Case-Based Advice System for Novice Programs Written in Assembly Language

针对用汇编语言编写的新手程序的基于案例的建议系统

• 批准号: 14580428
• 负责人: WATANABE Hiroyoshi
• 金额: $ 1.79万
• 依托单位: Teikyo University
• 依托单位国家: 日本
• 项目类别: Grant-in-Aid for Scientific Research (C)
• 财政年份: 2002
• 资助国家: 日本
• 起止时间: 2002 至 2003
• 项目状态: 已结题
• 来源: https://kaken.nii.ac.jp/en/grant/KAKENHI-PROJECT-14580428/
• 关键词: Programming Education; Advice; Assembly Language; CASL; Case-based Reasoning

We worked out a method of giving advice for novice programs using a case-based reasoning approach in programming education. Our target programs for the advice are called near-miss programs. A near-miss program is a program that does not run properly but is rather close to correct programs.We proposed a model of giving step-by-step advice. That is, the advice system gives small hints at first and then gives more hints or detailed advice later as needed. Students will gain more confidence when they solve problems with small hints than in cases where they got detailed advice.We also proposed two case-based methods of generating advice sentences. The first method is to generate advice sentences based on the similarity to advice cases, which include a near-miss program as a problem description and advice sentences on the program as a solution description. The second method is advice generation based on the difference between the target near-miss program and a correct program. There are many correct programs in evaluation case-bases of the evaluation assistant system we already implemented.Based on these ideas, we implemented a case-based advice system for a simple assembly language CASL II. The system had been utilized in actual classes and the results suggested that the case-based advice system was useful. In particular, we found that the difference-based method was very useful. Advice sentences were presented by the difference-based method even if case-bases were empty at first, because the difference-based method uses evaluation cases which are added to the case-base automatically when teachers evaluate students' submitted programs.

在程序设计教学中，我们采用基于案例的推理方法，提出了一种给新手程序提供建议的方法。我们的建议的目标程序被称为近距离程序。差一点的程序是一个不能正常运行的程序，但它与正确的程序相当接近。也就是说，建议系统首先给出小提示，然后根据需要给出更多提示或详细建议。学生在解决问题时，通过小的提示会比通过详细的建议获得更多的自信。我们还提出了两种基于案例的建议句生成方法。第一种方法是基于与建议案例的相似性生成建议语句，其中包括作为问题描述的未遂程序和作为解决方案描述的关于该程序的建议语句。第二种方法是建议生成的基础上的差异的目标接近错过的程序和正确的程序。在我们已经实现的评估辅助系统中，评估案例库中有许多正确的程序，基于这些思想，我们实现了一个简单的汇编语言CASL II的基于案例的建议系统。该系统已在实际课堂中使用，结果表明，基于案例的建议系统是有用的。特别是，我们发现基于差异的方法非常有用。基于差异的方法使用评价案例，当教师评价学生提交的程序时，评价案例会自动添加到案例库中，即使案例库一开始为空，也可以通过基于差异的方法给出建议句。

项目成果

渡辺博芳, 高井久美子, 荒井正之, 武井惠雄: "初等アセンブラプログラミングにおけるニアミスプログラムに対するアドバイス文生成法"情報技術レターズ. 1. 241-242 (2002)

Hiroyoshi Watanabe、Kumiko Takai、Masayuki Arai、Keio Takei：“基本汇编程序编程中的近似错误程序的建议语句生成方法”《信息技术快报》1. 241-242 (2002)。

Watanabe, H., Takai, K., Arai, M, Takei, S.: "Giving Advice Based on Differences between Near-Miss and Correct Programs"Journal of Information and Systems in Education. Vol.2, No.1, (to appear). (2003)

Watanabe, H.、Takai, K.、Arai, M、Takei, S.：“根据未遂计划和正确计划之间的差异提供建议”教育信息与系统杂志。

H.Watanabe, K.Takai, M.Arai, S.Takei: "Case-Based Adviser for Near-Miss Programs"Proceedings of The 11th International Conference on Artificial Intelligence in Education(AI-ED2003). 149-156 (2003)

H.Watanabe、K.Takai、M.Arai、S.Takei：第 11 届教育人工智能国际会议（AI-ED2003）“基于案例的未遂项目顾问”会议记录。

H.Watanabe, K.Takai, M.Arai, S.Takei: "Giving Advice Based on Differences between Near-Miss and Correct Programs"Journal of Information and Systems in Education. Vol.2NO.1. (2003)

H.Watanabe、K.Takai、M.Arai、S.Takei：“根据未遂计划和正确计划之间的差异提供建议”教育信息与系统杂志。

Watanabe, H., Takai, K., Arai, M., Takei, S.: "A Method of Generating Advice Sentences for Novice Programs Written in Assembly Language"Information Technology Letters. Vol.1. 241-242 (2002)

Watanabe, H.、Takai, K.、Arai, M.、Takei, S.：“为用汇编语言编写的新手程序生成建议句子的方法”信息技术快报。