Tablet computer learning and training system and method based on artificial intelligence

Abstract

The invention discloses a tablet computer learning training system and method based on artificial intelligence, and relates to the field of tablet computer application, and the method comprises the steps: inputting a preset learning material data set into a tablet computer local database; extracting texts in the learning material dataset and carrying out standardization processing to obtain a learning material text dataset, and carrying out named entity recognition through a first machine learning model; performing relation extraction on the named entity data set to obtain named entity relation classification, performing knowledge fusion based on a text feature semantic relation, adding an attribute tag and a relation tag, and establishing a learning material knowledge graph; recognizing the facial micro-expression image data of the user through a second machine learning model to obtain a user learning state recognition result, calculating distraction frequency and performing threshold judgment, and executing message prompt by the tablet computer based on a threshold judgment result; a learning training model is established, and an initial course, a derivative course, a derivative training question group and a personalized training question group are set for a user.

Description

Technical Field

[0001] The present invention relates to the technical field of tablet computer applications, and in particular to a tablet computer learning and training system and method based on artificial intelligence. Background Art

[0002] Thanks to the rapid development of electronic technology and industry, the tablet computer market has grown steadily and its penetration rate has continued to increase. In the education industry, the application of tablet computers as a new teaching aid has become a normal trend. Tablet computers can support multimedia learning, present teaching knowledge from multiple angles through text, images and videos, and can improve interactive operations, enhancing learners' understanding and memory. Tablet computers are lightweight and easy to carry around and use for learning in fragmented time. With the rise of artificial intelligence, tablet computers have provided new directions in learning and training.

[0003] At present, the Chinese invention patent with application number CN202311773999.9 discloses a method and system for implementing an education platform based on a tablet computer. The invention specifically includes: obtaining the capability information and indicator information of the tablet computer; collecting the keyword information of the user learning history of the learning application on the tablet computer; the education platform determines the keyword division granularity based on the capability information, indicator information and keyword information; classifies and associates all learning resources on the education platform according to the division granularity; obtains the similarity between the target user or target learning resource and the existing learning resource on the education platform; and recommends learning resources based on the similarity. The present invention can more accurately understand and organize the content of learning resources by classifying learning resources according to keywords; determines the division granularity according to the performance and storage capacity of the tablet computer, and flexibly adjusts it according to the specific situation. However, the invention adopts a fixed algorithm when integrating learning materials, lacks adaptability to different types of materials, relies only on keyword similarity for classification when classifying knowledge points, lacks the division and recommendation of learning materials from the perspective of attributes and relationships, and lacks personalization in arranging user courses and training questions, and lacks interaction with users, and fails to effectively judge the user's learning status and issue prompts to users based on the judgment results. Summary of the invention

[0004] The technical problem solved by the present invention is that the traditional method adopts a fixed algorithm when integrating learning material data using a tablet computer, which lacks adaptability to different types of materials. It fails to establish a data map from the perspective of the attributes and relationships of the named entities of the learning materials, and the degree of personalization of arranging user courses and training questions is insufficient. It also lacks interaction with users, fails to effectively judge the user's learning status and provide graded prompt messages based on the judgment results.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0006] A tablet computer learning and training method based on artificial intelligence, comprising:

[0007] Step S1, inputting preset video course data, textbook graphic data and training question data into the local database of the tablet computer to obtain a learning material data set;

[0008] Step S2, extracting text from the learning material data set and performing standardization processing to obtain a learning material text data set, and performing named entity recognition on the learning material text data set through a first machine learning model to obtain a named entity data set;

[0009] Step S3, extracting relationships from the named entity dataset to obtain named entity relationship classification results, generating text feature semantic relationships based on the named entity relationship classification results, performing knowledge fusion based on the text feature semantic relationships and adding attribute labels and relationship labels to the learning material dataset, and establishing a learning material knowledge graph through the Neo4j graph database;

[0010] Step S4, collecting the user's facial micro-expression image data through the front camera of the tablet computer, identifying the user's facial micro-expression image data through the second machine learning model to obtain a user learning state recognition result, calculating the distraction frequency based on the user learning state recognition result and performing a threshold determination, and the tablet computer executes the first message prompt and the second message prompt based on the threshold determination result;

[0011] Step S5, establish a learning training model, and set the initial course, derived courses, derived training question groups and personalized training question groups based on the subject selected by the user, the knowledge graph of the learning materials, the historical error rate of answering questions and the frequency of distraction.

[0012] As a preferred solution of the artificial intelligence-based tablet computer learning and training method described in the present invention, wherein: extracting the video course name index, teaching material text and training questions in the learning material data set, obtaining the learning material text data set by text format standardization processing, performing named entity recognition on the learning material text data set by a first machine learning model, and obtaining a named entity data set;

[0013] The named entity dataset includes concept entity set, question entity set, answer entity set, solution entity set, person entity set, event entity set, and time entity set.

[0014] As a preferred solution of the artificial intelligence-based tablet learning and training method described in the present invention, the first machine learning model is a pre-trained improved BERT named entity model, and the logic of the first machine learning model for identifying the learning material text data set is: the learning material text data set is embedded by word embedding and fragment embedding through the embedding layer to obtain an expanded data set, the text sentences of the expanded data set are vectorized and the sentences are marked to obtain a text vector set, the text vector set is weighted by the attention mechanism to obtain an updated vector set, the updated vector set is nonlinearly transformed and vector spliced ​​to obtain a text vector sequence, the text vector sequence is bidirectionally extracted through the memory network in the hidden layer, the text vector sequence is feature extracted according to the forward word order and the reverse word order respectively, the text feature vector sequence is obtained by vector splicing, the text feature vector sequence is probability labeled by the conditional random field algorithm in the output layer, and classification and decoding are completed by the joint probability of the text feature vector sequence to obtain a named entity data set.

[0015] As a preferred solution of the artificial intelligence-based tablet computer learning and training method described in the present invention, wherein: relationship extraction is performed based on a named entity data set to obtain text feature semantic relationships, knowledge fusion is performed based on the text feature semantic relationships, and a learning material knowledge graph is established through a Neo4j graph database;

[0016] The processing logic of the relationship extraction is as follows: generate graph nodes based on a named entity data set, map each named entity in the named entity data set to a node, connect each node based on a text feature vector sequence to generate a text node graph, perform weighted calculation based on a graph attention mechanism through a gating unit, update the text node state to obtain an updated text node graph, and obtain a named entity relationship classification result by performing global graph pooling and classification processing on the updated text node graph;

[0017] The named entity relationship classification results include: causal relationship, inclusion relationship, derivation relationship, opposition relationship, equivalence relationship, temporal relationship and spatial relationship.

[0018] As a preferred solution of the artificial intelligence-based tablet computer learning and training method described in the present invention, the processing logic of knowledge fusion is as follows: entity alignment processing is performed on the named entity data set through the Word2Vec semantic similarity algorithm to obtain an aligned named entity data set, relationship fusion is performed based on the named entity relationship classification result, similar relationships are merged to obtain a named entity fusion relationship, the aligned named entity data set and the named entity fusion relationship are formatted in a unified manner according to the Neo4j data import format, attribute labels and relationship labels are added to each named entity and the corresponding relationship based on the aligned named entity data set and the named entity fusion relationship, and a learning material knowledge graph is generated;

[0019] The attribute labels include subject index, primary importance, secondary importance, primary difficulty, secondary difficulty and tertiary difficulty;

[0020] The relationship labels include causal relationships, inclusion relationships, derivative relationships, opposition relationships, equivalence relationships, temporal relationships, and spatial relationships.

[0021] As a preferred solution of the artificial intelligence-based tablet computer learning training method described in the present invention, wherein: identifying the user's learning state and giving a message prompt to the user, collecting the user's facial micro-expression image data through the front camera of the tablet computer, analyzing the user's facial micro-expression image data through a second machine learning model to obtain the user's learning state recognition result, and the user's learning state recognition result includes a focused state and a distracted state;

[0022] The frequency of distraction per unit time is calculated based on the result of the user's learning state recognition, and a threshold is determined based on the distraction frequency. The threshold determination logic is: when the distraction frequency is greater than or equal to a first concentration threshold, the tablet computer executes a first message prompt; when the distraction frequency is greater than or equal to a second concentration threshold, the tablet computer executes a second message prompt;

[0023] The first message prompt is: the tablet computer microphone plays the first prompt sound effect and vibrates, and the screen display brightness changes back and forth between the original brightness and the maximum brightness at a uniform speed, and the first message prompt lasts for 2 seconds;

[0024] The second message prompt is: the tablet computer microphone plays a second prompt sound effect, stops learning and training, and pops up a rest notification.

[0025] As a preferred solution of the artificial intelligence-based tablet computer learning and training method described in the present invention, the second machine learning model is a pre-trained improved Yolov5 model, and the logic of the second machine learning model analyzing the user's facial micro-expression image is: Mosaic is used at the input end to enhance the image data, and the facial micro-expression image data of four consecutive video frames are fused to obtain a fused image;

[0026] The Focus module is introduced into the backbone network, the neighboring domain downsampling processing is added to the fused image, and the pixels of the fused image are subjected to interval sampling and segmentation convolution processing to obtain preliminary image features;

[0027] The BiFPN structure is introduced into the neck network, and weights are added to the input values ​​of each channel in the preliminary image features. The shallow fusion features and deep fusion features are obtained through weighted calculation. A cross-scale bidirectional connection is established between the convolutional layers. The shallow fusion features are transferred to the deep layer and the position information is added to the deep fusion features. The deep fusion features are transferred to the shallow layer and the semantic information is added to the shallow fusion features. The weighted fusion image features are obtained by further performing feature fusion processing through the convolutional layer.

[0028] The decoupled detection head is used to replace the detection head in the head network. The bounding box prediction and category prediction of the weighted fusion image features are respectively realized through the parallel classifier and regressor in the decoupled detection head. The learning state recognition result is obtained by combining the maximum value suppression algorithm.

[0029] The second machine learning model uses the DioU-LOSS loss function instead of the original CioU-LOSS loss function during training, and measures the loss by calculating the paradigm distance and overlapping area between the target box and the prediction box. The training set and validation set of the second machine learning model are obtained by dividing a preset facial micro-expression image set in a ratio of 7:3. The preset facial micro-expression image set is a learning state image with pre-labeled labels. The facial micro-expression image set includes focus images, daze images, nap images, yawn images, and gaze deviation images. The annotated image labels include focus state and distraction state.

[0030] As a preferred solution of the artificial intelligence-based tablet computer learning and training method described in the present invention, it includes: establishing a learning and training model, setting initial courses based on subject data selected by users, setting derived courses and derived training question groups based on relationship labels, and setting personalized training question groups based on difficulty labels, historical error rates of answering questions, and distraction frequency of derived training question groups.

[0031] As a preferred solution of the artificial intelligence-based tablet computer learning and training method described in the present invention, the logic of setting the initial course is as follows: taking the subject data selected by the user as the subject index in the attribute label, searching through the learning material knowledge graph to obtain the user target named entity set, extracting the concept entities in the user target named entity set to obtain the user target concept entity set, and based on the attribute labels of the user target concept entity set, randomly extracting from the user target concept entity set in a ratio of 4:6 between the first-level importance and the second-level importance to obtain the initial course;

[0032] The logic of setting the derived courses and derived training question groups is as follows: obtain the derived named entity set by searching the relationship labels of the user target concept entity set, extract the question entity set in the derived named entity set to obtain the derived question entity set, and based on the attribute labels of the derived question entity set, randomly extract from the derived question entity set in the ratio of 3:4:3 for the first level difficulty, the second level difficulty and the third level difficulty to obtain the derived training question group;

[0033] The logic for setting up personalized training question groups is as follows: threshold screening is performed on the historical error rates of the question entities in the derived training question group, question entities with an error rate less than or equal to the first error rate threshold are removed, relationship label search is performed on question entities with an error rate greater than the second error rate threshold to obtain an error-prone named entity set, the question entity set in the error-prone named entity set is extracted to obtain an error-prone question entity set, and based on the attribute labels of the error-prone question entity set, random sampling is performed from the error-prone question entity set in a ratio of 2:1 between the second-level difficulty and the third-level difficulty to obtain a personalized training question group.

[0034] A tablet computer learning and training system based on artificial intelligence, comprising: a data storage module, a data processing module, a learning and training module, an interaction module and a power module;

[0035] The data storage module is used to store learning material data sets, named entity data sets, learning material knowledge graphs and user historical data. The data storage module is connected to the learning and training module and the data processing module, receives the learning material knowledge graph data sent by the data processing module, receives the user historical data sent by the learning and training module, receives request instructions from the learning and training module and the data processing module and sends the corresponding stored data to the learning and training module and the data processing module;

[0036] The data processing module is used to generate a learning material knowledge graph corresponding to the learning material data set through the first machine learning model, identify the user's facial micro-expression image through the second machine learning model to obtain the user's learning status result and calculate the message prompt data. The data processing module sends the learning material knowledge graph data and the user's learning status result data to the data storage module, and sends the message prompt data to the interaction module;

[0037] The learning and training module is used to set up initial courses, derived courses, derived training question groups and personalized training question groups, and provide courses and training questions for users. The learning and training module is connected to the data storage module, sends request instructions to the data storage module and receives storage data corresponding to the instructions from the data storage module. The learning and training module is connected to the interaction module and receives the user's answer result data from the interaction module;

[0038] The interaction module is used to interact with the user, receive the user's learning status results sent by the data processing module and send message prompts, receive the courses and training questions of the learning and training module and display them to the user, record the user's answer results and send them to the learning and training module, receive user instructions and send them to the data storage module and the data processing module. User instructions include searching for named entities, selecting subjects, selecting courses, answering questions and exiting the system;

[0039] The power module is used to provide power to the data storage module, data processing module, learning and training module and interactive module.

[0040] The beneficial effects of the present invention are as follows: the named entity data set is efficiently extracted from the learning material data set by the first machine learning model, which is conducive to integrating the learning material data text data of different disciplines and categories. By establishing a learning material relationship map, it is conducive to systematically building a knowledge system for users, quickly, efficiently and accurately locating the retrieval target when the user searches, and displaying the deep relationship of each knowledge point in the learning material in the form of a structure diagram through the named entity relationship classification and the attribute label of the named entity data set, which is helpful to provide support for the scientific arrangement of courses and questions. The second machine learning model is used to quickly analyze the user's facial micro-expression image data, which is conducive to real-time obtaining whether the user is in a focused state or a distracted state, and the interactive ability is enhanced through sound and light message prompts, which is conducive to improving the user's learning and training efficiency. The improved second machine learning model is conducive to fully extracting the tiny features of the pupil, eye and mouth in the micro-expression image data. In the execution prediction task step during training, not only the overlapping area of ​​the target box and the prediction box is considered, but also the Euclidean distance between the center points of the target box and the prediction box is considered, which improves the detection accuracy of the user's facial micro-expression changes. Setting user courses and training questions according to attribute labels, difficulty labels, historical error rates in answering questions, and frequency of distraction can help to specifically address users' weak links in learning and improve the level of personalized recommendations. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 A basic flow chart of a tablet computer learning and training method based on artificial intelligence provided by one embodiment of the present invention;

[0042] Figure 2 A basic framework diagram of an artificial intelligence-based tablet computer learning and training system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0043] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, but not all of the embodiments.

[0044] Example 1, reference Figure 1 , as an embodiment of the present invention, provides a tablet computer learning and training method based on artificial intelligence, comprising:

[0045] Step S1, inputting preset video course data, textbook graphic data and training question data into the local database of the tablet computer to obtain a learning material data set;

[0046] Step S2, extracting text from the learning material data set and performing standardization processing to obtain a learning material text data set, and performing named entity recognition on the learning material text data set through a first machine learning model to obtain a named entity data set;

[0047] Step S3, extracting relationships from the named entity dataset to obtain named entity relationship classification results, generating text feature semantic relationships based on the named entity relationship classification results, performing knowledge fusion based on the text feature semantic relationships and adding attribute labels and relationship labels to the learning material dataset, and establishing a learning material knowledge graph through the Neo4j graph database;

[0048] Step S4, collecting the user's facial micro-expression image data through the front camera of the tablet computer, identifying the user's facial micro-expression image data through the second machine learning model to obtain a user learning state recognition result, calculating the distraction frequency based on the user learning state recognition result and performing a threshold determination, and the tablet computer executes the first message prompt and the second message prompt based on the threshold determination result;

[0049] Step S5, establish a learning training model, and set the initial course, derived courses, derived training question groups and personalized training question groups based on the subject selected by the user, the knowledge graph of the learning materials, the historical error rate of answering questions and the frequency of distraction.

[0050] In this embodiment, the video course name index, textbook text and training questions in the learning material data set are extracted, the learning material text data set is obtained by text format standardization processing, and the named entity recognition is performed on the learning material text data set by the first machine learning model to obtain a named entity data set;

[0051] The named entity dataset includes concept entity set, question entity set, answer entity set, solution entity set, person entity set, event entity set, and time entity set.

[0052] In this embodiment, the first machine learning model is a pre-trained improved BERT named entity model, and the logic of the first machine learning model for identifying the learning material text data set is: the learning material text data set is embedded by word embedding and fragment embedding through the embedding layer to obtain an expanded data set, the text sentences of the labeled expanded data set are vectorized and the sentences are marked to obtain a text vector set, the text vector set is weighted by the attention mechanism to obtain an updated vector set, the updated vector set is nonlinearly transformed and vector spliced ​​to obtain a text vector sequence, the text vector sequence is bidirectionally extracted through the memory network in the hidden layer, the text vector sequence is feature extracted according to the forward word order and the reverse word order respectively, a text feature vector sequence is obtained by vector splicing, the text feature vector sequence is probability labeled by the conditional random field algorithm in the output layer, and classification and decoding are completed by the joint probability of the text feature vector sequence to obtain a named entity data set.

[0053] In this embodiment, relationship extraction is performed based on the named entity data set to obtain text feature semantic relationships, knowledge fusion is performed based on the text feature semantic relationships, and a learning material knowledge graph is established through the Neo4j graph database;

[0054] The processing logic of the relationship extraction is as follows: generate graph nodes based on a named entity data set, map each named entity in the named entity data set to a node, connect each node based on a text feature vector sequence to generate a text node graph, perform weighted calculation based on a graph attention mechanism through a gating unit, update the text node state to obtain an updated text node graph, and obtain a named entity relationship classification result by performing global graph pooling and classification processing on the updated text node graph;

[0055] The named entity relationship classification results include: causal relationship, inclusion relationship, derivation relationship, opposition relationship, equivalence relationship, temporal relationship and spatial relationship.

[0056] In this embodiment, the processing logic of knowledge fusion is as follows: entity alignment processing is performed on the named entity dataset through the Word2Vec semantic similarity algorithm to obtain an aligned named entity dataset, relationship fusion is performed based on the named entity relationship classification results, similar relationships are merged to obtain a named entity fusion relationship, the aligned named entity dataset and the named entity fusion relationship are formatted in a unified manner according to the Neo4j data import format, attribute labels and relationship labels are added to each named entity and the corresponding relationship based on the aligned named entity dataset and the named entity fusion relationship, and a learning material knowledge graph is generated;

[0057] The attribute labels include subject index, primary importance, secondary importance, primary difficulty, secondary difficulty and tertiary difficulty;

[0058] The relationship labels include causal relationships, inclusion relationships, derivative relationships, opposition relationships, equivalence relationships, temporal relationships, and spatial relationships.

[0059] The named entity dataset includes a concept entity set, a question entity set, an answer entity set, a solution entity set, a person entity set, an event entity set and a time entity set;

[0060] The named entity relationship classification results include: causal relationship, inclusion relationship, derivation relationship, opposition relationship, equivalence relationship, temporal relationship and spatial relationship.

[0061] Among them, the first machine learning model is used to efficiently extract the learning material dataset to obtain a named entity dataset, which is conducive to integrating the learning material data text data of different subjects and categories. By establishing a learning material relationship map, it is helpful to systematically build a knowledge system for users, and quickly, efficiently and accurately locate the retrieval target when users search. Through the named entity relationship classification and the attribute labels of the named entity dataset, the deep relationship between each knowledge point in the learning material is displayed in the form of a structure diagram, which helps to provide support for the scientific arrangement of courses and questions.

[0062] In this embodiment, the user's learning status is identified and a message prompt is given to the user. The user's facial micro-expression image data is collected by the front camera of the tablet computer. The user's facial micro-expression image data is analyzed by the second machine learning model to obtain the user's learning status recognition result. The user's learning status recognition result includes a focused state and a distracted state.

[0063] The frequency of distraction per unit time is calculated based on the result of the user's learning state recognition, and a threshold is determined based on the distraction frequency. The threshold determination logic is: when the distraction frequency is greater than or equal to a first concentration threshold, the tablet computer executes a first message prompt; when the distraction frequency is greater than or equal to a second concentration threshold, the tablet computer executes a second message prompt;

[0064] The first message prompt is: the tablet computer microphone plays the first prompt sound effect and vibrates, and the screen display brightness changes back and forth between the original brightness and the maximum brightness at a uniform speed, and the first message prompt lasts for 2 seconds;

[0065] The second message prompt is: the tablet computer microphone plays a second prompt sound effect, stops learning and training, and pops up a rest notification.

[0066] In this embodiment, the second machine learning model is a pre-trained improved Yolov5 model, and the logic of the second machine learning model for analyzing the user's facial micro-expression image is: Mosaic is used at the input end to enhance the image data, and the facial micro-expression image data of four consecutive video frames are fused to obtain a fused image;

[0067] The Focus module is introduced into the backbone network, the neighboring domain downsampling processing is added to the fused image, and the pixels of the fused image are subjected to interval sampling and segmentation convolution processing to obtain preliminary image features;

[0068] The BiFPN structure is introduced into the neck network, and weights are added to the input values ​​of each channel in the preliminary image features. The shallow fusion features and deep fusion features are obtained through weighted calculation. A cross-scale bidirectional connection is established between the convolutional layers. The shallow fusion features are transferred to the deep layer and the position information is added to the deep fusion features. The deep fusion features are transferred to the shallow layer and the semantic information is added to the shallow fusion features. The weighted fusion image features are obtained by further performing feature fusion processing through the convolutional layer.

[0069] The decoupled detection head is used to replace the detection head in the head network. The bounding box prediction and category prediction of the weighted fusion image features are respectively realized through the parallel classifier and regressor in the decoupled detection head. The learning state recognition result is obtained by combining the maximum value suppression algorithm.

[0070] The second machine learning model uses the DioU-LOSS loss function instead of the original CioU-LOSS loss function during training, and measures the loss by calculating the paradigm distance and overlapping area between the target box and the prediction box. The training set and validation set of the second machine learning model are obtained by dividing a preset facial micro-expression image set in a ratio of 7:3. The preset facial micro-expression image set is a learning state image with pre-labeled labels. The facial micro-expression image set includes focus images, daze images, nap images, yawn images, and gaze deviation images. The annotated image labels include focus state and distraction state.

[0071] Among them, the second machine learning model is used to quickly analyze the user's facial micro-expression image data, which is conducive to real-time understanding of whether the user is in a focused state or a distracted state. The interactive ability is enhanced through sound and light message prompts, which is conducive to improving the user's learning and training efficiency. The improved second machine learning model is conducive to fully extracting the tiny features of the pupil, eyes and mouth in the micro-expression image data. In the step of executing the prediction task during training, not only the overlapping area of ​​the target box and the prediction box is considered, but also the Euclidean distance between the center points of the target box and the prediction box is considered, which improves the detection accuracy of the user's facial micro-expression changes.

[0072] In this embodiment, a learning training model is established, an initial course is set based on the subject data selected by the user, derivative courses and derivative training question groups are set based on relationship labels, and personalized training question groups are set based on the difficulty labels, historical error rates and distraction frequencies of the derivative training question groups.

[0073] In this embodiment, the logic of setting the initial course is as follows: taking the subject data selected by the user as the subject index in the attribute label, searching through the learning material knowledge graph to obtain the user target named entity set, extracting the concept entities in the user target named entity set to obtain the user target concept entity set, and based on the attribute labels of the user target concept entity set, randomly extracting from the user target concept entity set in a ratio of 4:6 between the first-level importance and the second-level importance to obtain the initial course;

[0074] The logic of setting the derived courses and derived training question groups is as follows: obtain the derived named entity set by searching the relationship labels of the user target concept entity set, extract the question entity set in the derived named entity set to obtain the derived question entity set, and based on the attribute labels of the derived question entity set, randomly extract from the derived question entity set in the ratio of 3:4:3 for the first level difficulty, the second level difficulty and the third level difficulty to obtain the derived training question group;

[0075] The logic for setting up personalized training question groups is as follows: threshold screening is performed on the historical error rates of the question entities in the derived training question group, question entities with an error rate less than or equal to the first error rate threshold are removed, relationship label search is performed on question entities with an error rate greater than the second error rate threshold to obtain an error-prone named entity set, the question entity set in the error-prone named entity set is extracted to obtain an error-prone question entity set, and based on the attribute labels of the error-prone question entity set, random sampling is performed from the error-prone question entity set in a ratio of 2:1 between the second-level difficulty and the third-level difficulty to obtain a personalized training question group.

[0076] Among them, setting user courses and training questions according to attribute labels, difficulty labels, historical error rates in answering questions, and distraction frequency is conducive to targetedly making up for the user's weak links in learning and improving the level of personalized recommendations.

[0077] Example 2, reference Figure 2 , as another embodiment of the present invention, provides a tablet computer learning and training system based on artificial intelligence, including: a data storage module, a data processing module, a learning and training module, an interaction module and a power module;

[0078] The data storage module is used to store learning material data sets, named entity data sets, learning material knowledge graphs and user historical data. The data storage module is connected to the learning and training module and the data processing module, receives the learning material knowledge graph data sent by the data processing module, receives the user historical data sent by the learning and training module, receives request instructions from the learning and training module and the data processing module and sends the corresponding stored data to the learning and training module and the data processing module;

[0079] The data processing module is used to generate a learning material knowledge graph corresponding to the learning material data set through the first machine learning model, identify the user's facial micro-expression image through the second machine learning model to obtain the user's learning status result and calculate the message prompt data. The data processing module sends the learning material knowledge graph data and the user's learning status result data to the data storage module, and sends the message prompt data to the interaction module;

[0080] The learning and training module is used to set up initial courses, derived courses, derived training question groups and personalized training question groups, and provide courses and training questions for users. The learning and training module is connected to the data storage module, sends request instructions to the data storage module and receives storage data corresponding to the instructions from the data storage module. The learning and training module is connected to the interaction module and receives the user's answer result data from the interaction module;

[0081] The interaction module is used to interact with the user, receive the user's learning status results sent by the data processing module and send message prompts, receive the courses and training questions of the learning and training module and display them to the user, record the user's answer results and send them to the learning and training module, receive user instructions and send them to the data storage module and the data processing module. User instructions include searching for named entities, selecting subjects, selecting courses, answering questions and exiting the system;

[0082] The power module is used to provide power to the data storage module, data processing module, learning and training module and interactive module.

[0083] It should be understood by those skilled in the art that embodiments of the present invention may provide methods, systems or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program codes. Among them, the storage medium may be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (Static Random Access Memory, referred to as SRAM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, referred to as EEPROM), erasable programmable read-only memory (Erasable Programmable Read Only Memory, referred to as EPROM), programmable read-only memory (Programmable Red-Only Memory, referred to as PROM), read-only memory (Read-Only Memory, referred to as ROM), magnetic memory, flash memory, disk or optical disk. These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0084] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A tablet computer learning and training method based on artificial intelligence, characterized in that: include: Step S1, inputting preset video course data, textbook graphic data and training question data into the local database of the tablet computer to obtain a learning material data set; Step S2, extracting text from the learning material data set and performing standardization processing to obtain a learning material text data set, and performing named entity recognition on the learning material text data set through a first machine learning model to obtain a named entity data set; Step S3, extracting relationships from the named entity dataset to obtain named entity relationship classification results, generating text feature semantic relationships based on the named entity relationship classification results, performing knowledge fusion based on the text feature semantic relationships and adding attribute labels and relationship labels to the learning material dataset, and establishing a learning material knowledge graph through the Neo4j graph database; Step S4, collecting the user's facial micro-expression image data through the front camera of the tablet computer, identifying the user's facial micro-expression image data through the second machine learning model to obtain a user learning state recognition result, calculating the distraction frequency based on the user learning state recognition result and performing a threshold determination, and the tablet computer executes the first message prompt and the second message prompt based on the threshold determination result; Step S5, establish a learning training model, and set the initial course, derived courses, derived training question groups and personalized training question groups based on the subject selected by the user, the knowledge graph of the learning materials, the historical error rate of answering questions and the frequency of distraction.

2. The tablet computer learning and training method based on artificial intelligence as claimed in claim 1, characterized in that: Extract the video course name index, textbook text and training questions in the learning material data set, obtain the learning material text data set through text format standardization, and perform named entity recognition on the learning material text data set through the first machine learning model to obtain a named entity data set; The named entity dataset includes concept entity set, question entity set, answer entity set, solution entity set, person entity set, event entity set, and time entity set.

3. The artificial intelligence-based tablet computer learning and training method according to claim 2, characterized in that: The first machine learning model is a pre-trained improved BERT named entity model. The logic of the first machine learning model for identifying the learning material text data set is: the learning material text data set is embedded by words and fragments through the embedding layer to obtain an expanded data set, the text sentences of the labeled expanded data set are vectorized and the sentences are marked to obtain a text vector set, the text vector set is weighted through the attention mechanism to obtain an updated vector set, the updated vector set is nonlinearly transformed and vector concatenated to obtain a text vector sequence, the text vector sequence is bidirectionally extracted through the memory network in the hidden layer, features are extracted from the text vector sequence according to the forward word order and the reverse word order respectively, a text feature vector sequence is obtained through vector concatenation, the text feature vector sequence is probability labeled through the conditional random field algorithm in the output layer, classification and decoding are completed through the joint probability of the text feature vector sequence to obtain a named entity data set.

4. The tablet computer learning and training method based on artificial intelligence as claimed in claim 1, characterized in that: Based on the named entity dataset, we extract relations to obtain semantic relations of text features, perform knowledge fusion based on the semantic relations of text features, and build a knowledge graph of learning materials through the Neo4j graph database; The processing logic of the relationship extraction is as follows: generate graph nodes based on a named entity data set, map each named entity in the named entity data set to a node, connect each node based on a text feature vector sequence to generate a text node graph, perform weighted calculation based on a graph attention mechanism through a gating unit, update the text node state to obtain an updated text node graph, and obtain a named entity relationship classification result by performing global graph pooling and classification processing on the updated text node graph; The named entity relationship classification results include: causal relationship, inclusion relationship, derivation relationship, opposition relationship, equivalence relationship, temporal relationship and spatial relationship.

5. The tablet computer learning and training method based on artificial intelligence as claimed in claim 4, characterized in that: The processing logic of knowledge fusion is as follows: perform entity alignment processing on the named entity dataset through the Word2Vec semantic similarity algorithm to obtain an aligned named entity dataset, perform relationship fusion based on the named entity relationship classification results, merge similar relationships to obtain named entity fusion relationships, unify the formats of the aligned named entity dataset and the named entity fusion relationships according to the Neo4j data import format, add attribute labels and relationship labels to each named entity and corresponding relationship based on the aligned named entity dataset and the named entity fusion relationships, and generate a learning material knowledge graph; The attribute labels include subject index, primary importance, secondary importance, primary difficulty, secondary difficulty and tertiary difficulty; The relationship labels include causal relationships, inclusion relationships, derivative relationships, opposition relationships, equivalence relationships, temporal relationships, and spatial relationships.

6. The tablet computer learning and training method based on artificial intelligence as claimed in claim 1, characterized in that: Identify the user's learning status and give the user a message prompt, collect the user's facial micro-expression image data through the front camera of the tablet computer, and analyze the user's facial micro-expression image data through the second machine learning model to obtain the user's learning status recognition result, which includes the focused state and the distracted state; The frequency of distraction per unit time is calculated based on the result of the user's learning state recognition, and a threshold is determined based on the distraction frequency. The threshold determination logic is: when the distraction frequency is greater than or equal to a first concentration threshold, the tablet computer executes a first message prompt; when the distraction frequency is greater than or equal to a second concentration threshold, the tablet computer executes a second message prompt; The first message prompt is: the tablet computer microphone plays the first prompt sound effect and vibrates, and the screen display brightness changes back and forth between the original brightness and the maximum brightness at a uniform speed, and the first message prompt lasts for 2 seconds; The second message prompt is: the tablet computer microphone plays a second prompt sound effect, stops learning and training, and pops up a rest notification.

7. The artificial intelligence-based tablet computer learning and training method according to claim 6, characterized in that: The second machine learning model is a pre-trained improved Yolov5 model. The logic of the second machine learning model for analyzing the user's facial micro-expression image is: Mosaic is used at the input end to enhance the image data, and the facial micro-expression image data of four consecutive video frames are fused to obtain a fused image; The Focus module is introduced into the backbone network, the neighboring domain downsampling processing is added to the fused image, and the pixels of the fused image are subjected to interval sampling and segmentation convolution processing to obtain preliminary image features; The BiFPN structure is introduced into the neck network, and weights are added to the input values ​​of each channel in the preliminary image features. The shallow fusion features and deep fusion features are obtained through weighted calculation. A cross-scale bidirectional connection is established between the convolutional layers. The shallow fusion features are transferred to the deep layer and the position information is added to the deep fusion features. The deep fusion features are transferred to the shallow layer and the semantic information is added to the shallow fusion features. The weighted fusion image features are obtained by further performing feature fusion processing through the convolutional layer. The decoupled detection head is used to replace the detection head in the head network. The bounding box prediction and category prediction of the weighted fusion image features are respectively realized through the parallel classifier and regressor in the decoupled detection head. The learning state recognition result is obtained by combining the maximum value suppression algorithm. The second machine learning model uses the DioU loss function instead of the original CioU loss function during training, and measures the loss by calculating the paradigm distance and overlapping area between the target box and the prediction box. The training set and validation set of the second machine learning model are obtained by dividing the preset facial micro-expression image set in a ratio of 7:

3. The preset facial micro-expression image set is a learning state image with pre-labeled labels. The facial micro-expression image set includes focus images, daze images, nap images, yawn images and gaze deviation images. The annotated image labels include focus state and distraction state.

8. The tablet computer learning and training method based on artificial intelligence as claimed in claim 1, characterized in that: Establish a learning and training model, set the initial course based on the subject data selected by the user, set the derivative courses and derivative training question groups based on the relationship labels, and set the personalized training question groups based on the difficulty labels, historical error rates and distraction frequencies of the derivative training question groups.

9. The artificial intelligence-based tablet computer learning and training method according to claim 8, characterized in that: The logic of setting the initial course is as follows: take the subject data selected by the user as the subject index in the attribute label, search through the learning material knowledge graph to obtain the user target named entity set, extract the concept entities in the user target named entity set to obtain the user target concept entity set, and based on the attribute labels of the user target concept entity set, randomly extract from the user target concept entity set in a ratio of 4:6 between the first-level importance and the second-level importance to obtain the initial course; The logic of setting the derived courses and derived training question groups is as follows: obtain the derived named entity set by searching the relationship labels of the user target concept entity set, extract the question entity set in the derived named entity set to obtain the derived question entity set, and based on the attribute labels of the derived question entity set, randomly extract from the derived question entity set in the ratio of 3:4:3 for the first level difficulty, the second level difficulty and the third level difficulty to obtain the derived training question group; The logic for setting up personalized training question groups is as follows: threshold screening is performed on the historical error rates of the question entities in the derived training question group, question entities with an error rate less than or equal to the first error rate threshold are removed, relationship label search is performed on question entities with an error rate greater than the second error rate threshold to obtain an error-prone named entity set, the question entity set in the error-prone named entity set is extracted to obtain an error-prone question entity set, and based on the attribute labels of the error-prone question entity set, random sampling is performed from the error-prone question entity set in a ratio of 2:1 between the second-level difficulty and the third-level difficulty to obtain a personalized training question group.

10. A tablet computer learning and training system based on artificial intelligence, comprising: Data storage module, data processing module, learning and training module, interaction module and power supply module; The data storage module is used to store learning material data sets, named entity data sets, learning material knowledge graphs and user historical data. The data storage module is connected to the learning and training module and the data processing module, receives the learning material knowledge graph data sent by the data processing module, receives the user historical data sent by the learning and training module, receives request instructions from the learning and training module and the data processing module and sends the corresponding stored data to the learning and training module and the data processing module; The data processing module is used to generate a learning material knowledge graph corresponding to the learning material data set through the first machine learning model, identify the user's facial micro-expression image through the second machine learning model to obtain the user's learning status result and calculate the message prompt data. The data processing module sends the learning material knowledge graph data and the user's learning status result data to the data storage module, and sends the message prompt data to the interaction module; The learning and training module is used to set up initial courses, derived courses, derived training question groups and personalized training question groups, and provide courses and training questions for users. The learning and training module is connected to the data storage module, sends request instructions to the data storage module and receives storage data corresponding to the instructions from the data storage module. The learning and training module is connected to the interaction module and receives the user's answer result data from the interaction module; The interaction module is used to interact with the user, receive the user's learning status results sent by the data processing module and send message prompts, receive the courses and training questions of the learning and training module and display them to the user, record the user's answer results and send them to the learning and training module, receive user instructions and send them to the data storage module and the data processing module. User instructions include searching for named entities, selecting subjects, selecting courses, answering questions and exiting the system; The power module is used to provide power to the data storage module, data processing module, learning and training module and interactive module.

Images