智能导学系统人机互动的跨学科研究与设计

Sim-Agent and Student Interaction in ITS: From Interdisciplinary Research and Design

智能导学系统(ITS)是借助人工智能技术,让计算机扮演专家指导者向不同特征学习者传授知识、提供学习指导的适应性学习支持系统。它有着不同于传统导学系统的基本结构与运行机制,学生模型建构与适应性指导是其两大核心特征。学生模型建构过程是ITS导学过程中通过追踪学生学习过程逐步形成特定学生知识学习模型。按照建构方式不同,常见的学生模型有四种,分别是期望与错误概念调整适应的学生模型、贝叶斯网络的学生模型、学习追踪的学生模型以及基于限制的学生模型,它们采用不同的学习行为指标。有效的人机互动不仅要以学生模型为基础,教学模型、课程脚本数据库还需要储存丰富的言语指导行为策略。ITS人机互动教育心理学研究,以及教育技术学、计算机科学领域的探索性设计研究进展启示:我国要推动ITS发展应联合不同领域研究者在已有基础上进行跨学科研究与设计;多研究专家教师的言语指导行为,形成科学系统的互动结构、脚本与策略;重视情感对学生学习的影响,建构集认知与情感为一体的学生模型。

Intelligent tutoring system （ITS） is an adaptive tutoring system which guides the learning processand learning outcome of students with different characteristics by artificial intelligence technology as an experttutor. ITS has the specific structure and mechanism different from traditional tutoring system, and its core charactersare student modeling and adaptive tutoring. According to educational psychology, student modeling in ITS is theprocess of constructing specific student learning model by tracing his or her learning behavior. There are mainly fourdifferent student models in ITS which employ different student learning behavior： Expectation and MisconceptionTailoring, Model Tracing, Constraint based Modeling, and Bayesian Network Modeling. According to human tutoringstudy in educational psychology, the effective interaction between Sim-agent and student should not only bases onstudent module in ITS system, but also needs rich and effective tutoring strategies in teaching module andcurriculum script＇s database. ITS studies from educational psychology, educational technology, and computerscience give some implications to us. Researchers in our country should be combined in various areas to promote thedevelopment of ITS on the basis of existing ITS research and design. Expert teachers＇ tutoring should be studiedmore for constructing effective interaction structure, tutoring script and tutoring strategies in ITS. In addition, morefocus should be put on student＇s emotion because of its influence to learning, and student cognitive and affectivemodel should be constructed in the future of ITS＇ student model design.