Federated contrastive learning framework for cross-platform teaching quality assessment

The rapid growth of online education platforms has led to fragmented 'data silos' in teaching quality metrics, making it challenging for traditional evaluation methods to achieve cross-platform dynamic tracking and analysis while protecting data privacy.This paper proposes an innovative evaluation system based on federated contrastive learning.By introducing a federated learning framework to establish a distributed collaborative training mechanism, it extracts common features from cross-platform teaching data through contrastive learning without sharing raw data.Experimental validation on the public educational network dataset demonstrates that this system elevates teaching quality assessment accuracy to 94.2%, representing a 12.8% improvement over traditional methods, while enabling real-time tracking and dynamic feedback on teaching effectiveness.This research provides an innovative technical pathway and effective solution to the challenge of reconciling data privacy protection with enhanced evaluation efficacy.