CN119151744A

CN119151744A - Art teaching management system and method based on AI

Info

Publication number: CN119151744A
Application number: CN202411308179.7A
Authority: CN
Inventors: 黄志勇
Original assignee: Nanjing Jinxi Aesthetic Education Software Technology Co ltd
Current assignee: Nanjing Jinxi Aesthetic Education Software Technology Co ltd
Priority date: 2024-09-19
Filing date: 2024-09-19
Publication date: 2024-12-17

Abstract

The present invention discloses an AI-based art teaching management system and method thereof, including an intelligent user management module, a course generation module, a real-time interaction module, an evaluation module, an interactive control module and a data security module. Through the multimodal deep integration learning algorithm MMDILA and the reinforcement learning algorithm, the multimodal data of the user is analyzed to realize the dynamic generation and adjustment of personalized course content, and provide real-time learning feedback and personalized suggestions. Blockchain technology and differential privacy technology are used to ensure the security and privacy of user data. The present invention significantly improves the personalized teaching effect and learning efficiency, shortens the learning time, enhances user satisfaction, and greatly improves data security and privacy protection capabilities, overcoming the shortcomings of the prior art in personalization, real-time feedback and data security. The present invention is significantly superior to the prior art in personalized teaching and data security, and can meet the needs of different users.

Description

Art teaching management system and method based on AI

Technical Field

The invention relates to the technical field of chip testing, in particular to an AI-based art teaching management system and a method thereof.

Background

In the conventional art teaching management system, a common method is mainly course design based on fixed learning content and linear teaching paths. These systems often lack the ability to analyze and adjust in real time the learning behavior, progress and performance of the individual users, making it difficult for the teaching content to dynamically adapt to the learning needs and interest changes of the users. Furthermore, most systems lack an effective user data security protection mechanism, which is prone to privacy disclosure and data security risks when handling and storing user data.

There are several major problems with art teaching management systems in that, first, these systems typically use static and predefined learning paths, lack the ability to respond to the user's personalized needs, and are unable to adjust course content and teaching strategies based on the user's specific performance and feedback. Secondly, the interaction and feedback mechanisms in these systems are limited, and multi-modal learning data (such as text, images, voice, etc.) of the user cannot be captured and processed in real time, thereby affecting the effect of personalized teaching. Furthermore, existing systems have inadequate data security measures, especially when related to user sensitive data, lack of effective encryption, privacy protection, and anomaly detection mechanisms, which can lead to data leakage and abuse.

Disclosure of Invention

Aiming at the problems of limited individualized learning capability, lack of real-time interaction and feedback mechanism, insufficient data security and privacy protection measures and the like, the invention provides an art teaching management system and a method based on AI, which optimize functions in aspects of course generation, interaction feedback, data security and the like by introducing a plurality of advanced AI technologies such as deep learning, natural language processing, block chain, differential privacy and the like. The invention can dynamically analyze the multi-mode learning data of the user, automatically generate and adjust personalized course content, provide learning feedback in real time, and ensure the safety and privacy of the user data through various data protection technologies (such as block chain and differential privacy technology).

In order to achieve the above object, the present invention is realized by the following technical scheme:

an AI-based art teaching management system, the system comprising:

The intelligent user management module is used for identifying the identity of a user, dynamically adjusting the allocation and authority configuration of learning resources by adopting a Support Vector Machine (SVM) algorithm according to learning behavior data and history records of the user, and optimizing the preprocessing and analysis flow of the data by carrying out unified modeling, feature selection and data fusion on the multi-mode data based on an improved multi-mode deep integrated learning algorithm (MMDILA);

The course generation module is used for forming personalized course content through a Convolutional Neural Network (CNN) model and a random forest algorithm based on the learning progress and the performance of the user;

The real-time interaction module analyzes the interaction content of the user through a Convolutional Neural Network (CNN) and Natural Language Processing (NLP) technology and generates real-time learning feedback and personalized suggestions;

the evaluation module is used for automatically evaluating works submitted by users by adopting a reinforcement learning algorithm and generating feedback;

The interaction control module optimizes the learning interface by utilizing voice recognition, gesture recognition and image recognition technologies;

and the data security module is used for protecting user data by combining a block chain and a differential privacy technology and monitoring data access behaviors in real time.

Preferably, the intelligent user management module further comprises:

the user identification unit is used for carrying out efficient and accurate identity identification on a user by adopting a Deep Neural Network (DNN) AI algorithm and providing personalized access right control;

the data analysis unit is used for comprehensively analyzing multi-mode learning data (including texts, images and voices) to generate personalized learning paths and progress plans which accord with the characteristics of users;

the resource allocation unit is used for carrying out deep modeling and sequence analysis on the user data through a Graph Neural Network (GNN) and a transducer model, and dynamically adjusting the allocation of learning resources;

Wherein the resource allocation unit further models and analyzes sequences of user data through a Graph Neural Network (GNN) and a transducer model, realizes resource optimization configuration by combining element learning and an adaptive learning rate algorithm,

The specific implementation steps are as follows:

step 1, modeling a relation network among users by combining multi-modal learning (Multimodal Learning) with a graph neural network (Graph Neural Network, GNN), carrying out fusion analysis on multi-modal data of the users to generate user characteristic expression, wherein a calculation formula is as follows:

Wherein, Representing a characteristic representation of node v at layer k+1, σ being a nonlinear activation function, N (v) being a set of neighbor nodes of node v,

C _vu is a normalization factor, and W ^(k) is a learnable weight matrix of a k-th layer;

Step 2, using a transducer model based on a self-attention mechanism to carry out sequence modeling on historical behavior data of a user, generating a personalized learning path, optimizing parameters through meta-learning and a self-adaptive learning rate algorithm, and calculating the following formula:

Wherein Q is a query matrix, K is a key matrix, V is a value matrix, Is the dimension of the key vector;

Through Meta-Learning (Meta-Learning) and adaptive Learning rate algorithms (e.g., adagrad or RMSProp), the Learning path is dynamically adjusted, optimizing the parameter θ:

where Gt is the cumulative gradient squared and ε is the smoothing term;

And 3, initially modeling a user feature matrix and a resource utility matrix by using Non-negative matrix factorization (Non-negative Matrix Factorization, NMF), and further optimizing a resource allocation strategy by using a Diffusion model (Diffusion Models), wherein the calculation formula is as follows:

Wherein p theta (x _t-1∣x_t) generates the conditional probability of the previous state x _t-1 for a given current state x _t, mu theta (x _t, t) and The mean value and variance of the neural network prediction are calculated, and I is an identity matrix;

Step 4, dynamically adjusting learning content and feedback expression modes by combining a real-time emotion analysis technology, wherein a calculation formula is as follows;

wherein Q (s, a) is a state-action value function, s is the current state, a is the current action, alpha is the learning rate, r is the instant prize, gamma is the discount factor, a 'is the possible future action, s' is the next state;

And 5, protecting the user data by adopting a differential privacy (DIFFERENTIAL PRIVACY) and federal learning (FEDERATED LEARNING) method.

Preferably, the course generating module includes:

the content generation unit automatically generates diversified personalized course content according to the learning needs and interests of the user by adopting an improved Convolutional Neural Network (CNN) and a random gradient descent (SGD) optimization algorithm;

The data fusion unit is used for efficiently integrating multi-mode learning data (text, image and video), carrying out data processing and feature extraction through a deep learning technology, and generating course content conforming to the learning style of a user;

and the dynamic adjustment unit is used for intelligently adjusting course difficulty, teaching strategies and content presentation modes by using a deep reinforcement learning algorithm based on the learning progress and real-time performance of the user so as to adapt to the change of the learning state of the user.

Preferably, the real-time interaction module includes:

the interactive analysis unit is used for extracting interactive text features through a word embedding method based on a Bi-directional long-short-term memory network (Bi-LSTM) and a Convolutional Neural Network (CNN);

the emotion recognition unit fuses the extracted text features with the voice and visual data of the user through a multi-mode emotion fusion network;

and the feedback generation unit is used for generating personalized learning feedback based on the transducer model and dynamically adjusting a feedback strategy through a reinforcement learning algorithm.

Preferably, the evaluation module includes:

the multi-dimensional evaluation unit is used for automatically evaluating and analyzing the artwork submitted by the user from a plurality of aspects of composition, color, skill and creative based on computer vision technology and machine learning algorithm;

The self-adaptive evaluation engine unit utilizes the historical performance data and real-time feedback of the user to realize personalized evaluation standards by dynamically adjusting the weight of each evaluation dimension;

and the safety optimization unit is used for optimizing evaluation standards and strategies by adopting a depth reinforcement learning algorithm DQN.

Preferably, the interaction control module includes:

a recognition unit recognizing voice, gesture, and image input of a user using a Deep Neural Network (DNN) model, and converting the instruction into an operation command;

And the interface optimization unit is used for adjusting the layout, the content presentation sequence and the visual effect of the learning interface in real time according to the identification result.

Preferably, the data security module includes:

The encryption unit is used for encrypting and distributively storing the user data by combining an advanced encryption standard (AES-256) algorithm with a distributed blockchain technology, so that the integrity and the safety of the data are ensured;

The privacy protection unit is used for protecting user data based on a differential privacy technology and preventing sensitive information from being revealed in the data analysis process;

And the monitoring unit is used for monitoring and detecting data access behaviors in real time by using a behavior analysis algorithm and a zero knowledge proving method, identifying abnormal access and supporting multiparty cooperation based on federal learning.

An AI-based art teaching management method applied to a computing device, the method comprising the following steps:

Step S1, collecting multi-modal learning data of a user, including but not limited to text, image, voice and video data;

The text data is subjected to word segmentation and syntactic analysis by a natural language processing technology, the image data is acquired by an image sensor with the resolution of 256x256 pixels and is stored in a JPEG format, the voice data is stored in a WAV format at the sampling rate of 16kHz, the video data is coded in an MP4 format at the frame rate of 30 fps;

S2, performing feature extraction and fusion on the acquired multi-mode data through a multi-mode deep integration learning algorithm MMDILA, wherein a convolutional neural network is used for performing feature extraction on image data, a two-way long-short-time memory network is used for performing semantic analysis on text data, an emotion recognition model is used for processing voice data, and information of different data modes is fused through a transducer model of a self-attention mechanism to generate a comprehensive learning image of a user;

And step S3, dynamically adjusting personalized course content and teaching strategies based on the user learning portrait through a reinforcement learning algorithm, wherein the reinforcement learning algorithm is based on a strategy gradient method, the learning rate is 0.01, the discount factor is 0.9, and the optimization aim is to maximize the user learning effect and satisfaction.

Preferably, the integration process of the multi-level deep learning model further comprises the following steps:

The integration process of the multi-level deep learning model further comprises the following steps:

step 1, through a multi-task learning framework, the outputs of a CNN model and a Bi-LSTM model are subjected to joint training in a sharing layer;

step 2, pre-training the multi-mode interaction data of the user by applying a self-supervision learning technology;

And 3, adopting a self-attention mechanism to dynamically adjust the importance of different data modes.

Preferably, the step of dynamically adjusting personalized course content and teaching strategies includes:

Step 1, predicting user behavior by utilizing an improved multi-mode deep integrated learning algorithm (MMDILA) and an antagonism generation network (GAN) model;

Step 2, simulating potential behaviors of a user through a training generator, and distinguishing the generated behaviors from actual behaviors by a training discriminator;

And 3, based on the result of the resistance training, adjusting a learning path and a feedback mechanism of the user in real time.

Term interpretation:

The multimodal deep learning algorithm MMDILA refers to a deep learning method that can process multiple types of data (e.g., text, image, voice, and video) simultaneously. The method comprises the steps of integrating the characteristics of various data finally to generate personalized learning paths and suggestions of a user through fusing a Convolutional Neural Network (CNN) for image characteristic extraction, natural Language Processing (NLP) for text analysis and emotion recognition models for voice processing. MMDILA is innovative in that the characteristics of various data sources are deeply integrated, and the understanding and response capability of the system to the learning behaviors of the user are improved.

The reinforcement learning algorithm is a machine learning method used for dynamically adjusting the learning path and teaching strategy of the user. And obtaining feedback data to evaluate the effect of the current strategy through continuous interaction with the user. The system sets an objective function and rewarding mechanism to guide the learning process towards optimization of the best strategy direction. Specific implementations of the algorithm include models such as Q-learning or deep Q-network (DQN) to maximize learning and satisfaction of users through continuous optimization strategies.

And the block chain technology is used for guaranteeing the safety and transparency of user data. All data access and operation behaviors are recorded through the distributed account book, so that the non-tamper property and traceability of data are ensured. The invention adopts a alliance chain structure, combines a multiparty consensus mechanism, records and monitors the data access process, and ensures that all data operations in the system are safe and verifiable. The technology is further integrated in a data security module and works in conjunction with differential privacy technology.

The differential privacy technology is a statistical method for protecting the privacy of user data, and the unrecognizable property of individual information in data analysis is ensured by adding noise in a data set. The technique prevents leakage of sensitive information during data analysis and model training. The invention utilizes the differential privacy technology to ensure that the output result does not reveal the specific information of any specific user when analyzing and processing the user data. The system controls the amount of noise added by setting a privacy budget to balance data availability and privacy protection.

The user portrayal generation means that the system generates the comprehensive learning characteristics and the behavior portrayal of the user through multi-mode data fusion and machine learning technology. Specifically, the invention carries out fusion processing on multi-mode data (text, image and voice) through a transducer model, extracts learning habit, preference and performance characteristics of a user, and forms personalized user portraits as the basis for course content generation and adjustment.

The real-time interaction module comprises an interaction analysis unit, an emotion recognition unit and a feedback generation unit and is used for carrying out real-time interaction and feedback with a user through multi-mode data (voice, gestures, texts and the like). The interactive analysis unit is responsible for analyzing user input data in real time, the emotion recognition unit analyzes the emotion state of the user, and the feedback generation unit provides instant personalized chemistry Xi Jian and course adjustment according to the analysis result so as to improve learning effect and user experience.

The multi-dimensional evaluation unit and the self-adaptive evaluation engine are used for comprehensively evaluating the learning effect of the user according to a plurality of indexes (such as learning progress, accuracy, participation degree and the like). The self-adaptive evaluation engine dynamically adjusts the learning path and course content based on the result of the multi-dimensional evaluation unit so as to realize the optimal learning effect. The engine combines a reinforcement learning algorithm, continuously learns and optimizes the evaluation standard and strategy, and ensures the accuracy and reliability of the evaluation result.

Compared with the prior art, the invention has the beneficial effects that:

1. The invention obviously improves the personalized teaching effect by introducing the multi-mode deep integrated learning algorithm (MMDILA) and the improved Support Vector Machine (SVM) algorithm, thereby obviously enhancing the response capability of the system to the personalized learning requirement of the user. By analyzing the multi-mode data (such as text, voice, images and the like) of the user, the invention can dynamically generate and adjust personalized course content and adapt to the learning progress and the performance of the user in real time. The innovative method overcomes the defect of fixed learning path in the prior art, so that the teaching process is more flexible and effective, and the learning effect and the user experience are improved.

2. Optimizing real-time interaction and feedback mechanism by utilizing Convolutional Neural Network (CNN) and Natural Language Processing (NLP) technology, the invention realizes real-time interaction and feedback with users. The system can generate instant feedback and personalized suggestions according to real-time input (such as voice instructions, gesture operations and the like) of the user, and automatically evaluate the learning effect of the user through a reinforcement learning algorithm. The mechanism not only enhances the interactivity and the interestingness of teaching, but also greatly improves the teaching efficiency, overcomes the defect of lack of real-time feedback in the prior art, and remarkably improves the participation feeling and satisfaction of users.

3. The invention innovatively combines the blockchain technology and the differential privacy technology, and effectively protects the safety and the privacy of user data. The block chain technology is used for recording and monitoring the access behavior of the data, ensuring the transparency and the non-falsification of the data operation, and the differential privacy technology is used for protecting the sensitive data of the user and preventing the sensitive data from being revealed in the data analysis and use processes. Compared with the prior art, the invention obviously improves the safety in the data storage and transmission process, makes breakthrough progress in the aspects of anomaly detection and data access monitoring, and comprehensively solves the defects of the prior art in the aspect of data safety protection.

4. The invention can efficiently fuse and process learning information (such as voice, image and text input of a user) from a plurality of data sources through an improved multi-modal deep integration learning algorithm (MMDILA) and a Convolutional Neural Network (CNN). The innovative data fusion method remarkably improves the accuracy and speed of data analysis, so that the system can know the learning habit and preference of the user faster and more accurately, and a more accurate personalized learning path is generated. This improvement over existing data processing methods significantly enhances the level of intelligence and the adaptive capabilities of the system.

5. The invention adopts modularized system architecture design, and each functional module (such as user management, course generation, interactive feedback, data safety and the like) is mutually independent and mutually coordinated, so that the system has high expansibility and compatibility in subsequent development and application. Compared with the prior art, the innovative architecture design not only facilitates the expansion and updating of functions, but also reduces the maintenance cost of the system, is beneficial to wider application and popularization, and improves the market competitiveness and commercialization potential of the invention.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings that are needed in the description of the embodiments will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

Wherein:

FIG. 1 is a system architecture diagram of an embodiment of the present invention;

FIG. 2 is a schematic diagram of an intelligent user management module according to an embodiment of the present invention;

FIG. 3 is a schematic diagram of a course generating module according to an embodiment of the present invention;

FIG. 4 is a schematic diagram of a real-time interaction module according to an embodiment of the present invention;

FIG. 5 is a schematic diagram of an evaluation module according to an embodiment of the invention;

FIG. 6 is a flow chart of a method according to an embodiment of the invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the embodiments of the present invention more clear, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. It will be apparent that the described embodiments are some, but not all, embodiments of the invention. All other embodiments, which are obtained by a person skilled in the art based on the described embodiments of the invention, fall within the scope of protection of the invention.

Example 1

As shown in fig. 1, an embodiment of the present invention provides an AI-based art teaching management system, including:

The intelligent user management module has the core functions of identifying the identity of the user and dynamically adjusting the allocation and authority configuration of learning resources according to learning behavior data and history records of the user. The module analyzes the behavior data of the user by using a Support Vector Machine (SVM) algorithm, and ensures that the system can flexibly respond to different requirements of the user.

The module optimizes the preprocessing and analysis flow of data by uniformly modeling, feature selection, data fusion, and the like, of multi-modal data (such as text, images, voice, and the like) based on an improved multi-modal deep integrated learning algorithm (MMDILA). MMDILA is improved by integrating features of multiple data types, enabling the system to more precisely understand the learning behavior and preferences of the user, thereby significantly optimizing the prior art in terms of the diversity and complexity of data entry.

Specifically, as shown in fig. 2, the intelligent user management module further includes:

The specific implementation steps are as follows:

where Gt is the cumulative gradient squared and ε is the smoothing term;

Further, as shown in fig. 3, the course generation module personalizes course content based on learning progress and performance of the user through a Convolutional Neural Network (CNN) model and a random forest algorithm;

The course generation module is responsible for generating personalized course content and providing dynamically adjusted learning materials for the user according to the learning progress and performance of the user. By analyzing the user's performance data (e.g., learning speed, accuracy, interaction, etc.), the system can continuously optimize the course content.

The module uses a Convolutional Neural Network (CNN) model to perform feature extraction on the user's image and visual data, and analyzes the user's performance in combination with a random forest algorithm to generate personalized lesson content. The random forest algorithm is used for evaluating the learning effect of the user in the process, and selecting proper learning paths and materials according to different characteristics, so that the user can learn in an optimal mode.

Specifically, the course generation module includes:

Further, as shown in fig. 4, the real-time interaction module analyzes the interaction content of the user through Convolutional Neural Network (CNN) and Natural Language Processing (NLP) technologies, and generates real-time learning feedback and personalized advice;

the real-time interaction module is used for analyzing the interaction content of the user in the learning process and generating real-time learning feedback and personalized suggestions so as to improve the user participation degree and the learning effect.

The module analyzes multimodal data, such as speech, text, and gestures, entered by a user using Convolutional Neural Network (CNN) and Natural Language Processing (NLP) techniques. The CNN model is used to process image and gesture data of the user, and the NLP technique is used to parse text and voice inputs of the user. Through the combination of the technologies, the system can capture the learning state and emotion change of the user in real time, generate adaptive learning suggestions, and improve the effectiveness and interestingness of learning interaction.

Specifically, the real-time interaction module includes:

Further, as shown in fig. 5, the evaluation module adopts a reinforcement learning algorithm to automatically evaluate the work submitted by the user and generate feedback;

the assessment module is responsible for automatically assessing works submitted by users and generating feedback so as to help the users to know learning progress and performance in time.

The module adopts a reinforcement learning algorithm to automatically evaluate learning works submitted by users. The reinforcement learning algorithm sets a reward mechanism according to the quality, innovation and complexity of the user work, and generates more accurate and personalized feedback by continuously optimizing the evaluation model. The evaluation method overcomes subjectivity of the traditional evaluation mode and provides more objective and scientific evaluation results.

Specifically, the evaluation module includes:

Further, the interaction control module optimizes a learning interface by utilizing voice recognition, gesture recognition and image recognition technologies;

the interaction control module optimizes the interaction interface between the user and the system, and ensures that the user can learn by using the system naturally and smoothly.

The module utilizes speech recognition, gesture recognition and image recognition techniques to create a more intuitive and human-friendly user interface. The voice recognition technology is used for recognizing voice commands of users, the gesture recognition technology is used for detecting gesture operations of the users, the image recognition technology is used for processing learning materials and works uploaded by the users, the system is ensured to be capable of accurately understanding and responding to the operations of the users, and user experience is improved.

Specifically, the interaction control module includes:

Further, the data security module is used for protecting user data by combining a block chain and a differential privacy technology and monitoring data access behaviors in real time.

The data security module is responsible for protecting the security and privacy of user data, and simultaneously monitoring the data access behavior in real time to prevent data leakage and unauthorized access.

The module combines blockchain technology and differential privacy technology to provide multiple data protection measures. Blockchain technology is used to record and monitor all data access behavior, ensuring transparency and non-tamper-ability of data. The differential privacy technology adds random noise in the data analysis process, so that sensitive information of a user is protected from being revealed. In addition, the data security module also integrates an abnormal behavior detection mechanism, and data access behaviors in the system are monitored in real time, so that potential data leakage and illegal access are prevented.

Specifically, the data security module includes:

Example 2

As shown in fig. 6, an embodiment of the present invention provides an AI-based art teaching management method applied to a computing device, the method including the steps of:

Further, the integration process of the multi-level deep learning model further comprises the following steps:

Further, the step of dynamically adjusting the personalized course content and the teaching strategy includes:

step 1, predicting user behavior by utilizing an improved multi-mode deep integrated learning algorithm MMDILA and an antagonism generation network model;

Example 3

The embodiment realizes a more intelligent and personalized art teaching management system through an improved multi-mode deep integrated learning algorithm (MMDILA) and a reinforcement learning algorithm. The invention can dynamically analyze various learning behaviors and data of the user, generate course content meeting individual requirements and adjust learning paths and strategies in real time. Meanwhile, the invention ensures the safety and privacy of the user data through a plurality of advanced data security technologies.

In the prior art, art teaching systems generally use only a single data source (such as text or simple learning progress data) for course generation, and lack the capability of fusion processing of multi-modal data. The depth and breadth of data fusion and analysis are greatly improved through an improved multi-mode deep integration learning algorithm (MMDILA).

The invention can process various types of data such as text, image, voice, video and the like at the same time, extract image characteristics through a Convolutional Neural Network (CNN), carry out semantic analysis through a Natural Language Processing (NLP) technology, and process voice data through an emotion recognition model. Through the fusion of the multi-mode data, the system can more accurately understand the learning habit, interest and emotion state of the user, so that more personalized learning content is generated. Compared with the prior art, the method and the device can more comprehensively capture the learning behaviors and preferences of the user, and effectively improve the pertinence and the effectiveness of personalized teaching.

The existing art teaching management system lacks a real-time interactive feedback mechanism, generally depends on fixed rules or a simple scoring system, and cannot adjust a learning path according to instant feedback of a user. The invention introduces a reinforcement learning algorithm to realize the real-time monitoring and dynamic optimization of the learning behavior of the user.

According to the invention, the learning effect and behavior of the user are evaluated in real time through the reinforcement learning algorithm, and the system is continuously self-adjusted to improve the learning effect. For example, when a user presents difficulty at a certain knowledge point, the system may automatically increase learning resources for that portion of content or adjust teaching strategies. Through the dynamic optimization mechanism, the system can generate a personalized path which better meets the learning requirement of the user according to the feedback and the performance of the user.

Compared with the prior art, the real-time feedback capability and the dynamic optimization mechanism of the invention remarkably improve the teaching effect and the user satisfaction, greatly reduce the learning time and increase the learning investment of the user.

In the prior art, user data security measures of art teaching systems are based on simple encryption or access control. On the basis of the method, the block chain technology and the differential privacy technology are combined, so that the data security and privacy protection capability are greatly enhanced. The invention adopts the blockchain technology to record and monitor the data access behavior, and ensures the transparency and the non-tamper property of the data operation. Meanwhile, sensitive data of a user is protected by utilizing a differential privacy technology, and personal privacy is prevented from being revealed in the data analysis and use process. In addition, the system integrates a behavior anomaly detection model, and potential data leakage and illegal access can be monitored and detected in real time. Compared with the prior art, the invention not only provides higher security in the data storage and transmission process, but also makes breakthrough progress in the aspects of privacy protection and data anomaly monitoring, so that the user data is safer and more reliable.

According to the embodiment, the personalized teaching capability, the real-time feedback mechanism and the data security of the conventional art teaching management system are obviously improved by introducing a multi-mode deep integrated learning algorithm, a reinforcement learning algorithm and various data security technologies. Through the innovative technical improvements, the invention realizes comprehensive improvement in the aspects of user experience, learning effect and data protection.

Example 4

In order to verify the remarkable advantages of the invention in the aspects of personalized teaching and data security, two groups of experiments are carried out, namely, firstly, the personalized teaching effect of the system is compared with that of the traditional art teaching system, and secondly, the performance of the data security module in the aspects of preventing data leakage and privacy protection is verified.

Design of experiment

The experimental samples were selected from 100 participants, and divided into an experimental group and a control group, each group having 50 persons. The experimental group uses the AI art teaching management system of the invention, and the control group uses the traditional art teaching management system.

The experimental period is 4 weeks.

And the test indexes comprise learning effect scoring, learning time, user satisfaction, data leakage detection success rate and privacy protection rate.

Experimental results

Index (I)	Experimental group (inventive system)	Control group (traditional system)	Rate of rise
				Learning effect score (full 100)	85.2	63.1	35.00%
Average learning time (hours)	28.6	35.8	-20.10%
				User satisfaction (full 5)	4.7	3.5	34.30%
Success rate of data leakage detection (secondary)	20/20	12 Months and 20 days	66.70%
				Privacy protection (percentage)	98.50%	85.00%	15.90%

The experimental data show that the learning effect score is 85.2 points in the 4-week learning period of the experimental group, which is obviously higher than 63.1 points of the control group, so that the system of the invention obviously improves the personalized teaching effect.

Average learning time the average learning time of the experimental group was 28.6 hours, which was 20.1% less than 35.8 hours of the control group. This shows that the inventive system effectively reduces the time required for learning through real-time feedback and personalized path optimization.

User satisfaction-Experimental groups using the system of the invention user satisfaction scores 4.7 points, 3.5 points higher than control groups. The result shows that the system of the invention provides better user experience and learning participation.

The success rate of data leakage detection is that in 20 times of simulated attack tests, the system successfully detects all 20 times of data leakage attempts, the success rate is 100 percent, and the success rate of comparison components is 60 percent (12/20). The system of the invention has stronger safety protection capability.

The privacy protection rate of the system is 98.5%, which is obviously better than 85.0% of the control group, and the system is excellent in protecting the user data privacy.

Experimental data shows that the AI art teaching management system is remarkably superior to the prior art in the aspects of personalized teaching effect, learning time optimization, user satisfaction improvement, data safety and privacy protection.

In summary, the invention innovatively uses an improved multi-modal deep ensemble learning algorithm (MMDILA) to significantly improve data fusion and analysis capabilities. Unlike the prior art, the present invention is capable of simultaneously processing multiple data types (such as text, image, voice and video) of a user and efficiently analyzing the data through Convolutional Neural Network (CNN), natural Language Processing (NLP) and emotion recognition model. The innovative data processing mode realizes more comprehensive and more accurate user behavior understanding, so that the invention can generate and dynamically adjust the learning path according to the personalized requirements of the user, and the effect and pertinence of personalized teaching are obviously improved.

The invention introduces a reinforcement learning algorithm, establishes a real-time feedback and dynamic optimization mechanism, and can continuously adjust course content and teaching strategies according to real-time performance and feedback of users. Compared with the feedback mode relying on fixed rules or simple scores in the prior art, the method can provide more instant and accurate personalized suggestions, so that the learning time is effectively shortened, and the learning efficiency is improved. The innovative interaction mechanism enhances the user experience and increases the learning investment and satisfaction of the user.

The invention provides a brand new data security and privacy protection scheme through an integrated block chain technology and a differential privacy technology. The block chain technology ensures transparency and non-tamper property of data operation, and the differential privacy technology effectively prevents leakage of sensitive information of users in the data analysis and use processes. In addition, the invention also introduces an abnormality detection model, and can monitor and cope with potential data leakage and illegal access risks in real time. Compared with the single encryption or access control measures in the prior art, the method and the device realize higher level protection in terms of data security and privacy protection.

The invention adopts a modularized system architecture design, and each functional module (such as user management, course generation, interactive feedback, data security and the like) works independently and cooperatively. The architecture design not only improves the expansibility and compatibility of the system and is convenient for subsequent function development and system upgrading, but also reduces the maintenance cost and complexity of the system. Compared with a relatively closed and inflexible system architecture in the prior art, the design of the invention can better adapt to different user requirements and use scenes, and has wide market application potential.

The invention verifies the remarkable advantages of the invention in the aspects of personalized teaching effect, learning time optimization, user satisfaction improvement, data safety and privacy protection through a large amount of experimental data. Experimental results show that the personalized teaching effect score of the system is improved by 35%, the learning time is reduced by 20%, the user satisfaction is increased by 34%, and the system is excellent in data leakage detection and privacy protection.

In the description of the present specification, a description referring to terms "one embodiment," "some embodiments," "examples," "specific examples," or "some examples," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present application. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, the different embodiments or examples described in this specification and the features of the different embodiments or examples may be combined and combined by those skilled in the art without contradiction.

Any process or method description in a flowchart or otherwise described herein may be understood as representing modules, segments, or portions of code which include one or more executable instructions for implementing specific logical functions or steps of the process. And the scope of the preferred embodiments of the present application includes additional implementations in which functions may be performed in a substantially simultaneous manner or in an opposite order from that shown or discussed, including in accordance with the functions that are involved.

The foregoing is merely illustrative of the present application, and the present application is not limited thereto, and any person skilled in the art will readily recognize that various changes and substitutions are possible within the scope of the present application. Therefore, the protection scope of the application is subject to the protection scope of the claims.

Claims

1. An AI-based art teaching management system, characterized in that the system includes:

An intelligent user management module is used to identify the user identity and dynamically adjust the allocation of learning resources and permission configuration using a support vector machine (SVM) algorithm based on the user's learning behavior data and historical records. The intelligent user management module is based on an improved multimodal deep integration learning algorithm MMDILA, which optimizes the data preprocessing and analysis process by performing unified modeling, feature selection, and data fusion on multimodal data;

The course generation module generates personalized course content based on the user's learning progress and performance through the convolutional neural network (CNN) model and random forest algorithm;

The real-time interaction module uses convolutional neural networks (CNN) and natural language processing (NLP) to analyze user interaction content and generate real-time learning feedback and personalized suggestions.

The evaluation module uses reinforcement learning algorithms to automatically evaluate user-submitted works and generate feedback;

Interactive control module, which uses speech recognition, gesture recognition and image recognition technologies to optimize the learning interface;

The data security module combines blockchain and differential privacy technology to protect user data and monitor data access behavior in real time.

2. The AI-based art teaching management system according to claim 1, characterized in that the intelligent user management module further comprises:

The user identification unit uses a deep neural network AI algorithm to efficiently and accurately identify users and provide personalized access rights control;

The data analysis unit uses multimodal learning data, including text, images, and voice, for comprehensive analysis to generate personalized learning paths and progress plans that meet user characteristics;

The resource allocation unit uses graph neural networks and Transformer models to perform deep modeling and sequence analysis on user data and dynamically adjust the allocation of learning resources;

The resource allocation unit further models and performs sequence analysis on user data through graph neural networks and Transformer models, and combines meta-learning and adaptive learning rate algorithms to achieve resource optimization configuration. The specific implementation steps are as follows:

Step 1: Use multimodal learning combined with graph neural networks to model the relationship network between users, perform fusion analysis on the multimodal data of users, and generate user feature representation. The calculation formula is as follows:

in, represents the feature representation of node v at the k+1th layer, σ is the nonlinear activation function, N(v) is the set of neighbor nodes of node v, c _vu is the normalization factor, and W ^(k) is the learnable weight matrix of the kth layer;

Step 2: Use the Transformer model based on the self-attention mechanism to perform sequence modeling on the user's historical behavior data, generate a personalized learning path, and optimize the parameters through meta-learning and adaptive learning rate algorithm. The calculation formula is as follows:

Among them, Q is the query matrix, K is the key matrix, and V is the value matrix. is the dimension of the key vector;

Through meta-learning and adaptive learning rate algorithm, the learning path is dynamically adjusted to optimize the parameter θ:

Where Gt is the accumulated squared gradient, ∈ is the smoothing term;

Step 3: Use non-negative matrix factorization to preliminarily model the user feature matrix and resource utility matrix, and further optimize the resource allocation strategy through the diffusion model. The calculation formula is as follows:

where pθ(x _t-1 |x _t ) is the conditional probability of generating the previous state x _t-1 given the current state x _t , μθ(x _t ,t) and is the mean and variance predicted by the neural network, and I is the identity matrix;

Step 4: Combine real-time sentiment analysis technology to dynamically adjust learning content and feedback expression. The calculation formula is as follows;

Among them, Q(s,a) is the state-action value function, s is the current state, a is the current action, α is the learning rate, r is the immediate reward, γ is the discount factor, a′ is the possible future action, and s′ is the next state;

Step 5: Use differential privacy and federated learning methods to protect user data.

3. The AI-based art teaching management system according to claim 1, characterized in that the course generation module comprises:

The content generation unit uses convolutional neural networks and stochastic gradient descent optimization algorithms to automatically generate diverse personalized course content based on user learning needs and interests;

Data fusion unit, which efficiently integrates multimodal learning data, processes data and extracts features through deep learning technology, and generates course content that suits the user's learning style;

The dynamic adjustment unit uses deep reinforcement learning algorithms to intelligently adjust course difficulty, teaching strategies and content presentation methods based on the user's learning progress and real-time performance.

4. The AI-based art teaching management system according to claim 1, characterized in that the real-time interaction module comprises:

The interaction analysis unit extracts interactive text features through word embedding method based on bidirectional long short-term memory network and convolutional neural network; the emotion recognition unit fuses the extracted text features with the user's voice and visual data through a multimodal emotion fusion network; the feedback generation unit generates personalized learning feedback based on the Transformer model, and dynamically adjusts the feedback strategy through a reinforcement learning algorithm.

5. The AI-based art teaching management system according to claim 1, characterized in that the evaluation module comprises: a multi-dimensional evaluation unit for automatically evaluating and analyzing the art works submitted by users from multiple aspects such as composition, color, technique and creativity based on computer vision technology and machine learning algorithms;

The adaptive evaluation engine unit uses the user's historical performance data and real-time feedback to dynamically adjust the weight of each evaluation dimension to achieve personalized evaluation standards;

The safety optimization unit is used to optimize the evaluation criteria and strategies using the deep reinforcement learning algorithm DQN.

6. The AI-based art teaching management system according to claim 1, characterized in that the interactive control module comprises:

The recognition unit uses a deep neural network model to recognize the user's voice, gesture, and image input and converts the instructions into operation commands;

The interface optimization unit adjusts the layout, content presentation order and visual effects of the learning interface in real time according to the recognition results.

7. The AI-based art teaching management system according to claim 1, characterized in that the data security module comprises:

The encryption unit uses the Advanced Encryption Standard AES-256 algorithm combined with distributed blockchain technology to encrypt and distribute user data;

The privacy protection unit protects user data based on differential privacy technology to prevent the leakage of sensitive information during data analysis. The monitoring unit uses behavioral analysis algorithms and zero-knowledge proof methods to monitor and detect data access behavior in real time, identify abnormal access, and support multi-party collaboration based on federated learning.

8. An AI-based art teaching management method according to any one of claims 1 to 7, applied to a computing device, the method comprising the following steps:

Step S1: Collecting multimodal learning data of the user, including but not limited to text, image, voice and video data;

Among them, text data is segmented and parsed using natural language processing technology; image data is collected by an image sensor with a resolution of 256x256 pixels and saved in JPEG format; the voice data sampling rate is 16kHz and stored in WAV format; the video data frame rate is 30fps and is encoded in MP4 format;

Step S2: extracting and fusing the collected multimodal data through the multimodal deep integrated learning algorithm MMDILA, wherein a convolutional neural network is used to extract features from image data, a bidirectional long short-term memory network is used to perform semantic analysis on text data, and a sentiment recognition model is used to process speech data. The Transformer model with a self-attention mechanism is used to fuse information from different data modalities to generate a comprehensive learning profile of the user;

Step S3: Based on the user learning portrait, dynamically adjust the personalized course content and teaching strategy through the reinforcement learning algorithm, wherein the reinforcement learning algorithm is based on the policy gradient method, the learning rate is 0.01, the discount factor is 0.9, and the optimization goal is to maximize the user learning effect and satisfaction.

9. The AI-based art teaching management method according to claim 8, characterized in that the integration process of the multi-level deep learning model further comprises the following steps:

Step 1: Through the multi-task learning framework, the outputs of the CNN model and the Bi-LSTM model are jointly trained in the shared layer;

Step 2: Apply self-supervised learning technology to pre-train users’ multimodal interaction data;

Step 3: Use the self-attention mechanism to dynamically adjust the importance of different data modalities.

10. The AI-based art teaching management method according to claim 8, characterized in that the step of dynamically adjusting personalized course content and teaching strategies comprises:

Step 1: Use the improved multimodal deep ensemble learning algorithm MMDILA and adversarial generative network model to predict user behavior;

Step 2: Train the generator to simulate the user's potential behavior and train the discriminator to distinguish the generated behavior from the actual behavior;

Step 3: Based on the results of the above adversarial training, adjust the user's learning path and feedback mechanism in real time.