医学PBL的智能体矩阵Smart PBL的设计与应用
王涛 , 罗肖 , 张莉 , 孙波 , 郭媛 , 刘进军 , 王渊
中国医学教育技术 ›› 2026, Vol. 40 ›› Issue (3) : 308 -313.
医学PBL的智能体矩阵Smart PBL的设计与应用
Design and application of Smart PBL AI agent matrix for medical PBL
基于问题的学习(problem-based learning, PBL)是培养医学生批判性思维、自主学习与团队协作等高阶能力的重要教学模式,但在当前实践中,学生在专业表达、学习主题构建与资料获取方面仍需提供支持,指导教师(tutor)的引导效果与反馈质量仍有提升空间,教学组织与评估效率亟待升级。为解决上述问题,项目团队以“角色维度×流程维度×赋能层次”为设计框架,基于低代码智能体平台,构建了覆盖学生、指导教师、管理者三类角色,贯穿PBL全流程的Smart PBL智能体矩阵,包含PBL partner、PBL cotutor、PBL admin三个子系列共9 个智能体,实现全角色、全流程、多层次的智能支持。在西安交通大学临床医学专业PBL教学中对这些智能体开展探索性应用,结果显示:学生的学习主题构建能力、专业表达能力与资料获取效率显著提升,指导教师普遍反映备课效率、课堂引导水平和评语撰写质量有所提升,管理者亦认可智能体在合理分组、讨论质量评估和教学案例生成方面的支持作用。本项目为AI融入医学教学提供了系统化的实施路径与可借鉴的实践范式。
Problem-based learning (PBL) is an important pedagogical mode for cultivating medical students’ higher-order competencies such as critical thinking, self-directed learning, and teamwork. However, in current practice, students still need support in professional expression, learning topic development, and information retrieval, tutors have room for improvement in discussion facilitation and feedback quality, and the efficiency of teaching organization and evaluation requires upgrading. To address these issues, the project team developed the Smart PBL agent matrix on a low-code agent platform, using a design framework of “role dimension × process dimension × empowerment level”. The matrix covers three user roles—students, tutors, and administrators—and spans the entire PBL workflow, comprising three sub-series (PBL Partner, PBL Cotutor, and PBL Admin) with a total of nine agents, providing intelligent support across all roles, all processes, and multiple empowerment levels. An exploratory application was conducted in PBL teaching among students majoring in Clinical Medicine at Xi’an Jiaotong University. Results showed that students’ learning topic formulation ability, professional expression skills, and literature searching efficiency were significantly improved. Tutors generally reported improvements in lesson preparation efficiency, discussion facilitation, and quality of written feedback. Administrators also recognized the agents’ supportive role in rational grouping, discussion quality evaluation, and teaching case generation. This project provides a systematic implementation approach and a replicable practical paradigm for the integration of AI into medical education.
| [1] |
|
| [2] |
黎加厚. 生成式人工智能对课程教材教法的影响[J]. 课程·教材·教法, 2024, 44(2): 14-21. |
| [3] |
|
| [4] |
|
| [5] |
刘明, 杨闽, 吴忠明, |
| [6] |
刘思宇, 谢春晖, 陶鑫, |
| [7] |
张澍, 孟列素, 刘进军, |
| [8] |
王艳凤, 苏小敏, 赵国星. 数智赋能“基础医学概论”实验教学创新与实践[J]. 中国生物化学与分子生物学报, 2025,41(12): 1870-1882. |
| [9] |
唐懿文, 陈奇峰, 肖寒, |
| [10] |
|
| [11] |
王维民. 第四代医学教育的深度剖析与实施策略[J]. 中华医学教育杂志, 2024,44(10): 721-725. |
| [12] |
郑娅峰, 赵亚宁, 黄璟玥, |
| [13] |
刘明, 郭烁, 吴忠明, |
| [14] |
毛刚, 王良辉. 人机协同: 理解并建构未来教育世界的方式[J]. 教育发展研究, 2021,41(1): 16-24. |
| [15] |
|
| [16] |
吴永和, 姜元昊, 陈圆圆, |
/
| 〈 |
|
〉 |