Evaluation of Error-Based Simulation by Using Qualitative Reasoning Techniques

 

Tsukasa Hirashima† and Tomoya Horiguchi††

†Department of Artificial Intelligence, Kyushu Institute of Technology 680-4, Kawatsu, Iizuka, Fukuoka 820-0067 JAPAN Phone: +81-948-29-7618, E-mail: tsukasa@ai.kyutech.ac.jp

††Faculty of Information Engineering, Kobe University of Mercantile Marine 5-1-1, Fukaeminami, Higashinada, Kobe 658-0022 JAPAN Phone: +81-78-431-6263, E-mail: horiguti@ti.kshosen.ac.jp

 

Abstract: This paper introduces an application case of qualitative reasoning techniques to realize Intelligent Learning Environment with Error-Based Simulation (EBS). EBS is a method to generate feedback for learning from mistakes, targeting an erroneous equation made in solving a mechanics problem. It is a behavior simulation that is generated by changing the value of "velocity" or "acceleration" of an object by using an erroneous equation. In the EBS, the object behaves in a strange manner. The difference between an EBS and normal simulation visualizes the error. Although EBS is a promising method to make a learner aware of a mistake in an equation, it isn't always effective. Therefore, evaluation and management of EBS is necessary. In this paper, the framework to evaluate EBS by using QSIM and DQ-analysis are reported.

 

Full Paper (PDF 8602 KB)