智慧科技

CounterTutor：基于错因诊断与反事实干预的个性化 AI 教学系统

国 刚

青岛开放大学

DOI: https://doi.org/10.59429/zhkj.v3i3.14393

Keywords: AI 教育；智能教学系统；错因诊断；误概念建模；反事实教学；概念修复

---

Abstract

大语言模型已能生成即时解析和个性化反馈，但多数 AI 教学系统仍把学生错误视为待纠正的结果，较少显式 建模错误背后的稳定认知规则。本文提出 CounterTutor，一个面向概念修复的错因感知反事实教学框架。系统首先 根据题目、学生错误答案、解题过程与历史记录诊断潜在误概念；随后生成数值代入、规则边界、结构展开、表征 转换与迁移验证等最小反事实干预，使学生发现自身错误规则的失效条件；最后通过自适应闭环更新错因状态并验 证迁移效果。在错因诊断、反馈质量和学习效果三个层面评估的结果显示该框架在诊断准确性、反馈针对性、迁移 正确率和误概念复发控制上优于标准解析、相似题推荐和普通 LLM 反馈。本文将 AI 教学目标从“解释正确答案” 推进到“修复错误规则”，为可解释、可验证的生成式教育系统提供了一条可扩展路径。

---

References

[1] Corbett A. T., Anderson J. R. Knowledge tracing: Modeling the acquisition of procedural knowledge. User Modeling and User-Adapted Interaction, 1994.

[2] Piech C., Bassen J., Huang J., et al. Deep Knowledge Tracing. NeurIPS, 2015.

[3] Wang Z., Lamb A., Saveliev E., et al. Diagnostic Questions: The NeurIPS 2020 Education Challenge, 2020.

[4] Eedi. Mining Misconceptions in Mathematics. Kaggle Competition, 2024.

[5] VanLehn K. The relative effectiveness of human tutoring, intelligent tutoring systems, and other tutoring systems. Educational Psychologist, 2011.

[6] Hattie J., Timperley H. The power of feedback. Review of Educational Research, 2007.

[7] Chi M. T. H., Bassok M., Lewis M. W., et al. Self-explanations in learning to solve problems. Cognitive Science, 1989.

[8] Pearl J. Causality: Models, Reasoning, and Inference. Cambridge University Press, 2009.

[9] Wachter S., Mittelstadt B., Russell C. Counterfactual explanations without opening the black box. Harvard Journal of Law & Technology, 2017.

[10] Kasneci E., Sessler K., Küchemann S., et al. ChatGPT for good? Learning and Individual Differences, 2023.