Leveraging Federated Learning for Privacy-Preserving Collaborative Learning Platforms in Education

Abstract

The goal of this research is to provide a collaborative educational learning framework that supports adaptive learning performance while ensuring the privacy of sensitive learner information in distributed educational environments by implementing Federated Learning. A centralized education system might experience issues with data privacy, potential security threats, and a lack of seamless collaboration among educational institutions in sharing knowledge and intelligence. To overcome these drawbacks, the suggested framework incorporates Federated Learning, Differential Privacy, Secure Aggregation, Adaptive Learning mechanisms, and Federated Averaging for decentralized collaborative model training without sharing raw educational data. It was implemented with the help of TensorFlow Federated, PyTorch, and Scikit-learn in a distributed educational simulation environment. Three metrics were selected for the evaluation of performance: Accuracy, Precision, Recall, F1-Score, Student Engagement Rate, Learning Completion Rate, Privacy Preservation Score, and Communication Efficiency. Experimental results showed that the proposed framework can be achieved with 94.8% Accuracy, 93.7% Precision, 92.9% Recall, and 93.3% F1-Score, surpassing the conventional centralized and distributed learning models. There was an increase in Student Engagement Rate to 91.3% and Learning Completion Rate to 89.6%. After conducting a privacy evaluation, the Data Leakage Risk decreased from 28.5% to 6.4%, and the Privacy Preservation Score increased to 95.2%. Communication Overhead was also improved from 920 MB to 410 MB, has high scalability, and is efficient for distributed training. Given the current digital learning environment, Federated Learning can bridge the gap between education outcomes, learning engagement, communication efficiency, and privacy protection, making it a secure, scalable, and efficient method for delivering intelligent, collaborative learning systems.