Artificial intelligence model training system based on federal learning

Abstract

The invention relates to the technical field of artificial intelligence model training, in particular to an artificial intelligence model training system based on federal learning. The system comprises a data acquisition and preprocessing module, a demand perception and analysis module, an adaptive communication and model updating module and a global model aggregation and distribution module, the data acquisition and preprocessing module collects learning behavior data of students from a plurality of learning terminals and converts the learning behavior data into a standardized data set; the demand perception and analysis module identifies learning modes and trends of students by using the data, accurately captures learning demands through an intelligent demand synchronization protocol, not only pays attention to learning progress and test results, but also quantifies urgency and importance of different demands in combination with multi-dimensional information such as interactive feedback and the like, and improves learning efficiency. According to the invention, an intelligent demand synchronization protocol is constructed in the demand perception and analysis module, so that the problem of balance between model consistency and personalization in a dynamic environment is solved, and the most suitable learning experience is provided for each student.

Description

Technical Field

[0001] The present invention relates to the technical field of artificial intelligence model training, and more specifically, to an artificial intelligence model training system based on federated learning. Background Art

[0002] With the continuous progress of Internet technology and artificial intelligence, online education platforms have gradually become an important part of modern education. As a key technology therein, artificial intelligence model training can process and analyze data from different participants, thereby providing personalized learning experiences and improving teaching quality. However, the educational content and forms change very rapidly, requiring the model to quickly adapt to new demands. Existing technologies usually rely on fixed-cycle data synchronization and model update strategies, which are difficult to cope with rapidly changing demands, resulting in slow system response and inability to adjust in a timely manner to meet personalized learning needs.

[0003] Therefore, although the current systems have achieved data management and preliminary personalized services to a certain extent, there are still significant deficiencies in the balance between model consistency and personalization in a dynamic environment. Specifically, traditional systems usually rely on fixed periodic data synchronization and model update strategies, which are difficult to cope with rapidly changing learning demands. When the system focuses too much on maintaining the consistency of the global model, it may ignore the unique needs of different user groups, thus reducing the learning effect; on the contrary, if personalization is overemphasized, it may lead to model fragmentation, affecting the overall performance and generalization ability. This imbalance ultimately affects the user experience and learning effect, making it difficult for the system to provide a truly efficient, flexible, and personalized learning experience. Summary of the Invention

[0004] The purpose of the present invention is to provide an artificial intelligence model training system based on federated learning to solve the balance problem between model consistency and personalization in a dynamic environment. Specifically, the demand perception and analysis module first constructs an intelligent demand synchronization protocol, identifies students' learning patterns and trends based on a standardized data set, and calculates demand weights using this data to quantify the urgency and importance of different demands. Then, the adaptive data update mechanism automatically adjusts the data exchange frequency and content according to the demand weights, optimizes resource allocation, and generates an enhanced data set to ensure that the system meets personalized needs while maintaining the consistency of the global model. The prediction model further analyzes and predicts students' future learning needs based on the enhanced data set, thereby generating detailed student learning demand data. This series of steps ensures that the system can dynamically respond to changes in students' demands, provide personalized learning resources and support in a timely manner, and maintain the consistency of the global model, ultimately improving the overall education quality and student satisfaction.

[0005] To achieve the above object, an artificial intelligence model training system based on federated learning is provided, including a data collection and preprocessing module. The data collection and preprocessing module collects students' learning behavior data from multiple learning terminals, and the data collection and preprocessing module preliminarily processes the students' behavior data and generates a standardized data set. It also includes:

[0006] A demand perception and analysis module. The demand perception and analysis module receives the standardized data set from the data collection and preprocessing module and constructs an intelligent demand synchronization protocol. The intelligent demand synchronization protocol uses a dynamic weight adjustment formula based on the standardized data set to quantify the urgency of different demands and obtains demand weights. The intelligent demand synchronization protocol includes an adaptive data update mechanism. The adaptive data update mechanism automatically adjusts the data exchange frequency based on the demand weights and obtains an enhanced data set. The intelligent demand synchronization protocol also includes a prediction model. The prediction model analyzes and predicts students' future learning needs based on the enhanced data set to generate students' learning demand data. The demand perception and analysis module transmits the enhanced data set and the students' learning demand data to the adaptive communication and model update module for providing customized learning content according to the students' actual learning progress.

[0007] As a further improvement of this technical solution, the learning behavior data includes but is not limited to learning time, types of tasks completed, and grades.

[0008] As a further improvement of this technical solution, the intelligent demand synchronization protocol calculates the comprehensive weight of each demand to reflect the urgency of each demand at a specific time point.

[0009] As a further improvement of this technical solution, the adaptive data update mechanism identifies the learning needs of students in a specific subject area and automatically adjusts the update priority of teaching resources based on the learning needs.

[0010] As a further improvement of this technical solution, the prediction model predicts students' future learning needs based on the enhanced data set.

[0011] As a further improvement of this technical solution, the adaptive communication and model update module receives the enhanced data set and the students' learning demand data from the demand perception and analysis module and dynamically adjusts the data exchange strategy according to the network conditions.

[0012] As a further improvement of this technical solution, the adaptive communication and model update module regularly performs model update operations, re-estimates the model parameters using the latest enhanced data set, and ensures that the model accurately reflects the learning trends and demand changes of students.

[0013] As a further improvement of this technical solution, the adaptive communication and model update module includes an adaptive communication mechanism for evaluating network quality.

[0014] As a further improvement of this technical solution, the adaptive communication and model update module further includes a feedback loop mechanism for continuously optimizing the model, calculating the prediction error, and finally generating the optimized model parameters and student learning requirement data.

[0015] As a further improvement of this technical solution, the global model aggregation and distribution module receives the model parameters and student learning requirement data transmitted from the adaptive communication and model update module, comprehensively analyzes them, identifies broader learning trends and patterns, and is used to optimize the allocation strategy of educational resources to ensure that each student can obtain the most suitable educational resources and services for themselves.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0017] 1. In this artificial intelligence model training system based on federated learning, by integrating data standardization processing and intelligent requirement analysis, the accuracy and efficiency of personalized learning support are significantly improved, and at the same time, the balance problem between model consistency and personalization in a dynamic environment is solved. First, in terms of personalized learning support, the demand perception and analysis module uses the standardized data set to identify the learning patterns and trends of students, and calculates the demand weights by combining multi-dimensional information such as interactive feedback, quantifying the urgency and importance of different demands. The adaptive data update mechanism automatically adjusts the data exchange frequency and content according to these weights, optimizes resource allocation, and generates an enhanced data set. Based on this data set, the prediction model can further analyze and predict the future learning requirements of students, so as to provide detailed student learning requirement data. This process not only ensures that the system can quickly respond to the actual demand changes of students, but also makes the allocation of educational resources more reasonable and efficient, providing the most suitable learning experience and support for each student.

[0018] 2. In this artificial intelligence model training system based on federated learning, by constructing an intelligent demand synchronization protocol and an adaptive data update mechanism, it can not only monitor and respond to the behavior patterns and feedback information of students in real time, but also dynamically adjust the educational resource allocation strategy on the premise of ensuring the global model consistency. This enables the system to quickly adapt to new learning requirements while maintaining high-quality teaching standards, greatly improving the flexibility and robustness of the system. This balanced strategy not only improves the effect of personalized learning, but also enhances the adaptability and competitiveness of the entire educational platform, bringing higher-quality and more personalized educational services to the majority of student groups. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a schematic flow chart of the artificial intelligence model training system based on federated learning of the present invention;

[0020] Figure 2Schematic diagram of the demand perception and analysis module of the present invention;

[0021] Figure 3 Schematic diagram of the adaptive communication and model update module of the present invention.

[0022] The meanings of the various labels in the figure are as follows:

[0023] Among them: 100, data acquisition and preprocessing module; 200, demand perception and analysis module; 300, adaptive communication and model update module; 400, global model aggregation and distribution module. Specific embodiments

[0024] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0025] At the same time, some technical terms are explained here:

[0026] The ARIMA model parameters are a commonly used time series analysis model. In this system, they refer to the coefficients of the autoregressive term, the differencing term, and the moving average term. These parameters are estimated based on historical data and are used to predict future demand trends.

[0027] See Figure 1 As shown, the purpose of this embodiment is to provide an artificial intelligence model training system based on federated learning, aiming to integrate the data acquisition and preprocessing module 100, the demand perception and analysis module 200, the adaptive communication and model update module 300, and the global model aggregation and distribution module 400 to provide an efficient, flexible, and responsive personalized learning experience for the online education platform. Thus, while ensuring the consistency of the global model, it meets the personalized needs of different user groups and provides a more efficient, flexible, and responsive online education solution.

[0028] In the data acquisition and preprocessing module 100, raw data is first obtained from each participating party. This data includes students' learning progress, test scores, interaction feedback, etc. To ensure the quality and consistency of the data, these raw data must be preprocessed. The specific steps include data cleaning, missing value handling, outlier detection, and standardization, etc. For example, assume we have a dataset of students' test scores , where represents the score of the th student.

[0029] To standardize the data so that it has the same scale and distribution, we use the following formula:

[0030] ;

[0031] wherein, represents the th data point in the original dataset (e.g., the test score of a certain student);

[0032] represents the dataset the minimum value in;

[0033] represents the dataset the maximum value in;

[0034] represents the th data point after standardization processing.

[0035] The standardized data will have a unified range (usually between 0 and 1), thus ensuring that data from different sources can be fairly compared and effectively utilized in subsequent analysis. The dataset after standardization processing is transmitted to the demand perception and analysis module 200. For example, the performance data of students in a certain region in a certain subject is standardized and passed as a standardized dataset to the next module. This not only ensures the consistency and quality of the data, solves the problems of uneven distribution and magnitude differences of the original data, but also provides a reliable basis for subsequent demand perception and model training.

[0036] In this way, the data collection and preprocessing module 100 lays a solid data foundation for the entire system, enabling subsequent modules to more effectively perform demand perception, model training, and optimization. This process ensures the efficient operation of the system and ultimately improves the personalized learning experience and overall education quality.

[0037] However, merely having high-quality data is not sufficient to address the dynamically changing learning needs, especially in terms of the personalized learning needs existing among different regions or groups. Existing systems often fail to respond to these changes in a timely manner. Therefore, to overcome the deficiencies of traditional systems in quickly adapting to new requirements, that is, being unable to effectively balance the global model consistency and personalized needs, this system introduces a demand perception and analysis module 200. This module identifies the most urgent current needs by real-time monitoring of user behavior patterns (such as learning progress, test scores), feedback information, etc., and dynamically adjusts the data exchange strategy and frequency according to these needs. This process not only improves the flexibility and response speed of the system but also lays a foundation for the subsequent adaptive communication and model update module 300. In this way, the data collection and preprocessing module 100 provides a solid data foundation for the entire system, while the demand perception and analysis module 200 further ensures that the system can flexibly respond to the ever-changing learning needs, ultimately achieving a more efficient and personalized online education solution.

[0038] See Figure 2 As shown, the demand perception and analysis module 200, as the core improvement point of the present invention, realizes the accurate capture and rapid response to students' learning needs by constructing an intelligent demand synchronization protocol. This module first aims to solve the deficiencies of existing systems in adapting to personalized learning needs, especially the problem of being unable to cope when dealing with the diverse learning needs existing among different regions or groups.

[0039] The demand perception and analysis module 200 receives the standardized data set transmitted by the data collection and preprocessing module 100, which is the basis for subsequent analysis. This module aims to solve the deficiencies of existing systems in adapting to personalized learning needs, especially the problem of being unable to cope when dealing with the diverse learning needs existing among different regions or groups. First, it uses the information in the received standardized data set, including students' learning progress, test scores, and interaction feedback, etc., to identify the learning patterns and trends of individuals and groups. These data are standardized to ensure the consistency and quality of the data, laying a solid foundation for accurately evaluating students' needs. Based on these standardized data sets, the intelligent demand synchronization protocol uses a dynamic weight adjustment formula to quantify the urgency and importance of different needs:

[0040] ;

[0041] In the formula, represents the comprehensive weight of the th demand, and this value is used to quantify the overall importance and urgency of each demand;

[0042] represents the urgency weight coefficient, which is a regulation parameter used to balance the contribution degree of urgency in the comprehensive weight;

[0043] represents the urgency score of the th requirement, and this value indicates the degree of urgency for this requirement to be processed;

[0044] represents the importance weight coefficient, which is a tuning parameter used to balance the contribution of importance in the comprehensive weight;

[0045] represents the th requirement's importance score, and this value indicates the degree of impact of this requirement on the overall goal or the learning effect of students.

[0046] This quantification method enables the system to automatically adjust the resource allocation strategy according to the changes in actual requirements, ensuring that students in the greatest need of help can receive timely support. This process directly depends on an accurate standardized dataset to ensure the accuracy of the requirement weight calculation.

[0047] To further respond to these changing requirements, the intelligent requirement synchronization protocol implements a set of adaptive data update mechanisms. Based on the changes in requirement weights, the system automatically adjusts the data exchange frequency and content, optimizes resource allocation, and improves efficiency. For example, when it detects a common difficulty among a large number of students in a certain subject area, the system will prioritize updating the teaching resources related to this and push personalized learning suggestions or supplementary materials to relevant students. This process not only depends on the previously determined requirement weights but also further enhances the ability to respond to personalized requirements. All operations are carried out based on the information provided by the standardized dataset and form an enhanced dataset by integrating the latest feedback and interaction data, thus more accurately reflecting the learning status and requirements of students.

[0048] In addition, to more comprehensively meet the learning needs of students, the intelligent requirement synchronization protocol introduces a prediction model that uses the enhanced dataset containing historical data and the latest interaction information to analyze and predict future possible learning requirements. This helps to prepare the corresponding educational resources in advance and relieve the service pressure during peak periods. By combining the rich information in the enhanced dataset, the prediction model can not only respond immediately to current requirements but also foresee future trends, providing more forward-looking support for educational services. This method ensures that the system can continuously optimize its response strategy to better serve the vast student population and improve the overall educational quality and personalized learning experience. The application of the enhanced dataset is an indispensable part of this process, which provides the necessary depth and accuracy to make prediction and resource allocation more effective. The prediction model is based on time series analysis techniques and uses the ARIMA model for modeling:

[0049] ;

[0050] In the formula, represents the demand at time ;

[0051] represents the constant term;

[0052] represents the autoregressive coefficient;

[0053] represents the demand at the past time points;

[0054] represents the moving average coefficient;

[0055] represents the error term at the past time points;

[0056] represents the error term at time ;

[0057] By combining demand weights and historical data, the prediction model can not only respond immediately to current demand but also foresee future trends, providing more forward-looking support for educational services. All of this is based on the effective analysis of the received standardized data set.

[0058] In summary, through its precise data analysis capabilities and flexible resource allocation strategies, the intelligent demand synchronization protocol obtains students' learning demand data after quantification and weight adjustment, ensuring seamless connection from data collection, demand quantification to resource optimal allocation, and significantly enhancing the personalized learning experience and educational quality. By effectively utilizing the standardized data set, the protocol can not only identify the learning patterns and trends of individuals and groups but also dynamically adjust resource allocation according to this information to meet the personalized learning needs of students. In addition, through the application of the prediction model, the system can foresee future possible learning demands in advance and make preparations for them, further enhancing the support ability for diverse learning environments.

[0059] However, although the demand perception and analysis module 200 has achieved accurate capture and rapid response to students' learning needs, there are still some challenges and deficiencies in practical applications. For example, in the face of unstable network conditions or data transmission delays, how to ensure real-time performance and accuracy becomes an urgent problem to be solved. In addition, with the continuous growth of students' data volume and the continuous change of learning needs, the existing models may not always maintain the optimal state and need to be continuously updated and optimized to adapt to new situations. To solve these problems, the system introduces the adaptive communication and model update module 300. This module focuses on improving the communication mechanism of the system to ensure efficient and stable data exchange under various network conditions. At the same time, it will also be responsible for regularly evaluating and updating the existing prediction models to ensure that they can accurately reflect the latest learning trends and demand changes. This can not only enhance the overall robustness and flexibility of the system, but also further improve the quality and efficiency of personalized learning services, providing students with a better learning experience. Therefore, the adaptive communication and model update module 300 will further improve the overall system architecture on the existing basis and promote the optimal utilization of educational resources.

[0060] See Figure 3 As shown, the adaptive communication and model update module 300 receives the quantified and weight-adjusted students' learning need data from the demand perception and analysis module 200, as well as the enhanced data set generated by this module. These data not only contain basic information such as students' learning progress, test scores, and interaction feedback, but also incorporate the urgency and importance scores of needs calculated based on the dynamic weight adjustment formula. Through such information transfer, the adaptive communication and model update module 300 can accurately identify the learning needs that need to be prioritized and timely update the system model according to the latest learning trends.

[0061] To ensure efficient and stable data exchange under different network conditions, this module designs an adaptive communication mechanism. First, according to the network condition indicators in the received students' learning need data, a comprehensive network quality index is calculated:

[0062] ;

[0063] In the formula, represents the comprehensive network quality index, which is used to evaluate the overall quality of the current network condition;

[0064] represents the comprehensive network quality index, which is used to evaluate the overall quality of the current network condition;

[0065] represents the delay time, indicating the delay of data transmission;

[0066] represents an adjustment parameter used to balance the impact of packet loss rate;

[0067] represents the packet loss rate, that is, the proportion of data packets lost during data transmission.

[0068] Based on different ranges of values, the system will automatically adjust the data transmission strategy. For example, in the case of high latency or high packet loss rate, a more robust coding method or a lower data sending frequency will be adopted to ensure the quality and efficiency of data transmission.

[0069] In addition, to achieve continuous optimization of the model, this module introduces a feedback loop mechanism that is closely connected to the data collection and preprocessing module 100 and the demand perception and analysis module 200. Specifically, when there is a deviation between the model prediction result and the actual learning outcome, the system will trigger a feedback process to re-evaluate and adjust the previously used weight coefficients and . By comparing the original predicted value and the actual observed value, the following formula can be used to calculate the error

[0070] ;

[0071] where represents the prediction error;

[0072] represents the actual observed value;

[0073] represents the original predicted value.

[0074] Then, based on the results of the error analysis, the system will automatically adjust the parameter settings, thereby improving the accuracy and response speed of the model. This two-way interaction not only enhances the synergistic effect among the modules but also provides a more accurate basis for subsequent data analysis.

[0075] Furthermore, the adaptive communication and model update module 300 will also regularly perform model update operations, using the latest enhanced dataset as training samples. This includes, but is not limited to, re-estimating the ARIMA model parameters to better reflect the changing trends of the current learning environment. In this way, the adaptive communication and model update module 300 ensures that the entire system can flexibly respond to the ever-changing learning needs and technical challenges, providing students with a continuously optimized learning experience. After obtaining the student learning demand data, the adaptive communication and model update module 300 transmits this data and the updated model parameters to the global model aggregation and distribution module 400 to further optimize the allocation of educational resources and personalized learning services.

[0076] The global model aggregation and distribution module 400 first receives the optimized model parameters and student learning requirement data from the adaptive communication and model update module 300, conducts in-depth analysis by synthesizing this information, and identifies broader trends and patterns. This module is not only responsible for integrating the data collected from various local models but also ensures that personalized requirements in different educational scenarios are met. Specifically, it aggregates data such as students' learning behaviors, preferences, and achievements scattered across different locations, and uses this rich information to adjust the educational resource allocation strategy, improving the relevance and effectiveness of teaching content.

[0077] In addition, the global model aggregation and distribution module 400 promotes cross-user knowledge sharing, enabling the system to dynamically respond to the needs of different learners, thereby providing more personalized learning paths and support. In this way, even while protecting user privacy, the system can continuously evolve and improve, bringing benefits to all participants. Ultimately, the artificial intelligence model training system based on federated learning successfully achieves the efficient utilization of educational resources, improves the quality of education, and promotes the development of personalized learning, enabling each student to obtain the most suitable learning experience and support.

[0078] In summary, by constructing an intelligent requirement synchronization protocol in the requirement perception and analysis module 200 to accurately capture and analyze students' learning requirements, and providing customized learning content and support according to the actual learning situation and progress of students, this system effectively solves the problems of uneven distribution of educational resources and insufficient personalized learning support. Ultimately, the artificial intelligence model training system based on federated learning not only enhances the satisfaction and learning effects of individual learners but also promotes the optimization of the overall educational environment, ensuring that a wider student population can benefit from high-quality, personalized educational experiences.

[0079] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art of this industry should understand that the present invention is not limited by the above embodiments. The above embodiments and the descriptions in the specification are only preferred examples of the present invention and are not used to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of protection claimed by the present invention is defined by the appended claims and their equivalents.

Claims

1. An artificial intelligence model training system based on federated learning, comprising a data collection and preprocessing module (100), wherein the data collection and preprocessing module (100) collects students' learning behavior data from multiple learning terminals, and the data collection and preprocessing module (100) performs preliminary processing on the student behavior data and generates a standardized data set, characterized in that: Also includes: A demand perception and analysis module (200), wherein the demand perception and analysis module (200) receives the standardized data set from the data acquisition and preprocessing module (100) and constructs an intelligent demand synchronization protocol, wherein the intelligent demand synchronization protocol uses a dynamic weight adjustment formula based on the standardized data set to quantify the urgency of different demands and derive the demand weight, wherein the intelligent demand synchronization protocol includes an adaptive data update mechanism, wherein the adaptive data update mechanism automatically adjusts the data exchange frequency based on the demand weight and obtains an enhanced data set, wherein the intelligent demand synchronization protocol also includes a prediction model, wherein the prediction model analyzes and predicts the students' future learning needs based on the enhanced data set to generate student learning needs data, wherein the demand perception and analysis module (200) transmits the enhanced data set and the student learning needs data to the adaptive communication and model update module (300) for providing customized learning content according to the students' actual learning progress.

2. The artificial intelligence model training system based on federated learning according to claim 1, characterized in that: The learning behavior data include but are not limited to learning time, completed task types and scores, and are standardized by the following formula: ； In the formula, Represents the first data points; Representation dataset The minimum value in ; Representation dataset The maximum value in ; After standardization, data points.

3. The artificial intelligence model training system based on federated learning according to claim 1, characterized in that: The intelligent demand synchronization protocol calculates the comprehensive weight of each demand through the following formula to reflect the urgency of each demand at a specific time point: ； In the formula, Indicates The comprehensive weight of each demand; represents the urgency weight coefficient; Indicates The urgency rating of each requirement; represents the importance weight coefficient; Indicates The importance rating of each requirement.

4. The artificial intelligence model training system based on federated learning according to claim 1, characterized in that: The adaptive data update mechanism identifies students' learning needs in specific subject areas and automatically adjusts the updating priority of teaching resources based on the learning needs.

5. The artificial intelligence model training system based on federated learning according to claim 1, characterized in that: The prediction model predicts students' future learning needs based on the enhanced data set and calculates the demand using the following formula: ； In the formula, Indicates at time Quantity of demand; represents a constant term; represents the autoregressive coefficient; Indicates the past The demand at a point in time; represents the moving average coefficient; Indicates the past The error term at each time point; Indicates at time The error term of .

6. The artificial intelligence model training system based on federated learning according to claim 1, characterized in that: The adaptive communication and model updating module (300) receives the enhanced data set and student learning demand data from the demand perception and analysis module (200), and dynamically adjusts the data exchange strategy according to the network status.

7. The artificial intelligence model training system based on federated learning according to claim 1, characterized in that: The adaptive communication and model updating module (300) regularly performs model updating operations, and re-estimates model parameters using the latest enhanced data set to ensure that the model accurately reflects the learning trends and demand changes of students.

8. The artificial intelligence model training system based on federated learning according to claim 7, characterized in that: The adaptive communication and model updating module (300) includes an adaptive communication mechanism for evaluating network quality, and calculates a comprehensive network quality index using the following formula: ； In the formula, Represents a comprehensive network quality index, which is used to assess the overall quality of current network conditions; Represents a comprehensive network quality index, which is used to assess the overall quality of current network conditions; Indicates delay time, indicating the delay in data transmission; Indicates the adjustment parameter used to balance the impact of packet loss rate; Indicates the packet loss rate, that is, the proportion of data packets lost during data transmission.

9. The artificial intelligence model training system based on federated learning according to claim 8, characterized in that: The adaptive communication and model updating module (300) also includes a feedback loop mechanism for continuously optimizing the model and calculating the prediction error through the following formula, and finally generating optimized model parameters and student learning demand data: ； In the formula, represents the prediction error; represents the actual observed value; Represents the original predicted value.

10. The artificial intelligence model training system based on federated learning according to claim 9, characterized in that: The global model aggregation and distribution module (400) receives the model parameters and student learning demand data transmitted from the adaptive communication and model updating module (300) and performs a comprehensive analysis on them to identify broader learning trends and patterns for optimizing the allocation strategy of educational resources, thereby ensuring that each student can obtain the educational resources and services that best suit him or her.

Images