Intelligent education activity management system and method based on large model

Abstract

The invention belongs to the technical field of information management, and discloses an intelligent education activity management system and method based on a large model. The system comprises four layers of data acquisition and access, processing and storage, intelligent analysis and decision, and user interaction and application. Education data are analyzed through fusion of transfer learning and a large model, an activity process is optimized based on a semantic network and the large model, and personalized learning services are provided by utilizing emotion interaction and the large model. The method is implemented through the steps of data acquisition, processing, model training, application and the like. Experiments show that the method can significantly improve student scores, improve educational activity organization efficiency and student satisfaction, provide powerful support for intelligent management of educational activities, and have good application prospects.

Description

Technical Field

[0001] The present invention relates to the technical field of information management, and particularly relates to an intelligent education activity management system and method based on large models. Background Art

[0002] With the continuous advancement of educational informatization, the forms and contents of educational activities have become increasingly rich and diverse. Traditional educational activity management methods mainly rely on manual operations and simple information systems, which expose many problems when facing a large amount of educational data and complex educational activity requirements. For example, it is difficult to comprehensively and accurately analyze students' learning situations, and it is impossible to provide personalized educational services according to students' individual differences; the organization and arrangement of educational activities lack scientificity and flexibility, easily resulting in waste of educational resources; the communication and interaction between teachers and students are not efficient enough, affecting the improvement of teaching effects.

[0003] The emergence of large model technology provides new ideas and methods for solving these problems. Large models have powerful language understanding, generation, and knowledge reasoning capabilities, and can deeply analyze and mine large-scale educational data, providing strong support for the management and decision-making of educational activities. However, the current research and practice of applying large model technology to educational activity management are relatively few, and no mature systems and methods have been formed yet. Summary of the Invention

[0004] The purpose of the present invention is an educational activity management system based on large model technology, including: A data collection and access layer, deploying a variety of data collection interfaces for docking with the school's internal educational administration system, student management system, teaching resource library, and external educational data, and using data encryption and desensitization technologies to ensure data security and privacy; A data processing and storage layer, performing cleaning, conversion, and integration preprocessing operations on the collected data, storing the data using distributed storage technology and establishing indexes, and using data warehouse technology for multidimensional analysis and management; An intelligent analysis and decision-making layer, including an educational data analysis platform integrating transfer learning and large models, an intelligent optimization algorithm for educational activity processes based on semantic networks and large models, and a personalized learning service system for emotional interaction and large models, deeply analyzing and mining educational data to provide intelligent decision-making support, and also including a model training and management module; A user interaction and application layer, providing a friendly human-computer interaction interface for teachers, students, and educational administrators to meet the operation needs of different users.

[0005] Furthermore, the educational data analysis platform integrating transfer learning and large models includes: A transfer learning module, built for the diversity and professionalism of educational data, uses a large model pre-trained on large-scale general data and transfers knowledge to the education field through a transfer learning method based on the attention mechanism; A large model adaptation and optimization module selects a large model suitable for educational data processing, such as a pre-trained language model based on the GPT architecture, and performs supervised fine-tuning training using a professional corpus in the education field; A data analysis process cleans, annotates, and performs feature engineering on educational data, then inputs it into local training of the large model, and uses transfer learning technology to optimize the model for student learning situation assessment and learning progress prediction.

[0006] Furthermore, the intelligent optimization algorithm for the educational activity process of the semantic network and the large model includes: An educational semantic network construction module collects and integrates knowledge in the education field, constructs structured semantic network data through entity extraction, relationship extraction, and semantic annotation technologies, and stores it in a graph database; An activity process modeling and optimization module abstracts the educational activity process into a directed graph, uses the large model to analyze the semantic network, and mines bottlenecks and optimization points in the process; An intelligent decision support module searches for information in the semantic network according to the educational activity status, and provides activity organization plans and decision-making suggestions in combination with the intelligent recommendation function of the large model.

[0007] Furthermore, the personalized learning service system for emotional interaction and the large model includes: An emotional interaction data collection and fusion module integrates voice emotion recognition, text emotion analysis, and facial expression recognition to collect and fuse students' emotional interaction data; A large model-driven personalized learning recommendation module uses the large model to combine students' learning situations, interests, and emotional states to provide personalized learning resource recommendations and learning path planning; A learning effect evaluation and feedback module evaluates students' learning effects through multi-faceted data, and adjusts personalized learning service strategies according to emotional feedback and learning experiences.

[0008] On the other hand, the present invention also provides an educational activity management method based on large model technology, including the following steps: Deploy data collection interfaces in each business system of the school to obtain original educational data; Clean, preprocess, annotate, and perform feature engineering on the collected data; Select a suitable large model, perform fine-tuning training using local annotated data, and set training parameters; After the training of each school node is completed, upload the model parameters to the transfer learning server, and the server updates the global model using a transfer learning algorithm based on the attention mechanism; Send the updated global model to each school node for educational data analysis.

[0009] Furthermore, it also includes steps for intelligent optimization of the educational activity process: Collect knowledge in the education field, and use natural language processing for entity extraction and relationship extraction; Convert the extracted knowledge into a semantic network and store it in a graph database to construct a semantic network for the educational activity process; Abstract the educational activity process into a directed graph and input it into the large model for analysis to mine optimization points; Query information in the semantic network according to the educational activity status, and combine with the intelligent recommendation function of the large model to provide an optimized activity organization plan and operation suggestions.

[0010] Furthermore, it also includes steps for personalized learning services: Integrate multi-modal interaction components such as voice emotion recognition, text emotion analysis, and facial expression recognition in the teaching platform and mobile applications; Perform multi-modal fusion processing on the student input data to accurately understand the student's needs and emotional state; Input the student's needs and emotional state into the large model to generate personalized learning recommendation content; Feed back the recommended content to the student, record the learning behavior and emotional feedback, and optimize the large model parameters and personalized learning service strategies.

[0011] Furthermore, the original educational data includes students' basic information, academic performance, and course arrangements.

[0012] Furthermore, the knowledge in the education field includes curriculum standards, teaching syllabuses, and education cases.

[0013] Furthermore, the multi-modal interaction components are used to collect students' voice, text, and expression information.

[0014] Beneficial effects: Efficient educational data analysis: Through the integration of transfer learning and large models, in-depth analysis and mining of educational data are achieved, breaking down the data barriers between schools, improving the data utilization efficiency, providing accurate data support for educational decision-making, and helping teachers and educational administrators better understand students' learning situations and needs.

[0015] Intelligent optimization of the educational activity process: Based on the algorithms of semantic networks and large models, it can deeply understand the logic and rules of educational activities, mine optimization points in the process, realize the intelligent management and optimization of educational activities, improve the efficiency of educational work, reduce educational costs, and make the educational resources more reasonably allocated.

[0016] High-quality personalized learning service: The personalized learning service system with emotional interaction and large models provides students with a convenient, efficient, and personalized learning experience, pays attention to students' emotional needs and learning status, enhances students' learning motivation and self-confidence, and promotes the all-round development of students.

[0017] Scientific education decision-making support: Utilize the powerful knowledge reasoning and data analysis capabilities of large models to provide multiple feasible solutions for education decision-making, predict and evaluate the effects of the solutions, assist education departments and schools in formulating scientific and reasonable education policies and teaching plans, and improve the scientificity and accuracy of education decision-making. Brief Description of the Drawings

[0018] Figure 1 System principle flowchart; Detailed Implementation Modes

[0019] The following further describes the implementation modes of the present invention in conjunction with the drawings and embodiments. The following embodiments are used to illustrate the present invention, but cannot be used to limit the scope of the present invention.

[0020] Embodiment 1 The object of the present invention is to provide an education activity management system and method based on large model technology, The system includes a data collection and access layer: Deploy a variety of data collection interfaces, support docking with the school's internal educational administration system, student management system, teaching resource library, etc., and at the same time be able to collect external education data, such as online education platform data, education data released by education research institutions, etc. Adopt technologies such as data encryption and desensitization to ensure the security and privacy of students' personal information and education data during the collection and transmission process. For example, obtain data such as students' learning achievements and attendance records through API interfaces, and use the SSL / TLS protocol for data transmission encryption.

[0021] Data processing and storage layer: Perform preprocessing operations such as cleaning, conversion, and integration on the collected data, remove noise and outliers, and unify the data format. Adopt distributed storage technology, such as the Ceph distributed storage system, to store massive education data, and establish a data index to improve data query efficiency. At the same time, use data warehouse technology to perform multidimensional analysis and management of the data, providing data support for subsequent intelligent analysis and decision-making.

[0022] Intelligent Analysis and Decision-making Layer: Based on the education data analysis platform integrating transfer learning and large models, the intelligent optimization algorithm for educational activity processes based on semantic networks and large models, and the personalized learning service system for emotional interaction and large models proposed in this invention, deeply analyze and mine educational data to provide intelligent support for educational activity decision-making. This layer also includes a model training and management module responsible for training, optimizing, and updating large models to adapt to changing educational needs.

[0023] User Interaction and Application Layer: Provide a friendly human-computer interaction interface for teachers, students, and educational administrators. Teachers can use this interface to formulate teaching plans, manage teaching resources, analyze students' learning situations, etc.; students can interact with the system through websites, mobile applications, etc. to obtain learning resources, participate in educational activities, and feedback learning problems, etc.; educational administrators can organize and arrange educational activities, allocate resources, analyze data, etc. The interface design follows the principles of simplicity and usability to improve user operation efficiency.

[0024] Construction of Transfer Learning Architecture: In view of the diversity and professionalism of educational data, build a transfer learning framework. Use large models pre-trained on large-scale general data and transfer their knowledge to the education field through transfer learning techniques. For example, adopt a transfer learning method based on the attention mechanism to quickly adapt to the specific tasks and data characteristics of the education field while maintaining the general knowledge of the large model.

[0025] Adaptation and Optimization of Large Models: Select large models suitable for educational data processing, such as pre-trained language models based on the GPT architecture, and fine-tune them according to the characteristics of the education field. Use professional corpora in the education field, including textbooks, academic papers, teaching cases, etc., to conduct supervised fine-tuning training on large models so that they can better understand and process educational data.

[0026] Data Analysis Process: First, clean, annotate, and perform feature engineering on educational data, and then input the processed data into a large model for local training. After training is completed, use transfer learning techniques to optimize the model for specific tasks in the education field. Through the analysis of students' learning data, such as learning achievements, learning time, learning behaviors, etc., provide support for students' learning situation assessment, learning progress prediction, etc.

[0027] Intelligent Optimization Algorithm for Educational Activity Processes Based on Semantic Networks and Large Models Construction of Educational Semantic Network: Collect and integrate various types of knowledge in the education field, including curriculum systems, teaching methods, educational policies, etc., to build a comprehensive educational semantic network. Through techniques such as entity extraction, relationship extraction, and semantic annotation, transform educational knowledge into structured semantic network data and store it in a graph database for convenient knowledge query and reasoning.

[0028] Activity Process Modeling and Optimization: Abstract the educational activity process into a directed graph, where nodes represent activity links and edges represent the logical relationships between links. Use large models to analyze the educational semantic network and identify bottlenecks and optimization points in the process. For example, through the inference ability of large models, it is found that more efficient teaching methods can be adopted for certain teaching links, or the allocation of certain educational resources can be more reasonable, thus improving the efficiency and quality of educational activities.

[0029] Intelligent Decision Support: When teachers or educational administrators organize educational activities, the system searches for relevant information in the semantic network based on the current activity status and provides them with the best activity organization plan and decision-making suggestions in combination with the intelligent recommendation function of the large model. For example, when arranging the class schedule, the system recommends the optimal class arrangement plan based on information such as teachers' teaching tasks, students' course requirements, and classroom resources.

[0030] Personalized Learning Service System Based on Emotional Interaction and Large Model Emotional Interaction Data Collection and Fusion: Integrate technologies such as speech emotion recognition, text emotion analysis, and facial expression recognition to achieve the collection and fusion of students' emotional interaction data. For example, during the learning process, the system can collect students' speech information through a voice interaction device and analyze the emotional tendency in their speech; collect students' facial expressions through a camera to judge their learning status and emotional changes; at the same time, conduct emotion analysis on students' text inputs on the learning platform. Integrate and process emotional data in different modalities to comprehensively understand students' emotional states.

[0031] Large Model-Driven Personalized Learning Recommendation: Utilize the powerful language understanding and generation capabilities of large models, combined with students' learning situations, interests, and emotional states, to provide students with personalized learning resource recommendations and learning path planning. The system first parses and semantically understands the input data of students, then retrieves relevant knowledge in the large model, and generates recommended content according to students' personalized needs. For example, for students who are interested in mathematics but have difficulties in a certain knowledge point, the system can recommend relevant mathematics learning materials, online courses, and problem-solving skills.

[0032] Learning Effect Evaluation and Feedback: Evaluate students' learning effects through multi-faceted data such as students' academic achievements, homework completion, and performance in educational activities. At the same time, adjust the personalized learning service strategy in a timely manner according to students' emotional feedback and learning experience to continuously improve students' learning effects and satisfaction. For example, if a student shows anxiety during the learning process, the system can appropriately adjust the learning progress and difficulty and provide more psychological counseling resources.

[0033] Deploy data collection interfaces in various school business systems, connect to the educational administration system, student management system, etc., and obtain original educational data, such as students' basic information, academic performance, course arrangements, etc.

[0034] Clean and preprocess the collected data to remove duplicate data, correct incorrect data, and fill in missing values. For example, use data cleaning tools to process incorrect scores and duplicate records in students' grades according to data quality rules.

[0035] Perform annotation and feature engineering on the preprocessed data to extract key features from the data. For example, quantify features such as students' study time and study frequency.

[0036] Select a suitable large model, such as the GPT-4 model, and perform fine-tuning training on the local node of the school using local annotated data. Set training parameters, such as a learning rate of 0.0001, 15 training epochs, and select Adagrad as the optimizer.

[0037] After the training of each school node is completed, upload the model parameters to the transfer learning server through an encrypted channel. The server uses a transfer learning algorithm based on the attention mechanism to fuse and optimize the parameters of each node and update the global model.

[0038] The transfer learning server distributes the updated global model to each school node, and each node uses the global model to analyze local educational data, such as predicting students' academic performance and assessing students' learning risks.

[0039] Collect various types of knowledge in the education field, including curriculum standards, teaching syllabuses, education cases, etc., and use natural language processing techniques to perform entity extraction and relationship extraction. For example, extract entities such as course names, teaching objectives, and teaching content from curriculum standards, as well as the association relationships between entities.

[0040] Convert the extracted knowledge into a semantic network and store it in the Neo4j graph database. Construct the nodes and edges of the semantic network. For example, use teaching links as nodes and the sequence between links as edges to establish a semantic network of the education activity process.

[0041] Abstract the education activity process into a directed graph and input it into the large model for analysis. Through the understanding and reasoning of the semantic network, the large model discovers optimization points in the education activity process, such as certain teaching links can be merged or the order can be adjusted.

[0042] When teachers or educational administrators organize educational activities, the system queries relevant information in the semantic network according to the current activity status and provides them with an optimized educational activity organization plan and operation suggestions in combination with the intelligent recommendation function of the large model. For example, when organizing a teaching seminar, the system recommends appropriate meeting processes and discussion topics based on information such as the theme of the seminar and the participants.

[0043] Integrate multi-modal interaction components such as speech emotion recognition, text emotion analysis, and facial expression recognition in the school's teaching platform and mobile applications. For example, students can input questions by voice, and the system uses speech emotion recognition technology to judge the emotional state of the students.

[0044] When students input learning needs or feedback learning problems, the system first performs multi-modal fusion processing on the input data, integrates information such as speech, text, and expressions, and accurately understands the needs and emotional states of the students. For example, when a student uploads a photo of their homework and asks about the problem-solving idea, the system obtains the homework content through image recognition and conducts comprehensive processing in combination with the text question and emotion analysis.

[0045] The system inputs the students' needs and emotional states into the large model, and the large model generates personalized learning recommendation content based on the pre-trained knowledge and fine-tuned educational domain knowledge. For example, for a student who has difficulties in Chinese writing and is in a low mood, the large model can recommend relevant writing skill tutorials, excellent essays, and provide some words of encouragement and psychological counseling.

[0046] The system feeds back the recommended content to the students and records the students' learning behaviors and emotional feedback. By analyzing the students' learning data and emotional feedback, continuously optimize the parameters of the large model and the personalized learning service strategy to improve the service quality. For example, if a student is not interested in a certain recommended content, the system analyzes the reasons, adjusts the recommendation algorithm, and improves the recommended content.

[0047] Select multiple schools in a certain region as experimental objects, including primary schools, middle schools, and universities, and deploy the educational activity management system proposed by the present invention in these schools.

[0048] Divide the experimental schools into an experimental group and a control group. The experimental group adopts the system and method of the present invention, and the control group adopts the traditional educational activity management method.

[0049] Determine the experimental indicators, including the improvement rate of students' learning achievements, the efficiency of educational activity organization, students' satisfaction, etc. The improvement rate of students' learning achievements is measured by comparing the exam scores of students before and after the experiment; the efficiency of educational activity organization is evaluated by statistics of indicators such as activity preparation time and resource utilization rate; students' satisfaction is collected by questionnaires and online evaluations to collect students' feedback.

[0050] Student learning achievement improvement rate: The average improvement rate of the learning achievements of the experimental group students after the experiment reached 20%, especially in major subjects such as mathematics and Chinese. For example, in the mathematics subject, the average score of the experimental group students increased by 10 points compared to the control group. Through personalized learning services, students can target their weak links for targeted learning, improving the learning effect.

[0051] Educational activity organization efficiency: The preparation time of educational activities in the experimental group was shortened by an average of 30%, and the resource utilization rate increased by 25%. For example, when organizing a campus culture activity, through the intelligent optimization of the educational activity process, reasonably arranging activity links and resource allocation, the preparation time of the activity was shortened from the original one week to four days, while reducing resource waste.

[0052] Student satisfaction: Through questionnaires and online evaluations, the student satisfaction of the experimental group reached over 90%, while that of the control group was only about 70%. The personalized learning service function of the student feedback system can meet their learning needs, improve their learning enthusiasm and initiative, and at the same time, the intelligent interaction function is convenient and fast, enhancing their communication and interaction with teachers and the system.

[0053] The experimental results show that the educational activity management system and method based on large model technology proposed in the present invention can significantly improve students' learning achievements, enhance the organization efficiency of educational activities, and increase students' satisfaction, having good application prospects and promotion value, providing strong technical support for the intelligent management of educational activities.

[0054] Example 2 Data collection and preprocessing: Collect data such as students' past course selection records, learning achievements, and interest survey results. Clean these data, remove missing values and incorrect data, and then perform normalization processing to unify different types of data into the same numerical range for subsequent analysis. For example, convert grades into values between 0 and 1, and digitally represent interests and hobbies through one-hot encoding and other methods.

[0055] Construct the student state space: Combine various characteristics of students into a state vector as the input of the reinforcement learning agent. The state vector includes information such as the courses selected by students, the current learning progress, the grade ranking, and interest preferences. For example, use a multi-dimensional array to represent the state vector, with each dimension corresponding to a characteristic.

[0056] Define the action space: The action space is the set of courses available for students to choose. Each course is an action, and the agent affects the learning state of students by selecting different course combinations.

[0057] Design the reward function: The reward function aims to measure the learning benefits after students select courses. For example, the reward is calculated comprehensively based on indicators such as subsequent academic performance improvement, course completion rate, and interest satisfaction of students. The greater the improvement in academic performance, the higher the course completion rate, and the higher the matching degree between the selected courses and interests, the higher the reward value.

[0058] Model training: Use reinforcement learning algorithms such as Deep Q-Network (DQN) for training. During the training process, the agent selects actions (courses) based on the current student state, observes the new state and the obtained reward after executing the action, and continuously updates the Q-value table or neural network parameters to optimize the strategy and maximize the long-term cumulative reward. After multiple iterative trainings, a policy model that can recommend the optimal course combination according to the student state is obtained.

[0059] Build a blockchain network: Select a suitable blockchain platform, such as Ethereum or a consortium blockchain. Each school acts as a blockchain node, deploys a blockchain client, and sets the public-private key pair of the node for data encryption and signature verification.

[0060] Data encryption and uploading to the blockchain: Encrypt data such as student grades and comprehensive quality evaluations, for example, using the AES symmetric encryption algorithm. Combine the encrypted data with metadata (such as the school to which the data belongs, data type, timestamp, etc.) into a transaction record. Each school node signs the transaction record and then broadcasts it to the blockchain network.

[0061] Write and deploy smart contracts: Write a smart contract for data access permission management using a smart contract programming language such as Solidity. The smart contract defines the access permission rules for different roles (teachers, education administrators, students, etc.). For example, teachers can view the grades of students in their taught classes, and education administrators can view the comprehensive data of all students in the school. Deploy the smart contract to the blockchain to obtain the contract address.

[0062] Data access and verification: When data needs to be accessed, the user sends an access request to the blockchain node, including user identity information and the identifier of the required data. The node verifies the user's identity and then calls the smart contract to query the access permission. If the permission is allowed, the node obtains the encrypted data from the blockchain, decrypts the data using the corresponding private key, and returns it to the user.

[0063] Select the teacher model and the student model: The teacher model is a complex large model trained on large-scale education data, such as a pre-trained language model based on the Transformer architecture. The student model selects a lightweight model with a simple structure and few parameters, such as a neural network structure like MobileNet.

[0064] Define the distillation loss function: The distillation loss function consists of two parts. One is the soft-label loss between the outputs of the student model and the teacher model, and the other is the hard-label loss between the output of the student model and the true label. The soft-label loss is measured by calculating the KL divergence of the output probability distributions of the student model and the teacher model, and the hard-label loss uses the cross-entropy loss function. For example, the distillation loss = α * soft-label loss + (1 - α) * hard-label loss, where α is a hyperparameter used to balance the weights of the two losses.

[0065] Knowledge distillation training: When training the student model, the same input data is input into both the teacher model and the student model simultaneously. The teacher model outputs soft labels (probability distributions), and the student model is trained based on the hard labels (true labels) and the soft labels of the teacher model. Through the backpropagation algorithm, the parameters of the student model are continuously adjusted to make its output as close as possible to the output of the teacher model while maintaining the ability to predict the true labels. After multiple rounds of training, the student model gradually learns the key knowledge of the teacher model and becomes a lightweight model that can operate efficiently on low-configured devices.

[0066] Agent modeling: Define multiple agents according to different links of educational activities. For example, the venue layout agent is responsible for planning the layout of the activity venue, and the program arrangement agent is responsible for determining the order and time of the activity programs. Each agent has its own goals, knowledge bases, and decision-making mechanisms. The knowledge base of the agent contains knowledge related to its own tasks. For example, the knowledge base of the venue layout agent stores information such as different venue sizes and equipment placement rules.

[0067] Establishment of communication mechanism: Design the communication protocol between agents. For example, use a message queue or a publish-subscribe mode. Agents share information and coordinate actions by exchanging messages. For example, after determining the program time, the program arrangement agent sends a message to the venue layout agent, informing the program duration and stage equipment requirements, and the venue layout agent adjusts the venue layout plan according to this information.

[0068] Task assignment and collaboration: The large model analyzes the overall process of the educational activity and formulates the overall task plan. Then the tasks are decomposed and assigned to each agent. Each agent makes autonomous decisions and executes tasks based on the assigned tasks and the information received from other agents. During the execution process, the agent continuously adjusts its actions according to new information to achieve the efficient collaborative organization of the educational activity. For example, when the time of a certain program is adjusted temporarily, the program arrangement agent promptly notifies the venue layout agent and other relevant agents, and each agent adjusts its plan accordingly.

[0069] Example 5: Generation of virtual teaching resources based on generative adversarial networks Data collection and preprocessing: Collect a large amount of art work data, including image data of paintings, audio data of music works, etc. Normalize the image data and adjust the pixel values to between 0 and 1; perform preprocessing operations such as sampling and framing on the audio data, and extract audio features such as Mel Frequency Cepstral Coefficients (MFCC).

[0070] Generator and discriminator design: The generator adopts a Convolutional Neural Network (CNN) or Recurrent Neural Network (RNN) structure. According to the information such as the style and theme of art works analyzed by the large model, it generates virtual art works. For example, for painting generation, the generator inputs a random noise vector and style and theme encoding vectors, and outputs a virtual painting image. The discriminator also adopts a CNN or RNN structure to judge the authenticity of the generated virtual works and real works.

[0071] Adversarial training: The generator and discriminator conduct adversarial training. The generator tries to generate realistic virtual works to deceive the discriminator; the discriminator endeavors to distinguish between real works and the generated virtual works. During the training process, the parameters of the generator and discriminator are updated alternately. The generator adjusts its generation strategy according to the discriminator's feedback to make the generated works more realistic; the discriminator continuously improves its discrimination ability based on the differences between real works and generated works. After multiple rounds of adversarial training, the generator can generate high-quality virtual teaching resources to meet the needs of art education.

Claims

1. A large-scale intelligent educational activity management system, characterized in that: include: The data collection and access layer deploys a variety of data collection interfaces to connect with the school's internal teaching system, student management system, teaching resource library, and external education data, and uses data encryption and desensitization technology to ensure data security and privacy; The data processing and storage layer cleans, converts, and integrates the collected data for preprocessing operations, uses distributed storage technology to store data and create indexes, and uses data warehouse technology for multi-dimensional analysis and management; Intelligent analysis and decision-making layer, based on the education data analysis platform integrating transfer learning and big models, the intelligent optimization algorithm of education activity process based on semantic network and big models, and the personalized learning service system based on emotional interaction and big models, it conducts in-depth analysis and mining of education data, provides intelligent decision-making support, and also includes model training and management modules; The user interaction and application layer provides a friendly human-computer interaction interface for teachers, students and educational administrators to meet the operational needs of different users.

2. The educational activity management system according to claim 1, characterized in that: The educational data analysis platform integrating transfer learning and large models includes: The transfer learning module is built to target the diversity and specialization of educational data. It uses a large model pre-trained on large-scale general data to transfer knowledge to the education field through a transfer learning method based on the attention mechanism. Large model adaptation and optimization module, selects large models suitable for educational data processing, pre-trains language models based on the GPT architecture, and uses professional corpora in the education field for supervised fine-tuning training; The data analysis process involves cleaning, labeling, and feature engineering the educational data before inputting it into a large model for local training. The model is then optimized using transfer learning technology for use in evaluating student learning situations and predicting learning progress.

3. The educational activity management system according to claim 1, characterized in that: The semantic network and large model educational activity process intelligent optimization algorithm includes: The educational semantic network construction module collects and integrates knowledge in the educational field, constructs structured semantic network data through entity extraction, relationship extraction and semantic annotation technology, and stores it in a graph database; The activity process modeling and optimization module abstracts the educational activity process into a directed graph, uses a large model to analyze the semantic network, and discovers bottlenecks and optimization points in the process; The intelligent decision support module searches for information in the semantic network according to the status of educational activities, and provides activity organization plans and decision suggestions in combination with the large model intelligent recommendation function.

4. The educational activity management system according to claim 1, characterized in that: The emotional interaction and large model personalized learning service system includes: Emotional interaction data collection and fusion module, which integrates speech emotion recognition, text emotion analysis, and facial expression recognition, and collects and integrates students' emotional interaction data; The personalized learning recommendation module driven by the big model uses the big model to combine students' learning situation, interests, hobbies and emotional state to provide personalized learning resource recommendations and learning path planning; The learning effect evaluation and feedback module evaluates students' learning effects through multi-faceted data and adjusts personalized learning service strategies based on emotional feedback and learning experience.

5. A method for managing educational activities based on a large model, characterized in that: The following steps are involved: Deploy data collection interfaces in various business systems of the school to obtain original education data; Clean, preprocess, annotate and feature engineer the collected data; Select a suitable large model, use local annotated data for fine-tuning training, and set training parameters; After the training of each school node is completed, the model parameters are uploaded to the transfer learning server, and the server uses the transfer learning algorithm based on the attention mechanism to update the global model; The updated global model will be sent to each school node for educational data analysis.

6. The educational activity management method according to claim 5, characterized in that: It also includes steps for intelligent optimization of educational activity processes: Collect knowledge in the field of education and use natural language processing for entity extraction and relationship extraction; The extracted knowledge is converted into a semantic network and stored in a graph database to construct a semantic network of educational activity processes; Abstract the educational activity process into a directed graph and input it into the big model for analysis to discover optimization points; Query information in the semantic network according to the status of educational activities, and provide optimized activity organization plans and operation suggestions in combination with the big model intelligent recommendation function.

7. The educational activity management method according to claim 5, characterized in that: It also includes personalized learning service steps: Integrate speech emotion recognition, text emotion analysis, and facial expression recognition multimodal interaction components into teaching platforms and mobile applications; Perform multimodal fusion processing on student input data to accurately understand students' needs and emotional states; Input students’ needs and emotional states into the big model to generate personalized learning recommendations; Feedback recommended content to students, record learning behaviors and emotional feedback, and optimize large model parameters and personalized learning service strategies.

8. The educational activity management method according to claim 5, characterized in that: The original education data includes students' basic information, academic performance, and course schedule.

9. The educational activity management method according to claim 6, characterized in that: The education field knowledge includes curriculum standards, teaching syllabuses, and education cases.

10. The educational activity management method according to claim 7, characterized in that: The multimodal interaction component is used to collect students' voice, text, and expression information.

Images