生源多样化背景下《人工智能技术》分层教学实践

Tiered Teaching Practice in the “Artificial Intelligence Technology” Course in the Context of Student Diversity

摘要: 在高等职业教育招生制度改革深化背景下，生源多样化已成为人工智能相关专业课程教学的突出挑战。本文针对机器人工程与智能制造专业混合编班中存在的双重差异，以《人工智能技术》课程为载体，提出“双维分流导向 + 梯队式项目驱动 + 增值性多元评价”的教学改革模式。通过模块化重构教学内容，采用“黑白盒”分级教学策略，并组建“1 + 1 + 1”异质化团队进行项目协作。本研究基于准实验研究设计，设立传统教学平行班作为对照组与课程改革班作为实验组，构建了贯穿课程前、中、后的“知识掌握、实践能力、创新思维”三维测评体系。通过方差分析与多元回归分析等统计方法对实证数据进行处理，统计结果表明：不同招生路径生源在基础理论与实践偏好上存在显著的双向结构性差异(P < 0.01)；在控制学生学业初始水平的条件下，实验组学生在工程实践能力与创新思维维度的学期增值幅度显著优于传统对照组(P < 0.05)。多元回归分析进一步证实，该分层互补协作模式对多元生源的综合能力增值具有显著正向解释力，为新工科背景下人工智能专业课程的差异化融合教学提供了严谨的量化实证范式。

Abstract: Against the backdrop of deepening reforms in the enrollment system of higher vocational education, the diversification of student backgrounds has become a prominent challenge in teaching courses related to artificial intelligence. This paper addresses the dual differences (horizontal differences in professional background and vertical differences in student level) existing in the mixed-class teaching of Robotics Engineering and Intelligent Manufacturing majors. Taking the “Artificial Intelligence Technology” course as a vehicle, it proposes a teaching reform model of “dual-dimensional diversion guidance + tiered project-driven + value-added multi-dimensional evaluation”. Through modular content reconstruction, the “black and white box” teaching strategy, and the formation of “1 + 1 + 1” heterogeneous teams for project collaboration, this course addresses these dual differences. Based on a quasi-experimental research design, this study established a parallel class using traditional teaching as the control group and the reformed class as the experimental group, constructing a three-dimensional assessment system covering “knowledge mastery, practical ability, and innovative thinking” across the pre-, mid-, and post-course stages. Empirical data were analyzed using statistical methods including analysis of variance (ANOVA) and multiple linear regression. The statistical results reveal significant bidirectional structural differences in foundational theory and practical preferences among students from different enrollment pathways (P < 0.01). Controlling for initial academic levels, students in the experimental group demonstrated significantly higher semester value-added growth in engineering practice and innovative thinking compared to the control group (P < 0.05). Multiple regression analysis further confirms that this tiered complementary collaboration model exerts a significant positive explanatory power on the comprehensive capacity enhancement of diverse students, providing a rigorous empirical paradigm for differentiated and integrated teaching of AI courses across new engineering disciplines.