to promoting the development of intelligent education.

[0004] In response to these problems, it is particularly important to build an automated and intelligent learning system by combining artificial intelligence technology and big data analysis. Through the application of AI technology, multiple modules can work together to accurately identify learners' needs, predict learning time, and provide effective learning suggestions and push reminders based on learners' personal circumstances. This system can achieve intelligent matching and push of content based on learners' real-time progress, knowledge mastery, and learning preferences, thereby improving learning efficiency, optimizing learning paths, and helping learners better conduct personalized learning.

[0005] Therefore, the automatic assisted learning system based on AI technology came into being. Through the collaboration of multiple modules, it can provide comprehensive intelligent support in terms of accurate identification of learning content, prediction of learning progress, personalized reminders, etc., and promote further innovation and development of educational technology.

[0006] In view of this, the present invention is proposed. Summary of the invention

[0007] In order to solve the above technical problems, the basic concept of the technical solution adopted by the present invention is: An AI-based automatic assisted learning system, comprising a data transmission module, a learning identification module, an identification and matching module, a learning prediction module, a learning push module and a personal learning library; the data transmission module is used to receive the current time, the learner number and the learning content that the learner needs to learn, and then transmit them to the learning identification module; the learning identification module is used to receive the information transmitted by the data transmission module, and input the learning content into the identification and matching module; Among them, the learning content includes identifying pictures, videos, and texts; the recognition and matching module is used to identify the learning content, and associate the identified learning content with the content in the personal learning library. When it is necessary to learn and train the knowledge content, the learning content is transmitted to the learning prediction module. The learning prediction module is used to predict the expected training time of the current learning content, and transmit the prediction result to the learning push module. The learning push module is used to remind the learners, and the reminder content includes the expected training time.

[0008] As a preferred embodiment of the present invention, the process of identifying the learning content of the learning by the learning recognition module is specifically expressed as follows: Step 1: When the learning content is a video or a picture, the video is extracted frame by frame through video recognition and picture recognition, assuming that the total number of video frames is S, and face recognition is performed on the extracted pictures to obtain the faces contained in the pictures and calculate the average pixel of any one of the faces, which will be stored, wherein is the picture information and is the number of faces contained in the picture; Step 2: When the learning content is text, the text is extracted word by word through text recognition, assuming that there are W texts in total, and any one of the texts is Ew, the learning recognition module will store and obtain the learning content corresponding to each face; then the learning recognition module will be matched with the personal learning library, wherein the personal learning library is the content that the learner has learned through the face recognition module or the text recognition module.

[0009] As a preferred embodiment of the present invention, the specific matching process of the recognition and matching module is: S1, obtaining the learning content recognized by the learning recognition module, by obtaining the learning content corresponding to the current face, and then by establishing a face recognition model, establishing a face analysis model, comparing the established face recognition model with the matched face, and establishing a face similarity model; S2, obtaining the video or text corresponding to the recognized learning content, by obtaining the learning content corresponding to the current text, and then establishing a text analysis model, comparing the established text analysis model with the matched text, and establishing a text similarity model; S3, obtaining the similarity value between the face and the text, and calculating the similarity between the learning content and the content in the personal learning library; S4, obtaining the expected training time corresponding to the similar learning content.

[0010] As a preferred embodiment of the present invention, the learning prediction module specifically adopts the following prediction process: the learning time corresponding to the learning content is converted through a time conversion model, wherein the learning conversion model is obtained through big data; the learning conversion model training process includes: collecting the learning content learned N times in the personal learning library and its corresponding learning time TZ as training data, and obtaining a learning time deep model through training data, and the formula of the learning time deep model is: FH=TZ+γ(1+Z); wherein, through multiple calculations and continuous updating of the independent variables in the learning time deep model formula, the final calculation result FH is obtained; wherein, γ is a calculation parameter.

[0011] As a preferred embodiment of the present invention, after the learning push module receives the estimated training time transmitted by the learning prediction module, the learning push module then reminds according to the estimated training time. The specific reminder process includes: S1, marking the estimated training time as FH, and obtaining the starting point corresponding to the current time, the rest time XS corresponding to the first rest point, the rest time Ds corresponding to the second rest point, and the rest time Dh corresponding to the third rest point through the learning push module; wherein the first rest point, the second rest point, and the third rest point are respectively 20%, 30%, and 80% of the total training time; S2, obtaining the first time W1 corresponding to the starting point, the second time W2 corresponding to the first rest point, the third time W3 corresponding to the second rest point, and the fourth time W4 corresponding to the third rest point; S3, The learning push module marks the training end time as the first time W1 when the training duration is greater than or equal to the starting point and less than the expected training duration corresponding to the first rest point; when the training duration is greater than or equal to the first rest point and less than the expected training duration corresponding to the second rest point, the training end time is marked as the second time W2; when the learning duration is greater than or equal to the second rest point and less than the expected training duration corresponding to the third rest point, the training end time is marked as the third time W3; when the learning duration is greater than or equal to the expected training duration corresponding to the third rest point, the training end time is marked as the fourth time W4; S4. During the training process, the learning push module reminds the trainees to compare the current training time with the corresponding expected training time, and pushes the learning end signal when the training time is greater than or equal to the expected training time.

[0012] As a preferred embodiment of the present invention, during the actual training process, the estimated training time is adjusted according to the actual training situation, wherein the time interval for adaptively adjusting the estimated training time according to the actual training situation is H, and the training time is adjusted within the interval of H.

[0013] As a preferred embodiment of the present invention, the learning push module is also used to adjust the learning content according to the actual learning situation of the learner. The specific adjustment process includes: S1, obtaining the learning content actually needed by the learner, wherein the required learning content is the content that the learner obtains from the personal learning library and needs to be improved; S2, obtaining the learning content actually needed by the learner for learning, and when the learning content is successfully trained within the interval time H, the learning adjustment probability of the learning content is calculated according to the formula, wherein when the learning content is successfully trained, the learning content is marked as A1, and when the learning content is not successfully trained, the learning content is marked as A2; wherein, when the content that needs to be improved is not obtained in the personal learning library, the actual needs of the learner are adjusted. The learning content is pushed; S3, obtaining the learning adjustment probability Wz of the most recent learning content required by the learner; S4, transmitting the learning adjustment probability Wz of the most recent required learning content, whether the learning content is successfully trained within the interval time of H, and the learning time required for successful learning to the learning push module, and when the learning content is successfully trained within the interval time of H, the success probability is calculated; S5, the learning adjustment probability of the learner is calculated according to the formula, wherein, when the learning adjustment probability of the learner is compared with the learning adjustment probability Wz of the most recent required learning content, if the learning adjustment probability of the learner is greater than the learning adjustment probability Wz of the most recent required learning content, the learning content of the learner is adjusted.

[0014] As a preferred embodiment of the present invention, the learning push module adjusts the learning content according to the learning adjustment probability of the learner, including the following steps: when the learning adjustment probability of the learner is greater than the learning adjustment probability Wz of the most recently required learning content, the estimated training time corresponding to the learner's last learning content is obtained; if the actual training time corresponding to the last learning content exceeds or equals the estimated training time, the last learning content is marked as the learning content that needs to be improved; if the actual training time corresponding to the last learning content is less than the estimated training time, the last learning content is marked as the learning content that can be improved; content; if the estimated training time corresponding to the last learning content is successfully trained within the interval time of H, the last learning content is marked as the learning content that needs to be improved; S1, remove the learning content that was most recently marked as needing to be improved from the above marked content; S2, transmit the learning adjustment probability Wz of the remaining most recent required learning content, whether the learning content is successfully trained within the interval time of H, and the learning time required for successful learning to the learning push module; S3, re-match the learning content required by the learner with the corresponding learning content in the personal learning library, and adjust the learning content.

[0015] As a preferred embodiment of the present invention, it also includes: when the learning adjustment probability of the learning content required by the learner has not changed, the learning content is pushed again for adjustment according to steps S1-S3 until all the required learning content of the current learner is pushed, and the required learning content is no longer adjusted.

[0016] As a preferred embodiment of the present invention, it also includes: obtaining the adjustment probability of all learning contents, setting a prediction time, and recalculating the adjustment probability when the time T is exceeded, wherein when the adjustment probability of any learning content is less than the threshold of the adjustment probability, the learning content is re-pushed, and the re-pushing of learning content includes: obtaining the average adjustment probability of the current learner, wherein the average adjustment probability is obtained by obtaining the adjustment probabilities of several learning contents and then calculating the average value, matching the learning content whose adjustment probability does not meet the standard and the corresponding adjustment probability with the re-pushing learning module, and the re-pushing learning module obtains the learning content whose adjustment probability does not meet the standard and the corresponding adjustment probability, and re-pushes them.

[0017] Compared with the prior art, the present invention has the following beneficial effects: 1. Build an automatic assisted learning system based on AI technology, which includes multiple modules working together.

[0018] 2. Accurate recognition and matching of learning content is achieved through data transmission module, learning recognition module and recognition and matching module.

[0019] 3. Introduce a learning prediction module to predict the estimated training time of the current learning content.

[0020] 4. Use the learning push module to remind individuals and store the pushed content in the personal learning library.

[0021] The specific implementation modes of the present invention are further described in detail below in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In the attached picture: Figure 1 A framework diagram of an AI-based automatic assisted learning system. DETAILED DESCRIPTION

[0023] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. The following embodiments are used to illustrate the present invention.

[0024] Embodiment 1: like Figure 1As shown, an AI-based automatic assisted learning system, 1. includes a data transmission module, a learning identification module, an identification and matching module, a learning prediction module, a learning push module and a personal learning library; the data transmission module is used to receive the current time, the learner number and the learning content that the learner needs to learn, and then transmit them to the learning identification module; the learning identification module is used to receive the information transmitted by the data transmission module, and input the learning content into the identification and matching module; Among them, the learning content includes identifying pictures, videos, and texts; the recognition and matching module is used to identify the learning content, and associate the identified learning content with the content in the personal learning library. When it is necessary to learn and train the knowledge content, the learning content is transmitted to the learning prediction module. The learning prediction module is used to predict the expected training time of the current learning content, and transmit the prediction result to the learning push module. The learning push module is used to remind the learners, and the reminder content includes the expected training time.

[0025] Further, the process of the learning recognition module identifying the learning content of the learning is specifically represented as follows: Step 1: When the learning content is a video or a picture, the video is extracted frame by frame through video recognition and picture recognition, assuming that the total number of video frames is S, and the extracted picture is subjected to face recognition, the faces contained in the picture are obtained and the average pixel of any one of the faces is calculated, and the faces are stored, wherein is the picture information and is the number of faces contained in the picture; Step 2: When the learning content is text, the text is extracted word by word through text recognition, assuming that there are a total of W texts, and any one of the texts is Ew, the learning recognition module will store and obtain the learning content corresponding to each face; then the learning recognition module will match with the personal learning library, wherein the personal learning library is the content that the learner has learned through the face recognition module or the text recognition module; It also includes: when the learning adjustment probability of the learning content required by the learner has not changed, the learning content is pushed again for adjustment according to steps S1-S3 until all the required learning content of the current learner is pushed, and the required learning content is no longer adjusted.

[0026] Furthermore, the specific matching process of the recognition and matching module is as follows: S1. Obtain the learning content after recognition by the learning recognition module, by obtaining the learning content corresponding to the current face, and then by establishing a face recognition model, establish a face analysis model, compare the established face recognition model with the matched face, and establish a face similarity model; S2. Obtain the video or text corresponding to the recognized learning content, by obtaining the learning content corresponding to the current text, and then establish a text analysis model, compare the established text analysis model with the matched text, and establish a text similarity model; S3. Obtain the similarity value between the face and the text, and calculate the similarity between the learning content and the content in the personal learning library; S4. Obtain the estimated training time corresponding to the similar learning content.

[0027] Furthermore, the learning prediction module specifically adopts the following prediction process: the learning time corresponding to the learning content is converted through a time conversion model, wherein the learning conversion model is obtained through big data; the learning conversion model training process includes: collecting the learning content learned N times in the personal learning library and its corresponding learning time TZ as training data, and obtaining a learning time deep model through training data. The formula of the learning time deep model is: FH=TZ+γ(1+Z); wherein, through multiple calculations and continuous updating of the independent variables in the learning time deep model formula, the final calculation result FH is obtained; wherein, γ is a calculation parameter.

[0028] Furthermore, after the learning push module receives the estimated training time transmitted by the learning prediction module, the learning push module then reminds according to the estimated training time. The specific reminder process includes: S1, marking the estimated training time as FH, and obtaining the starting point corresponding to the current time, the rest time XS corresponding to the first rest point, the rest time Ds corresponding to the second rest point, and the rest time Dh corresponding to the third rest point through the learning push module; wherein the first rest point, the second rest point, and the third rest point are respectively 20%, 30%, and 80% of the total training time; S2, obtaining the first time W1 corresponding to the starting point, the second time W2 corresponding to the first rest point, the third time W3 corresponding to the second rest point, and the fourth time W4 corresponding to the third rest point; S3 ... When the training duration is greater than or equal to the starting point and less than the expected training duration corresponding to the first rest point, the training end time is marked as the first time W1; when the training duration is greater than or equal to the first rest point and less than the expected training duration corresponding to the second rest point, the training end time is marked as the second time W2; when the training duration is greater than or equal to the second rest point and less than the expected training duration corresponding to the third rest point, the training end time is marked as the third time W3; when the training duration is greater than or equal to the expected training duration corresponding to the third rest point, the training end time is marked as the fourth time W4; S4. During the training process, the learning push module reminds the training personnel to compare the current training time with the corresponding expected training time, and when the training time is greater than or equal to the expected training time, the learning end signal is pushed.

[0029] Furthermore, during the actual training process, the estimated training time is adjusted according to the actual training situation, wherein the time interval for adaptively adjusting the estimated training time according to the actual training situation is H, and the training time is adjusted within the interval of H; it also includes: obtaining the adjustment probability of all learning contents, setting a predicted duration, and recalculating the adjustment probability when the time T is exceeded, wherein when the adjustment probability of any learning content is less than the threshold of the adjustment probability, the learning content is re-pushed, and the re-pushing of the learning content includes: obtaining the average adjustment probability of the current learner, wherein the average adjustment probability is obtained by obtaining the adjustment probabilities of several learning contents, and then calculating the average value, matching the learning content whose adjustment probability does not meet the standard and the corresponding adjustment probability with the re-pushing learning module, and the re-pushing learning module obtains the learning content whose adjustment probability does not meet the standard and the corresponding adjustment probability, and re-pushes them.

[0030] Furthermore, the learning push module is also used to adjust the learning content according to the actual learning situation of the learner. The specific adjustment process includes: S1, obtaining the learning content actually needed by the learner, wherein the required learning content is the content that the learner obtains from the personal learning library and needs to be improved; S2, obtaining the learning content actually needed by the learner for learning, and when the learning content is successfully trained within the interval time H, the learning adjustment probability of the learning content is calculated according to the formula, wherein when the learning content is successfully trained, the learning content is marked as A1, and when the learning content is not successfully trained, the learning content is marked as A2; wherein, when the content that needs to be improved is not obtained in the personal learning library, the learning content actually needed by the learner is adjusted. Push; S3, obtain the learning adjustment probability Wz of the most recent learning content required by the learner; S4, transmit the learning adjustment probability Wz of the most recent learning content required, whether the learning content is successfully trained within the interval time of H, and the learning time required for successful learning to the learning push module, and when the learning content is successfully trained within the interval time of H, calculate the success probability; S5, calculate the learning adjustment probability of the learner according to the formula, wherein, when the learning adjustment probability of the learner is compared with the learning adjustment probability Wz of the most recent learning content required, if the learning adjustment probability of the learner is greater than the learning adjustment probability Wz of the most recent learning content required, then adjust the learning content of the learner; Furthermore, the learning push module adjusts the learning content according to the learning adjustment probability of the learner, including the following steps: when the learning adjustment probability of the learner is greater than the learning adjustment probability Wz of the most recently required learning content, the estimated training time corresponding to the learner's last learning content is obtained, and if the actual training time corresponding to the last learning content exceeds or equals the estimated training time, the last learning content is marked as learning content that needs to be improved; if the actual training time corresponding to the last learning content is less than the estimated training time, the last learning content is marked as learning content that can be performed; if the last If the estimated training time corresponding to the last learning content is successfully trained within the interval time of H, the last learning content will be marked as the learning content that needs to be improved; S1. The most recent learning content marked as the learning content that needs to be improved will be eliminated from the marked content; S2. The learning adjustment probability Wz of the remaining most recent required learning content, whether the learning content is successfully trained within the interval time of H, and the learning time required for successful learning are transmitted to the learning push module; S3. The learning content required by the learner is re-matched with the corresponding learning content in the personal learning library, and the learning content is adjusted.

Claims

1. An AI-based automatic assisted learning system, characterized in that: It includes a data transmission module, a learning identification module, an identification and matching module, a learning prediction module, a learning push module and a personal learning library; the data transmission module is used to receive the current time, the learner number and the learning content that the learner needs to learn, and then transmit it to the learning identification module; The learning and recognition module is used to receive information transmitted by the data transmission module and input the learning content into the recognition and matching module; Among them, the learning content includes identifying pictures, videos, and texts; the recognition and matching module is used to identify the learning content, and associate the identified learning content with the content in the personal learning library. When it is necessary to learn and train the knowledge content, the learning content is transmitted to the learning prediction module. The learning prediction module is used to predict the expected training time of the current learning content, and transmit the prediction result to the learning push module. The learning push module is used to remind the learners, and the reminder content includes the expected training time.

2. The AI-based automatic assisted learning system according to claim 1, characterized in that: The process of the learning recognition module identifying the learning content of the learning is specifically expressed as follows: Step 1: When the learning content is a video or a picture, the video is extracted frame by frame through video recognition and picture recognition. It is assumed that the total number of video frames is S, and face recognition is performed on the extracted pictures to obtain the faces contained in the pictures and calculate the average pixel of any one of the faces, which will be stored, where is the picture information and is the number of faces contained in the picture; Step 2: When the learning content is text, the text is extracted word by word through text recognition. It is assumed that there are W texts in total, and any one of the texts is Ew. The learning recognition module will store and obtain the learning content corresponding to each face; then the learning recognition module will be matched with the personal learning library, where the personal learning library is the content that the learner has learned through the face recognition module or the text recognition module.

3. The AI-based automatic assisted learning system according to claim 2, characterized in that: The specific matching process of the recognition and matching module is as follows: S1, obtaining the learning content recognized by the learning recognition module, by obtaining the learning content corresponding to the current face, and then by establishing a face recognition model, establishing a face analysis model, comparing the established face recognition model with the matched face, and establishing a face similarity model; S2, obtaining the video or text corresponding to the recognized learning content, by obtaining the learning content corresponding to the current text, and then establishing a text analysis model, comparing the established text analysis model with the matched text, and establishing a text similarity model; S3, obtaining the similarity value between the face and the text, and calculating the similarity between the learning content and the content in the personal learning library; S4, obtaining the estimated training time corresponding to the similar learning content.

4. The AI-based automatic assisted learning system according to claim 1, characterized in that: The learning prediction module specifically adopts the following prediction process: the learning time corresponding to the learning content is converted through a time conversion model, wherein the learning conversion model is obtained through big data; the learning conversion model training process includes: collecting the learning content learned N times in the personal learning library and its corresponding learning time TZ as training data, and obtaining a learning time deep model through training data. The formula of the learning time deep model is: FH=TZ+γ(1+Z); wherein, through multiple calculations and continuous updating of the independent variables in the learning time deep model formula, the final calculation result FH is obtained; wherein, γ is a calculation parameter.

5. The AI-based automatic assisted learning system according to claim 1, characterized in that: After the learning push module receives the estimated training time transmitted by the learning prediction module, the learning push module then reminds according to the estimated training time. The specific reminder process includes: S1, marking the estimated training time as FH, and obtaining the starting point corresponding to the current time, the rest time XS corresponding to the first rest point, the rest time Ds corresponding to the second rest point, and the rest time Dh corresponding to the third rest point through the learning push module; wherein the first rest point, the second rest point, and the third rest point are 20%, 30%, and 80% of the total training time, respectively; S2, obtaining the first time W1 corresponding to the starting point, the second time W2 corresponding to the first rest point, the third time W3 corresponding to the second rest point, and the fourth time W4 corresponding to the third rest point; S3, the learning push module will When the training duration is greater than or equal to the starting point and less than the expected training duration corresponding to the first rest point, the training end time is marked as the first time W1; when the training duration is greater than or equal to the first rest point and less than the expected training duration corresponding to the second rest point, the training end time is marked as the second time W2, when the learning duration is greater than or equal to the second rest point and less than the expected training duration corresponding to the third rest point, the training end time is marked as the third time W3, when the learning duration is greater than or equal to the expected training duration corresponding to the third rest point, the training end time is marked as the fourth time W4; S4. During the training process, the learning push module reminds the training personnel to compare the current training time with the corresponding expected training time, and when the training time is greater than or equal to the expected training time, the learning end signal is pushed.

6. The AI-based automatic assisted learning system according to claim 5, characterized in that: During the actual training process, the estimated training time is adjusted according to the actual training situation, wherein the time interval for adaptively adjusting the estimated training time according to the actual training situation is H, and the training time is adjusted within the interval of H.

7. The AI-based automatic assisted learning system according to claim 1, characterized in that: The learning push module is also used to adjust the learning content according to the actual learning situation of the learner. The specific adjustment process includes: S1, obtaining the learning content actually needed by the learner, wherein the required learning content is the content that the learner obtains from the personal learning library and needs to be improved; S2, obtaining the learning content actually needed by the learner for learning, and when the learning content is successfully trained within the interval time H, the learning adjustment probability of the learning content is calculated according to the formula, wherein when the learning content is successfully trained, the learning content is marked as A1, and when the learning content is not successfully trained, the learning content is marked as A2; wherein, when the content that needs to be improved is not obtained in the personal learning library, the learning content actually needed by the learner is pushed Send; S3, obtain the learning adjustment probability Wz of the most recent learning content required by the learner; S4, transmit the learning adjustment probability Wz of the most recent learning content required, whether the learning content is successfully trained within the interval time of H, and the learning time required for successful learning to the learning push module, when the learning content is successfully trained within the interval time of H, calculate the success probability; S5, calculate the learning adjustment probability of the learner according to the formula, wherein, when the learning adjustment probability of the learner is compared with the learning adjustment probability Wz of the most recent learning content required, if the learning adjustment probability of the learner is greater than the learning adjustment probability Wz of the most recent learning content required, the learning content of the learner is adjusted.

8. The AI-based automatic assisted learning system according to claim 7, characterized in that: The learning push module adjusts the learning content according to the learning adjustment probability of the learner, including the following steps: when the learning adjustment probability of the learner is greater than the learning adjustment probability Wz of the most recent required learning content, the estimated training time corresponding to the learner's last learning content is obtained, and if the actual training time corresponding to the last learning content exceeds or equals the estimated training time, the last learning content is marked as the learning content that needs to be improved; if the actual training time corresponding to the last learning content is less than the estimated training time, the last learning content is marked as the learning content that can be carried out; if the estimated training time corresponding to the last learning content is successfully trained within the interval time of H, the last learning content is marked as the learning content that needs to be improved; S1, remove the learning content that is most recently marked as needing to be improved from the above marked content; S2, transmit the learning adjustment probability Wz of the remaining most recently required learning content, whether the learning content is successfully trained within the interval time of H, and the learning time required for successful learning to the learning push module; S3, re-match the learning content required by the learner with the corresponding learning content in the personal learning library, and adjust the learning content.

9. The AI-based automatic assisted learning system according to claim 1, characterized in that: Also includes: When the learning adjustment probability of the learning content required by the learner has not changed, the learning content is pushed again for adjustment according to steps S1-S3 until all the required learning content of the current learner is pushed and no further adjustment is made to the required learning content.

10. The AI-based automatic assisted learning system according to claim 6, characterized in that: Also includes: Obtain the adjustment probability of all learning contents, set a prediction time, and recalculate the adjustment probability when the time T is exceeded. When the adjustment probability of any learning content is less than the threshold of the adjustment probability, re-push the learning content. The re-pushing of learning content includes: obtaining the average adjustment probability of the current learner, wherein the average adjustment probability is obtained by obtaining the adjustment probabilities of several learning contents and then calculating the average value. The learning contents whose adjustment probabilities do not meet the standards and the corresponding adjustment probabilities are matched with the re-pushing learning module. The re-pushing learning module obtains the learning contents whose adjustment probabilities do not meet the standards and the corresponding adjustment probabilities, and re-pushes them.

Images