Hui Nie (Primary Contact)
Chengdu Shude Experimental Middle School (West Campus), China
Differential teaching, Federated Learning, Educational equity
31-12-2025
This study constructs an ethics-first lightweight artificial intelligence teaching framework aiming at the practical problems existing in junior middle school English differential teaching, such as rigid stratification mechanism, lack of cultural literacy cultivation and ethical disputes of technology application. The framework realizes non-inductive data collection based on smartphone sensors, uses federated learning technology to build a dynamic grouping model, and incorporates cultural competence as the core dimension into the hierarchical system. A one-semester teaching experiment among 326 students in the eastern, central and western regions shows that the grouping accuracy rate of this framework reaches 89%, students' cultural understanding ability is improved by 44%, the grammatical error rate is reduced by 31%, and the parent data withdrawal request rate is reduced to 0.9%. Research confirms that this framework provides a scalable practical scheme for promoting educational equity through ethical compliance design and lightweight technology path.
PDF
Copyright (c) 2025 Al lnnovations and Applications
This work is licensed under a Creative Commons Attribution 4.0 International License.
