Learning and Intelligent Systems: CIRCLE: Center for Interdisciplinary Research on Constructive Learning Environments

[9720359] Kurt Van Lehn匹兹堡大学CIRCLE：建设性学习环境跨学科研究中心该中心有三个主要目标。首先，本中心试图了解一种极其有效的教学方法，即人际辅导。这将通过使用协议分析、眼动追踪和控制实验来研究人类导师，并将他们与现有的计算机导师进行对比。最重要的假设是，当学生为自己构建知识时，他们学得最好。该中心的第二个目标是建立和测试新一代的计算机辅导系统，鼓励学生构建目标知识，而不是告诉他们。学生的自我建构可能需要通过提示、苏格拉底式提问或其他方式进行微妙的引导，这一目标将通过在现有的智能辅导系统中添加先进的规划和自然语言组件，并在合作学校中测试新的导师来实现。该中心的第三个目标是帮助将这种新技术整合到现有的教育实践中。这一目标将通过在整个开发过程中与从业者和学校合作，通过升级已经被数千名学生使用的导师，以及通过与其他项目、中心和行业建立伙伴关系来实现。实现中心的三个目标将影响多个学科。理解为什么人类导师如此有效，应该会对学习心理学产生重大影响。计算机科学将通过开发能够反应性地规划自然语言提示和问题以促使学生学习的系统而得到发展。随着中心的建构主义技术和教学方法在实际应用中被证明是有效的，并得到广泛应用，教育将得到改善。

9720359 Kurt Van Lehn University of Pittsburgh CIRCLE: Center for Interdisciplinary Research on Constructive Learning Environments The Center has three main goals. First, the Center seeks to understand an extremely effective pedagogy, human tutoring. This will be achieved by using protocol analysis, eye tracking and controlled experiments to study human tutors and contrast them to existing computer tutors. The overarching hypothesis is that students learn best when they construct knowledge for themselves. The second goal of the Center is to build and test new generation of computer tutoring systems that encourage students to construct the target knowledge instead of telling it to them. The student's self-construction may need to be delicately guided via prompting, Socratic questioning or other means, This goal will be achieved by adding advanced planning and natural language components to existing intelligent tutoring systems and testing the new tutors in collaborating schools. The Center's third goal is to help integrate this new technology into existing educational practices. This goal will be achieved by working with practitioners and schools throughout the development process, by upgrading tutors that are already in use by thousands of students, and by forming partnerships with other projects, centers and industries. Achieving the three Center goals will impact multiple disciplines. Understanding why human tutors are so effective should have a significant impact on the psychology of learning. Computer science will be advanced by developing systems that can reactively plan natural language hints and questions that get a student to learn. Education will be improved as the Center's constructivist technology and instructional methods prove to be effective in actual use and become widely available.