人机协同视域下研究生《高级机器学习》课程多维教学体系重构研究

Research on the Reconstruction of Multi-Dimensional Teaching System for Postgraduate Course Advanced Machine Learning from the Perspective of Human-AI Collaboration

摘要: 随着生成式人工智能技术的快速发展，研究生教育面临人才培养模式转型的迫切需求。针对当前《高级机器学习》研究生课程中存在的“理论与实践脱节”“通用案例与行业特色脱节”“传统评价与AI时代需求错位”等问题，本文提出了“人工智能+”赋能课程教学改革的整体方案。该方案以OBE成果导向教育理念为指导，从教学内容、教学模式、评价体系三个维度进行系统性重构：在教学内容层面，构建“AI素养导论 + 行业特色案例”的双轮驱动内容体系；在教学模式层面，设计“课前AI预习–课中精讲研讨–课后协作实验”的三阶递进式人机协同教学模式；在评价体系层面，建立“过程性评价 + 结构化互评 + AI使用规范”的多元评价机制。通过在北京物资学院2025级研究生《高级机器学习》课程为期一学期的教学实践，依托作业文本、问卷、课程成绩等多源材料，可以初步观察到学生在AI工具规范使用意识、算法实践完成质量方面存在改善迹象，课程学生成绩整体优良，优秀率达到80%以上。受单组行动研究设计约束，改革效果尚待更多教学轮次进一步校验；该方案可为同类型地方行业特色院校研究生专业课改革提供一种可供参考的实践框架。

Abstract: With the rapid advancement of generative artificial intelligence technologies, postgraduate education faces an urgent demand for the transformation of talent training models. Aiming at the prominent problems existing in the postgraduate course Advanced Machine Learning, including the disconnection between theoretical knowledge and practical application, the separation between general teaching cases and industry-specific characteristics, and the mismatch between traditional evaluation systems and the demands of the AI era, this paper proposes an integrated teaching reform scheme empowered by “Artificial Intelligence Plus”. Guided by the Outcome-Based Education (OBE) philosophy, the scheme systematically reconstructs teaching from three dimensions: teaching content, teaching mode and evaluation system. In terms of teaching content, a two-wheel driven content system integrating an introductory module on AI literacy and industry-specific characteristic cases is constructed. In terms of teaching mode, a three-stage progressive human-AI collaborative teaching framework is designed, consisting of pre-class AI-assisted preview, in-class intensive lecturing and seminar discussion, and after-class collaborative experiments. In terms of evaluation system, a diversified evaluation mechanism combining formative assessment, structured peer review and standardized rules for AI utilization is established. Through one-semester teaching practice of the Advanced Machine Learning course for 2025-entry postgraduates at Beijing Wuzi University, multisource materials including assignment texts, questionnaires and course grades suggest preliminary signs of improvement in students’ awareness of standardized AI-tool usage and the quality of their algorithm-oriented practical work. The students achieved generally satisfactory course performance, with an excellence rate of over 80%. Constrained by the singlegroup action-research design, the effects of this reform need further verification across more teaching cycles. The proposed scheme can serve as a practical framework for reference in the reform of specialized postgraduate courses at similar local industry-featured universities.