Adaptive Instructional Systems for Intelligent Computer-Assisted Instruction

人工智能(AI)衍生的技术和理论将应用于基于计算机的教育领域。智能计算机辅助教学(ICAI)是该领域的研究方向之一，其目标是实现在学生解决问题的过程中接受学生的解释，推理学生的知识状态，并自适应地引导他们的系统。要实现这一目标，必须解决学生与系统之间的知识交流建构心理模型的问题。为了建立这种智能教学系统的组织方法，我们开发了一个由以下模块构成的系统：(1)自然语言界面：通过语法分析和格语法分析，将学生对问题的解释和求助转化为框架形式的内部表示。(2)专家知识和推理机制：利用系统中的原始知识和过程性知识，将学生的行为重构为学生模型。在重构过程中，诊断学生的误解和错误，并通过直接辅导策略回答学生的提问。(3)基于脚本的监控：通过脚本来表达问题解决的过程。在一个工作站上组织了一个辅导系统，并通过比较三组60名学生对该系统进行了评估。结果表明，该系统对于初学者来说是有效的，他们可以用句子来解释自己的行为和未知的知识。

Techniques and theories derived from Artificial Intelligence (AI) are going to be applied to computer-based educational fields. Intelligent Computer-Assisted Instruction (ICAI) is one of the researches in the fields and the goal of ICAI is to realize the systems that accept students' explanation in their problem-solving process, reason their knowledge state and guide them adaptively. The issue of constructing mental model from knowledge communication between a student and a system must be solved for pursuing the goal. To establish the organization method of such intelligent tutoring systems, we developed the system that is constructed by following modules.(1) Natural Language Interface : Student's explanation for solving problems and their asking for helps are translated to the frame-formed internal representation by the syntactic analysis and the case grammar analysis.(2) Expert knowledge and Inference Mechanism : The student's action is reconstructed to the student model by using the primitive knowledge and the procedural knowledge in the system. During reconstruction, student's misunderstanding and errors are diagnosed and his asking is answered by the direct tutorial strategy.(3) Script-Based Monitor : Problem-solving process is expressed by scripts. Diagnosis and assistance can be attained using proposed monitor.A tutoring system was organized on a workstation and the system evaluation was done by comparing three groups of 60 students. The result shows the efficiency of the system for novice students who can explain their action and unknown knowledge by sentences.