AI digital textbook interaction method and device based on knowledge graph enhancement and medium

Abstract

The invention discloses an AI digital textbook interaction method and device based on knowledge graph enhancement and a medium, and can be applied to the technical field of artificial intelligence and education fusion. According to the method, after the digital textbook knowledge graph and the vector retrieval library are constructed according to the digital textbook content and the heuristic education large model is constructed according to the teacher teaching corpus, the personalized learning path recommendation information of the students is generated according to the digital textbook knowledge graph, the vector retrieval library and the real-time interaction behavior data; therefore, a systematized knowledge network and a learning path can be provided for students, reply statements corresponding to the question text are generated according to the question text, the digital textbook knowledge graph, the vector retrieval library and the heuristic education large model, and intelligent question answering can be performed in real time. And generating teaching record suggestion information according to the teaching design demand information and the digital textbook knowledge graph to provide intelligent teaching suggestions for teachers.

Description

Technical Field

[0001] The present application relates to the technical field of integration of artificial intelligence and education, and in particular to an AI digital teaching material interaction method, device and medium based on knowledge graph enhancement. Background Art

[0002] Among the related technologies, digital textbook technology has become a key driving force for the digital transformation of education. Existing digital textbook platforms mainly include content display, interactive design, knowledge management and other aspects. However, the current digital textbook platforms are still unable to conduct real-time intelligent Q&A in terms of intelligent interaction and personalized learning assistance. Moreover, due to the use of simple directory structures or keyword indexes, it is impossible to build a systematic knowledge network and learning path for the object, and it is also difficult to provide teachers with intelligent teaching design suggestions.

[0003] In summary, the technical problems existing in the relevant technologies need to be improved. Summary of the invention

[0004] The main purpose of the embodiments of the present application is to propose an AI digital teaching material interaction method, device and medium based on knowledge graph enhancement, which can perform real-time intelligent question and answer, provide a systematic knowledge network and learning path to the object, and provide teachers with intelligent teaching suggestions.

[0005] To achieve the above objectives, one aspect of an embodiment of the present application proposes an AI digital teaching material interaction method based on knowledge graph enhancement, the method comprising:

[0006] Construct a digital textbook knowledge graph based on the digital textbook content;

[0007] Constructing a vector retrieval library according to the digital teaching material content;

[0008] Construct a heuristic education model based on the teacher's teaching corpus;

[0009] The AI ​​tutor module is used to obtain students’ real-time interactive behavior data, question texts, and teachers’ teaching design requirements information;

[0010] Generate personalized learning path recommendation information for the student based on the digital teaching material knowledge graph, the vector retrieval library and the real-time interactive behavior data;

[0011] Generate a reply sentence corresponding to the question text according to the question text, the digital teaching material knowledge graph, the vector retrieval library and the heuristic education macro model;

[0012] Generate teaching record suggestion information according to the teaching design requirement information and the digital teaching material knowledge graph;

[0013] The digital teaching material knowledge graph is updated according to the feedback information from the students or the teachers.

[0014] In some embodiments, constructing a digital teaching material knowledge graph based on the digital teaching material content includes:

[0015] Preprocessing the digital teaching material content to obtain a digital teaching material text corpus;

[0016] Extracting knowledge point entities from the digital teaching material text corpus through a named entity recognition model; and identifying semantic relationships between the knowledge point entities through a dependency syntax analysis algorithm;

[0017] The digital teaching material knowledge graph is constructed using the knowledge point entities as nodes and the semantic relationships as edges.

[0018] In some embodiments, the step of constructing a vector search library based on the digital teaching material content includes:

[0019] Dividing the digital teaching material content into texts to obtain a plurality of text segments;

[0020] Mapping the plurality of text segments into semantic vector representations;

[0021] The vector retrieval library is constructed according to all the semantic vector representations.

[0022] In some embodiments, the step of constructing a heuristic education model based on a teacher's lecture corpus includes:

[0023] Acquire the teacher's teaching corpus, which includes real corpus data from the teacher's daily teaching and guidance process;

[0024] Constructing a multi-round heuristic teaching dialogue dataset based on the teacher's teaching corpus;

[0025] According to the multi-round heuristic teaching dialogue data set and the digital teaching material knowledge graph, a general large language model is trained to obtain the heuristic education large model.

[0026] In some embodiments, generating the personalized learning path recommendation information of the student according to the digital teaching material knowledge graph, the vector retrieval library and the real-time interactive behavior data includes:

[0027] Determining text of interest based on the real-time interactive behavior data;

[0028] Converting the text of interest into a first question vector;

[0029] Mapping the first question vector to the digital textbook knowledge graph to obtain the contextual semantic context corresponding to the text of interest;

[0030] Mapping the question vector to the vector search library to obtain the text content corresponding to the context semantic context;

[0031] The personalized learning path recommendation information of the student is generated according to the contextual semantic context and the text content.

[0032] In some embodiments, the step of generating a reply statement corresponding to the question text according to the question text, the digital teaching material knowledge graph, the vector retrieval library and the heuristic education macro model includes:

[0033] Converting the question text into a second question vector;

[0034] Mapping the question vector to the digital teaching material knowledge graph, and determining the knowledge graph node corresponding to the question text;

[0035] Mapping the question vector to the vector retrieval library to obtain the textbook text paragraph corresponding to the knowledge graph node;

[0036] Generate context information based on the question text, the knowledge graph node and the textbook text paragraph;

[0037] The context information is input into the heuristic educational model, and natural language answer information is predicted as the reply sentence.

[0038] In some embodiments, generating teaching record suggestion information according to the teaching design requirement information and the digital teaching material knowledge graph includes:

[0039] Extracting core knowledge points from the digital teaching material knowledge graph according to the teaching design requirement information;

[0040] Generate a knowledge point recommendation list based on the core knowledge points;

[0041] Performing teaching topological sorting on the knowledge points in the teaching recommendation list;

[0042] The teaching record suggestion information is generated according to the teaching topological sorting result.

[0043] In some embodiments, updating the digital teaching material knowledge graph according to the feedback information of the student or the teacher includes:

[0044] Generate an analysis report based on the feedback information, the analysis report including the number of questions asked about knowledge points and satisfaction with answers;

[0045] Determine the knowledge points to be optimized that meet the preset requirements according to the analysis report;

[0046] The digital teaching material knowledge graph is expanded and optimized according to the knowledge points to be optimized.

[0047] To achieve the above objective, another aspect of the present application provides a computer device, including:

[0048] at least one processor;

[0049] at least one memory for storing at least one program;

[0050] When the at least one program is executed by the at least one processor, the at least one processor implements the above method.

[0051] To achieve the above objective, another aspect of an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program implements the above method when executed by a processor.

[0052] The embodiments of the present application include at least the following beneficial effects: The present application provides an AI digital textbook interaction method, device and medium based on knowledge graph enhancement. The scheme constructs a digital textbook knowledge graph and a vector retrieval library according to the digital textbook content, and constructs a heuristic education model according to the teacher's teaching corpus. Then, the AI ​​tutor module obtains the students' real-time interaction behavior data, question text and the teacher's teaching design demand information, and generates personalized learning path recommendation information for students according to the digital textbook knowledge graph, the vector retrieval library and the real-time interaction behavior data, so as to provide students with a systematic knowledge network and learning path, and generates a reply sentence corresponding to the question text according to the question text, the digital textbook knowledge graph, the vector retrieval library and the heuristic education model, so as to conduct intelligent question and answer in real time, and generate teaching record suggestion information according to the teaching design demand information and the digital textbook knowledge graph, so as to provide teachers with intelligent teaching suggestions; at the same time, the present application also updates the digital textbook knowledge graph according to the feedback information, which can improve the accuracy of intelligent question and answer, learning path generation and teaching suggestions during the application process. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 It is a flow chart of an AI digital teaching material interaction method based on knowledge graph enhancement provided in an embodiment of the present application;

[0054] Figure 2 It is a schematic diagram of the application framework of the AI ​​digital teaching material interaction method based on knowledge graph enhancement provided in an embodiment of the present application;

[0055] Figure 3 It is a training diagram of the heuristic education model provided in the embodiment of the present application;

[0056] Figure 4It is a schematic diagram of the working principle of the AI ​​tutor module provided in the embodiment of the present application;

[0057] Figure 5 is a schematic diagram of application interaction provided by an embodiment of the present application;

[0058] Figure 6 It is a schematic diagram of the hardware structure of the computer device provided in the embodiment of the present application. DETAILED DESCRIPTION

[0059] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application is further described in detail below in conjunction with the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the embodiments of the present application, they are only examples of devices and methods consistent with some aspects of the embodiments of the present application.

[0060] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0061] Among the related technologies, digital textbook technology has become a key driving force for the digital transformation of education. Existing digital textbook platforms mainly include content display, interactive design, knowledge management and other aspects. However, the current digital textbook platforms still have the following problems in terms of intelligent interaction and personalized learning assistance:

[0062] First, the intelligent interaction technology is insufficient: the existing digital teaching material systems mostly adopt preset interaction modes and lack real-time intelligent question-and-answer technology based on natural language processing, which cannot meet students' instant question-and-answer needs; the interactive window can only interact with students in the form of "one question and one answer", which is a single method and cannot provide inspiring guidance to the objects; it is difficult to understand students' complex problems and cannot provide targeted answers and knowledge recommendations.

[0063] Second, the limitation of knowledge representation technology: the current knowledge organization of digital textbooks mostly adopts a simple directory structure or keyword index, lacking deep semantic representation and association analysis technology for knowledge content. This makes it difficult for the system to support complex knowledge reasoning and personalized recommendations, and it is impossible to build a systematic knowledge network and learning path for students.

[0064] Third, teaching assistance technology is weak: the existing system has limited technical support for teaching design assistance and lacks intelligent teaching plan generation technology based on textbook content. It is difficult for teachers to quickly retrieve and associate knowledge points in textbooks, and they cannot obtain intelligent teaching design suggestions, which greatly limits the improvement of teaching efficiency.

[0065] In view of this, an embodiment of the present application provides an AI digital teaching material interaction method, device and medium based on knowledge graph enhancement. The present application integrates the digital teaching material interaction process with the knowledge graph and the generative AI algorithm, thereby providing real-time intelligent question-and-answer functions for designated students, and providing students with a systematic knowledge network and learning path, while providing teachers with intelligent teaching suggestions.

[0066] The AI ​​digital teaching material interaction method based on knowledge graph enhancement provided in the embodiment of the present application relates to the field of artificial intelligence and education integration technology. The AI ​​digital teaching material interaction method based on knowledge graph enhancement provided in the embodiment of the present application can be applied to a terminal, can also be applied to a server, and can also be software running in a terminal or a server. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart speaker, a smart watch, and a car terminal, etc., but is not limited to this; the server side can be configured as an independent physical server, or it can be configured as a server cluster or distributed system composed of multiple physical servers, and can also be configured to provide cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The cloud server of basic cloud computing services, the server can also be a node server in the blockchain network; the software can be an application that implements the AI ​​digital teaching material interaction method based on knowledge graph enhancement, etc., but is not limited to the above forms.

[0067] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0068] It should be noted that in each specific implementation of the present application, when it comes to the need to perform relevant processing based on data related to the identity or characteristics of the object, such as object information, object behavior data, object historical data, and object location information, the permission or consent of the object will be obtained first, and the collection, use, and processing of these data will comply with relevant laws, regulations, and standards. In addition, when the embodiment of the present application needs to obtain sensitive personal information of the object, the separate permission or separate consent of the object will be obtained through a pop-up window or by jumping to a confirmation page. After clearly obtaining the separate permission or separate consent of the object, the necessary object-related data for the normal operation of the embodiment of the present application will be obtained.

[0069] Figure 1 is an optional flowchart of the AI ​​digital teaching material interaction method based on knowledge graph enhancement provided in an embodiment of the present application. Figure 1 The method may include but is not limited to steps S101 to S106.

[0070] Step S110, constructing a digital textbook knowledge graph according to the digital textbook content;

[0071] Step S120, constructing a vector retrieval library according to the digital teaching material content;

[0072] Step S130, constructing a heuristic education model based on the teacher's teaching corpus;

[0073] Step S140: obtaining students' real-time interactive behavior data, question texts, and teachers' teaching design demand information through the AI ​​tutor module;

[0074] Step S150: Generate personalized learning path recommendation information for students based on the digital teaching material knowledge graph, vector retrieval library and real-time interactive behavior data;

[0075] Step S160, generating a reply statement corresponding to the question text according to the question text, the digital teaching material knowledge graph, the vector retrieval library and the heuristic education macro model;

[0076] Step S170: Generate teaching record suggestion information according to teaching design requirement information and digital teaching material knowledge graph;

[0077] Step S180: Update the digital teaching material knowledge graph based on feedback from students or teachers.

[0078] In the embodiment of the present application, the process of constructing a digital textbook knowledge graph based on the digital textbook content includes but is not limited to the following steps:

[0079] Step S111, preprocessing the digital textbook content to obtain a digital textbook text corpus;

[0080] Step S112: extracting knowledge point entities from the digital teaching material text corpus through a named entity recognition model; and identifying semantic relationships between knowledge point entities through a dependency syntax analysis algorithm;

[0081] Step S113: construct a digital teaching material knowledge graph using knowledge point entities as nodes and semantic relationships as edges.

[0082] It is understandable that if Figure 2 As shown, this embodiment can automatically obtain digital teaching material content such as each unit, section, knowledge point, etc. of the teaching material from the digital teaching material platform through the API interface. The digital teaching material content includes text content, structural information (such as chapter level) and metadata (such as difficulty level). The obtained digital teaching material content is preprocessed, and the preprocessing process includes but is not limited to word segmentation, part-of-speech tagging, and named entity recognition.

[0083] In this embodiment, a BERT-based named entity recognition (NER) model may be used. The training process of the named entity recognition model includes the following steps:

[0084] Step 1: Data preparation: Manually annotate 5,000 textbook text samples and mark the knowledge point entities.

[0085] Step 2: Model selection: Use the pre-trained BERT-Base model as the basis.

[0086] Step 3: Fine-tune training: Fine-tune the BERT model on the labeled data, train for 10 epochs, and set the learning rate to 5e-5.

[0087] Step 4: Validation and testing: Use 20% of the labeled data as a validation set and prepare 1,000 independent test sets to evaluate model performance.

[0088] The trained named entity recognition model of this embodiment achieved an F1 score of 95% on the test set.

[0089] It is understandable that the present embodiment uses a relation extraction method based on dependency algorithm analysis to identify the semantic relationship between knowledge point entities. The algorithm uses the Stanford CoreNLP tool for dependency syntactic analysis, and then extracts the semantic relationship between entities through predefined syntactic patterns. The extracted knowledge point entities and semantic relationships are stored in the Neo4j graph database to form a digital teaching material knowledge graph. Among them, the knowledge point entities are used as nodes of the graph, and the semantic relationship between entities is used as the edge of the graph. The attributes of each node include the name of the knowledge point, type (such as concept, definition, formula, etc.), and difficulty level (1-5 levels). The attributes of each edge include relationship type (such as "contains", "lead", "related", etc.), relationship strength (0-1 floating point number).

[0090] In the embodiment of the present application, the process of constructing a vector search library based on the digital teaching material content includes but is not limited to the following steps:

[0091] Step S121, dividing the digital teaching material content into text segments to obtain a plurality of text segments;

[0092] Step S122, mapping a plurality of text segments into semantic vector representations;

[0093] Step S123: construct a vector retrieval library based on all semantic vector representations.

[0094] It is understandable that if Figure 2 As shown, in this embodiment, after the digital teaching material content is divided into several text segments according to natural paragraphs or semantic integrity, each text segment is mapped to a semantic vector representation through pre-trained semantic vector models such as Embedding-Ada-002, M3E and Word2Vec, and all semantic vector representations are stored in a vector retrieval library. The content in the vector retrieval library can be retrieved in the form of vectors.

[0095] In the embodiment of the present application, the process of constructing a heuristic education model based on the teacher's teaching corpus includes but is not limited to the following steps:

[0096] Step S131, obtaining a teacher's teaching corpus, wherein the teacher's teaching corpus includes real corpus data in the teacher's daily teaching and guidance process;

[0097] Step S132: construct a multi-round heuristic teaching dialogue dataset based on the teacher's teaching corpus;

[0098] Step S133: train the general large language model based on the multi-round heuristic teaching dialogue data set and the digital teaching material knowledge graph to obtain the heuristic education large model.

[0099] It is understandable that if Figure 3As shown, this embodiment analyzes the real corpus of teachers' (teachers') daily teaching and guidance processes to construct a multi-round heuristic teaching dialogue dataset, and then combines it with the digital teaching material knowledge graph to conduct self-supervised pre-training of the general large language model in a targeted manner, thereby further improving the language understanding and generation capabilities of the general large language model in the field of education, thereby forming a heuristic education large model in the field of education.

[0100] In the embodiment of the present application, after constructing the digital textbook knowledge graph, vector retrieval library and heuristic education model, the present embodiment deploys the digital textbook and related resources in the background of the designated software, and provides students with an interactive learning interface through a web page, which presents the textbook directory structure and text content, as well as the AI ​​tutor module. The working principle of the AI ​​tutor module is as follows: Figure 4 shown.

[0101] In the embodiment of the present application, the process of generating personalized learning path recommendation information for students based on the digital teaching material knowledge graph, vector retrieval library and real-time interactive behavior data includes but is not limited to the following steps:

[0102] Step S151, determining the text of interest according to the real-time interactive behavior data;

[0103] Step S152, converting the text of interest into a first question vector;

[0104] Step S153: Map the first question vector to the digital textbook knowledge graph to obtain the contextual semantic context corresponding to the text of interest;

[0105] Step S154, mapping the question vector to a vector search library to obtain the text content corresponding to the context semantic context;

[0106] Step S155: Generate personalized learning path recommendation information for students based on the contextual semantic context and the text content.

[0107] It is understandable that when students are reading textbooks, the AI ​​tutor module obtains students' interactive behavior data in real time by combining front-end JavaScript event monitoring and back-end log recording. And embed event listeners in the front-end page to capture students' clicks, word markings, page dwell time and other behaviors, and transmit these data to the back-end server in real time to identify the corresponding text of interest. When it is detected that a student marks a certain knowledge point, the knowledge point node is queried from the digital textbook knowledge graph to obtain its associated preceding knowledge points and subsequent knowledge points, forming a local knowledge network centered on the knowledge point to describe the contextual semantic context. This local knowledge network is integrated with the relevant text content in the vector database and the rich media resources of the pre-screened external popular science website, and is input into the heuristic education model as a retrieval result to generate personalized learning path recommendation information for students, and can be pushed to students through the student interface to guide students to carry out systematic learning.

[0108] In the embodiment of the present application, the process of generating a reply statement corresponding to the question text according to the question text, the digital teaching material knowledge graph, the vector retrieval library and the heuristic education model includes but is not limited to the following steps:

[0109] Step S161, converting the question text into a second question vector;

[0110] Step S162: Map the question vector to the digital textbook knowledge graph, and determine the knowledge graph node corresponding to the question text;

[0111] Step S163: Map the question vector to the vector search library to obtain the textbook text paragraph corresponding to the knowledge graph node;

[0112] Step S164, generating context information according to the question text, the knowledge graph node and the textbook text paragraph;

[0113] Step S165: input the context information into the heuristic educational model, and predict the natural language answer information as the reply sentence.

[0114] In an embodiment of the present application, the text content can be retrieved from the vector database through the retrieval enhancement model. Specifically, the question text can be semantically understood. Among them, the semantic understanding can use the pre-trained Sentence-BERT model to encode the question text into a 768-dimensional semantic vector representation, and then query the Milvus vector database for the top 10 text vectors that are most similar to the question vector corresponding to the real-time interactive behavior data, and then return the original text content corresponding to these text vectors.

[0115] It is understood that the training process of the Sentence-BERT model includes but is not limited to the following steps:

[0116] Step 1: Data preparation: This embodiment collects a large amount of question-and-answer data in the field of education, including student questions, textbook content, knowledge point explanations, etc., and cleans and formats the data.

[0117] Step 2: Training data construction: This embodiment uses question-answer data to construct positive and negative sample pairs. Positive sample pairs consist of semantically similar questions and answers, while negative sample pairs consist of semantically unrelated questions and answers. Such data helps the model learn semantic similarity judgment.

[0118] Step 3: Model fine-tuning: This embodiment uses the pre-trained Sentence-BERT model as a basis and uses the constructed positive and negative sample data to fine-tune the model. By setting appropriate training hyperparameters (such as batch size, learning rate, etc.), the model gradually learns the semantic matching pattern in the educational scenario.

[0119] Step 4: Model evaluation: This embodiment uses a portion of the reserved data to evaluate the trained model, using cosine similarity as the evaluation indicator. After multiple rounds of tuning, the model achieved a high accuracy rate in sentence similarity judgment on education question-and-answer data.

[0120] This embodiment can use the trained Sentence-BERT model to encode the question text entered by the student into a semantic vector, query the Milvus vector database for the top 10 text vectors most similar to the question vector, and then return the original text content corresponding to these vectors. At the same time, the system encodes all nodes (knowledge point entities) in the knowledge graph into semantic vectors through the same Sentence-BERT model. Further, the cosine similarity between the question vector and all node vectors is calculated. The top 5 nodes with the highest similarity are selected as candidate answer nodes.

[0121] In order to optimize the quality of answer generation, the graph neural network (GNN) method is used to mine the context information of the candidate answer nodes. Specifically, the graph attention network (GAT) model is used, which takes the candidate answer node as the center and integrates the semantic information of its first-order and second-order neighbor nodes to form an extended semantic representation of the answer node.

[0122] Finally, the expanded answer node representation, related paragraph content and question text are input into the heuristic education model through the Langchain framework to generate interactive text in natural language. During the decoding process, the system introduces the attention mechanism and Beam Search strategy to dynamically select the information of the answer node and its neighboring nodes and organize them into coherent reply sentences.

[0123] In the embodiment of the present application, an example of a multi-round heuristic dialogue dataset is as follows:

[0124] instruction: "Why can't toilet cleaner and 84 disinfectant be mixed?"

[0125] output: "Please think about what are the ingredients of toilet cleaner and 84 disinfectant?"

[0126] Input: "It is hydrochloric acid and sodium hypochlorite"

[0127] output: "Very good! What are the properties of hydrochloric acid and sodium hypochlorite? What reaction will occur when the two are mixed?"

[0128] Input: "will react to produce chlorine gas"

[0129] Output: "That's right! When hydrochloric acid and sodium hypochlorite are mixed, they react to produce toxic chlorine gas, which can easily poison people. This reaction is the well-known redox reaction."

[0130] After the above processing, the heuristic dialogue model built with the Sentence-BERT model can not only accurately match the knowledge points related to the question, but also mine the contextual semantic information of the knowledge points. At the same time, it combines the rich media resources of external popular science websites to generate complete and logically coherent answer texts, thereby improving the satisfaction of intelligent question answering. The system also records the students' question and answer interaction data for continuous optimization of the knowledge graph.

[0131] In the embodiment of the present application, the process of generating teaching record suggestion information according to the teaching design requirement information and the digital teaching material knowledge graph includes but is not limited to the following steps:

[0132] Step S171, extracting core knowledge points from the digital teaching material knowledge graph according to teaching design requirement information;

[0133] Step S172: Generate a knowledge point recommendation list based on the core knowledge points;

[0134] Step S173, performing teaching topological sorting on the knowledge points in the teaching recommendation list;

[0135] Step S174: Generate teaching record suggestion information according to the teaching topology sorting result.

[0136] It is understandable that when teachers prepare for classes, they propose teaching design requirements to the AI ​​tutor module through the interactive interface. The AI ​​tutor module extracts relevant concepts from the digital textbook knowledge graph, uses semantic retrieval to match teaching materials, and generates teaching preparation plan suggestions in combination with teaching activity templates and students' actual learning conditions. Specifically, according to the requirements of the course outline, the core knowledge points can be matched in the digital textbook knowledge graph through the graph pattern matching algorithm, and a recommended list of key knowledge and difficult knowledge can be automatically generated. Then, with the recommended key knowledge nodes as the starting nodes, the depth-first search (DFS) algorithm is used to generate the teaching topological sorting of the knowledge points as a planning reference for the teaching process. At the same time, according to the difficulty level attributes of the knowledge points, the knowledge points are divided into the basic layer (difficulty level 1-2), the advanced layer (difficulty level 3-4) and the extended layer (difficulty level 5) to provide stratified teaching content suggestions, and then the teaching record suggestion information is generated based on the teaching topological sorting results and the stratified teaching content suggestions, so that teachers can prepare personalized teaching design plans based on these suggestion information.

[0137] In addition, the embodiment of the present application can also use the above-mentioned heuristic education model to provide teachers with more intelligent teaching assistance. Teachers can upload natural language to describe teaching needs and background, automatically analyze the needs and generate teaching design suggestions based on knowledge graphs, including teaching objectives, key points and difficulties, teaching process, homework, etc., which greatly reduces the pressure on teachers to prepare lessons.

[0138] In the embodiment of the present application, the interactive interface is a responsive web interface built by the ReactJS framework, and the results of the AI ​​tutor's intelligent Q&A, knowledge recommendation, and teaching design are presented to students and teachers in the interactive interface. Students can provide scores and comments on the Q&A results and recommended knowledge through the feedback button in the interface, and teachers can edit and modify the teaching design plan through a drag-and-drop visual editor. This embodiment uses the Kafka message queue to record the interactive feedback information of students in real time, and performs real-time analysis through the Spark stream processing engine. An analysis report including the number of questions asked by students for each knowledge point, Q&A satisfaction and other indicators is generated every week. Based on these analysis reports, the knowledge points to be optimized containing weak knowledge points or missing knowledge points in teaching resources are identified, and the knowledge map is automatically expanded and optimized accordingly. Among them, the optimization process includes using an association rule mining algorithm to discover new knowledge point relationships, adjusting the difficulty level of knowledge points through feedback information, and using a graph structure analysis algorithm to detect and delete redundant or erroneous edges. This continuous optimization mechanism ensures that the effect and intelligence level of AI tutor services can be continuously improved.

[0139] In this embodiment, when the method of the present application is applied to Figure 5 The interactive system shown includes but is not limited to the following modules:

[0140] The digital textbook platform module is configured to use the mySQL database to store, manage and organize the multimedia content of digital textbooks, such as text, pictures, audio and video. It also provides an interactive web interface to allow students to browse textbook content and perform operations such as marking words and clicking. At the same time, it collects students' learning behavior data in real time through point-of-study technology.

[0141] The knowledge graph construction module is configured as follows: using the NLTK library for basic processing such as word segmentation and part-of-speech tagging, further adopting a BERT-based named entity recognition model to identify key concepts and definitions in the textbooks, and finally using a dependency-based syntactic analysis method to identify the semantic relationship between entities, and using the Neo4j graph database to visualize and store it, thereby extracting knowledge from the digital textbook content and building a knowledge graph.

[0142] The AI ​​tutor module is configured to provide intelligent interactive services based on knowledge graphs and large language models. First, the Sentence-BERT model is used to convert student questions into semantic vectors, and then the relevant textbook paragraphs and knowledge graph nodes are retrieved from the Milvus vector database based on the vectors. The heuristic education large model is used to generate answers in combination with the retrieved knowledge. At the same time, relevant knowledge points can be recommended based on the student's learning trajectory and knowledge graph.

[0143] The teaching assistance module is configured to generate the teaching sequence of knowledge points using a depth-first search algorithm. Based on the difficulty attribute in the knowledge graph, the teaching content is layered, and a preliminary teaching plan is automatically generated in combination with course planning and difficulty analysis, so as to provide teachers with teaching design support based on the knowledge graph.

[0144] The user interaction module is configured to: be responsible for the interaction between the system and the user, including receiving and parsing user input (such as questions, word marking operations, etc.); presenting the AI ​​tutor's answers, knowledge recommendations and other results to the user; and collecting user evaluations and feedback on the system's answers.

[0145] The data analysis and optimization module is configured to: use the Kafka message queue to collect user interaction data; use Spark for large-scale data processing and analysis; and automatically update and optimize the knowledge graph based on the analysis results.

[0146] based on Figure 5 The interactive system and the method of the embodiment of the present application include but are not limited to the following steps in the application process:

[0147] Step 1: Students access textbook content through the digital textbook platform module.

[0148] Step 2: The user interaction module captures students’ operations (such as asking questions and marking words).

[0149] Step 3: The AI ​​tutor module receives user input and calls the knowledge graph and large language model to generate answers or recommendations.

[0150] Step 4: The user interaction module presents the results to students.

[0151] Step 5: The data analysis and optimization module continuously collects and analyzes student feedback data to optimize the knowledge graph and AI model.

[0152] Step 6. Teachers can obtain teaching design suggestions through the teaching assistance module.

[0153] It can be seen that the method provided in this application can automatically generate teaching design plans for teachers based on the content of the textbooks and the knowledge graph through the AI ​​tutor module, including the reasoning of key teaching knowledge points, automatic generation of lesson preparation materials and teaching scripts, etc., greatly reducing the pressure of teacher preparation. For students, the method provided in this application can significantly improve the application effect of digital textbooks, create a more intelligent and personalized learning environment for learners, and provide teachers with intelligent teaching assistants to promote the development of smart education.

[0154] The embodiment of the present application also provides a computer device, which includes a memory and a processor, wherein the memory stores a computer program, and the processor implements the above-mentioned AI digital teaching material interaction method based on knowledge graph enhancement when executing the computer program. The computer device can be any intelligent terminal including a tablet computer, a car computer, etc.

[0155] It can be understood that the contents of the above method embodiments are all applicable to the present device embodiments, the functions specifically implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0156] See also Figure 6 , Figure 6 The hardware structure of a computer device according to another embodiment is shown, and the computer device includes:

[0157] The processor 610 may be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present application;

[0158] The memory 620 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 620 can store an operating system and other applications. When the technical solution provided in the embodiment of this specification is implemented by software or firmware, the relevant program code is stored in the memory 620, and the processor 610 calls and executes the AI ​​digital teaching material interaction method based on knowledge graph enhancement in the embodiment of this application;

[0159] Input / output interface 630, used to implement information input and output;

[0160] Communication interface 640, used to realize communication interaction between the device and other devices, which can be realized through wired mode (such as USB, network cable, etc.) or wireless mode (such as mobile network, WIFI, Bluetooth, etc.);

[0161] bus 650 , which transmits information between the various components of the device (e.g., processor 610 , memory 620 , input / output interface 630 , and communication interface 640 );

[0162] The processor 610 , the memory 620 , the input / output interface 630 , and the communication interface 640 are connected to each other in communication within the device via a bus 650 .

[0163] An embodiment of the present application also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements the above-mentioned AI digital teaching material interaction method based on knowledge graph enhancement.

[0164] It can be understood that the contents of the above method embodiments are all applicable to the present storage medium embodiments, the functions specifically implemented by the present storage medium embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0165] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely disposed relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0166] The embodiments described in the embodiments of the present application are intended to more clearly illustrate the technical solutions of the embodiments of the present application and do not constitute a limitation on the technical solutions provided in the embodiments of the present application. Those skilled in the art will appreciate that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.

[0167] Those skilled in the art will appreciate that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than shown in the figures, or a combination of certain steps, or different steps.

[0168] The device embodiments described above are merely illustrative, and the units described as separate components may or may not be physically separated, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0169] Those skilled in the art will appreciate that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices may be implemented as software, firmware, hardware, or a suitable combination thereof.

[0170] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0171] It should be understood that in the present application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the objects associated before and after are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0172] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the above units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0173] The units described above as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0174] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0175] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including multiple instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of various embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, referred to as ROM), random access memory (Random Access Memory, referred to as RAM), disk or optical disk and other media that can store programs.

[0176] The preferred embodiments of the present invention are described above with reference to the accompanying drawings, but the scope of the rights of the present invention is not limited thereto. Any modification, equivalent substitution and improvement made by a person skilled in the art without departing from the scope and essence of the present invention should be within the scope of the rights of the present invention.

Claims

1. An AI digital teaching material interaction method based on knowledge graph enhancement, characterized in that: The method comprises: Construct a digital textbook knowledge graph based on the digital textbook content; Constructing a vector retrieval library according to the digital teaching material content; Construct a heuristic education model based on the teacher's teaching corpus; The AI ​​tutor module is used to obtain students’ real-time interactive behavior data, question texts, and teachers’ teaching design requirements information; Generate personalized learning path recommendation information for the student based on the digital teaching material knowledge graph, the vector retrieval library and the real-time interactive behavior data; Generate a reply sentence corresponding to the question text according to the question text, the digital teaching material knowledge graph, the vector retrieval library and the heuristic education macro model; Generate teaching record suggestion information according to the teaching design requirement information and the digital teaching material knowledge graph; The digital teaching material knowledge graph is updated according to the feedback information from the students or the teachers.

2. The method according to claim 1, characterized in that: The method of constructing a digital textbook knowledge graph according to the digital textbook content includes: Preprocessing the digital teaching material content to obtain a digital teaching material text corpus; Extracting knowledge point entities from the digital teaching material text corpus through a named entity recognition model; and identifying semantic relationships between the knowledge point entities through a dependency syntax analysis algorithm; The digital teaching material knowledge graph is constructed using the knowledge point entities as nodes and the semantic relationships as edges.

3. The method according to claim 1, characterized in that The step of constructing a vector retrieval library according to the digital teaching material content includes: Dividing the digital teaching material content into texts to obtain a plurality of text segments; Mapping the plurality of text segments into semantic vector representations; The vector retrieval library is constructed according to all the semantic vector representations.

4. The method according to claim 2, characterized in that: The heuristic education model is constructed based on the teacher's teaching corpus, including: Acquire the teacher's teaching corpus, which includes real corpus data from the teacher's daily teaching and guidance process; Constructing a multi-round heuristic teaching dialogue dataset based on the teacher's teaching corpus; According to the multi-round heuristic teaching dialogue data set and the digital teaching material knowledge graph, a general large language model is trained to obtain the heuristic education large model.

5. The method according to claim 1, characterized in that The generating of the personalized learning path recommendation information of the student according to the digital teaching material knowledge graph, the vector retrieval library and the real-time interactive behavior data includes: Determining text of interest based on the real-time interactive behavior data; Converting the text of interest into a first question vector; Mapping the first question vector to the digital textbook knowledge graph to obtain the contextual semantic context corresponding to the text of interest; Mapping the question vector to the vector search library to obtain the text content corresponding to the context semantic context; The personalized learning path recommendation information of the student is generated according to the contextual semantic context and the text content.

6. The method according to claim 1, characterized in that The step of generating a reply statement corresponding to the question text according to the question text, the digital teaching material knowledge graph, the vector retrieval library and the heuristic education macromodel comprises: Converting the question text into a second question vector; Mapping the question vector to the digital teaching material knowledge graph, and determining the knowledge graph node corresponding to the question text; Mapping the question vector to the vector retrieval library to obtain the textbook text paragraph corresponding to the knowledge graph node; Generate context information based on the question text, the knowledge graph node and the textbook text paragraph; The context information is input into the heuristic educational model, and natural language answer information is predicted as the reply sentence.

7. The method according to claim 1, characterized in that Generating teaching record suggestion information according to the teaching design requirement information and the digital teaching material knowledge graph includes: Extracting core knowledge points from the digital teaching material knowledge graph according to the teaching design requirement information; Generate a knowledge point recommendation list based on the core knowledge points; Performing teaching topological sorting on the knowledge points in the teaching recommendation list; The teaching record suggestion information is generated according to the teaching topological sorting result.

8. The method according to claim 1, characterized in that The updating of the digital teaching material knowledge graph according to the feedback information of the student or the teacher includes: Generate an analysis report based on the feedback information, the analysis report including the number of questions asked about knowledge points and satisfaction with answers; Determine the knowledge points to be optimized that meet the preset requirements according to the analysis report; The digital teaching material knowledge graph is expanded and optimized according to the knowledge points to be optimized.

9. A computer device, characterized in that: include: at least one processor; at least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the method according to any one of claims 1 to 8.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 8 is implemented.

Images