Reconstruction and Practice of AI-Empowered Training Models for Electronic Information Engineering Professional Master's Students

Abstract

To address the insufficient interdisciplinary integration, research computing bottlenecks, and low industrial adaptability in the training of Electronic Information Engineering professional master's students, this project leveraged the first-level discipline doctoral program in Physics to systematically reform the AI-empowered training model. Following a trinity pathway of "structural optimization, intelligent empowerment, and mechanism innovation," the study constructed a cross-curricular knowledge-capability graph and developed a central agent teaching platform. With the "Optoelectronic Engineering" and "Communication Engineering" tracks as pilot cases, we collected over 1,000 hours of controlled experimental data. The results demonstrate that the experimental group (AI-empowered) significantly outperformed the control group in terms of course objective attainment, research training efficiency, and practical innovation capabilities. Furthermore, the core value of AI empowerment lies in improving the "process quality" and "high-order outputs" of learning, rather than merely boosting test scores. This project validates the effectiveness of the "Physics Foundation—AI Empowerment—Engineering Application" integration framework, providing a replicable model for the reform of engineering professional degree graduate education in regional universities.