Articulate Virtual Laboratories for Science and Engineering Education

9453078 Forbus该项目的目标是开发清晰的虚拟实验室（avl），通过脚手架和指导学生进行概念设计任务来教授科学和工程原理。我们正在测试的教育猜想是，清晰的虚拟实验室将使学生能够(a)从根本上更好地学习基本原理，否则他们将无法执行。为了验证这些猜想，我们的原型实验室将被西北大学和牛津大学的工程本科生以及埃文斯顿镇高中的高中生用作课程作业的一部分。我们相信，清晰的虚拟实验室可以极大地改善工程和科学教育。设计经验对工程教育至关重要，并为学习基本物理原理提供了强大的激励环境：如果不使用广泛的物理原理，就无法设计喷气发动机、冰箱或发电厂。设计环境可以支撑学生，让他们专注于基础知识，这对于基础科学和工程的教学来说是非常宝贵的，并且可以更好地激发学生对科学学习的兴趣。在典型的课堂环境中，很难提供设计经验，因为许多有趣的物理工件（如发电厂、喷气发动机和冰箱）的建造和实验都是昂贵的或危险的。清晰的虚拟实验室将通过使学生能够在模拟环境中设计、分析和测试工件来解决这些问题，成本低且安全。他们将为学生提供指导，以帮助他们理解基本原理，帮助他们练习建模、分析和设计物理系统所需的技能，并提供一种好的实验室助理所提供的监督，通过减少学生探索中不具启发意义的方面。创建模拟的虚拟实验室需要综合几种人工智能技术的进步。定性物理为科学家和工程师的隐性知识提供了形式化的表示，将他们的专业知识与基于经验的直觉联系起来，使软件能够使用类似于领域专家认为自然的方法和概念。组合建模为计算机辅助建模提供了表示和推理技术（例如，如何将一个领域的专业知识应用于建模真实世界的情况，以便对它们进行正式分析）。真理维护系统提供推理服务和原材料，用于构建系统结果和推理的解释，以帮助学生理解学习领域。符号代数和约束传播提供了数学解决方案。类比处理技术提供了检索和应用从设计和实例库的结果到实际情况的能力，以指导学生。我们正在开发两个原型清晰的虚拟实验室，与西北大学和牛津大学的工程学院合作，他们愿意与他们的学生一起使用它们。首先是热力学循环，这是电厂、推进系统、制冷系统和热泵概念设计中使用的一种理想化方法。要掌握热力学循环的设计和分析，需要对热力学的大量内容有深刻的理解。第二个是反馈控制器。反馈和控制理论的概念渗透到现代科学和工程中。虽然设计最优控制器需要数学分析，但我们相信，在适当的指导下，高中生可以掌握反馈的许多重要概念。* * *

9453078 Forbus The goal of this project is to develop articulate virtual laboratories (AVLs) that teach science and engineering principles by scaffolding and coaching students in conceptual design tasks. The educational conjecture we are testing is that articulate virtual laboratories will enable students to (a) learn fundamental principles radically be better than they would otherwise be unable to perform. To test these conjectures, our prototype laboratories will be used by engineering undergraduates from Northwestern University and Oxford University as part of their course work, and by high school students from Evanston Township High School. We believe that articulate virtual laboratories could dramatically improve engineering and science education. Design experience is essential to engineering education, and provides a powerful motivating context for learning fundamental physical principles: One cannot design a jet engine, refrigerator, or power plant without using a broad range of physical principles. Design environments that scaffold students, allowing them to focus on fundamentals, could prove invaluable for instruction in basic science as well as engineering, and could better motivate interest in science learning. Design experiences are difficult to provide in typical classroom settings, because many interesting physical artifacts (such as power plants, jet engines, and refrigerators) are expensive or dangerous to build and experiment with. Articulate virtual laboratories will address these problems by enabling students to design, analyze, and test artifacts in a simulated environment, cheaply and safely. They will provide coaching for students, in order to help them understand fundamental principles, to help them practice the skills needed to model, analyze, and design physical systems, and to provide the kind of supervision that a good laboratory assistant provides by way of minimizing unenlightening aspects of student explorations. Creating art iculate virtual laboratories requires synthesizing advances involving several AI technologies. Qualitative physics provides formal representations for the tacit knowledge of scientists and engineers that connects their professional knowledge to their experience-based intuitions, enabling software to use methods and concepts similar to those deemed natural by domain experts. Compositional modeling provides representations and reasoning techniques for computer-assisted modeling (e.g., how to apply professional knowledge of a domain to modeling real-world situations so that they can be formally analyzed). Truth- maintenance systems provide reasoning services and the raw material for constructing explanations of the system's results and reasoning in terms that help students understand the domain of study. Symbolic algebra and constraint propagation provide mathematical solutions. Analogical processing techniques provide the ability to retrieve and apply results from libraries of worked out designs and examples to notel situations, in order to coach students. We are developing two prototype articulate virtual laboratories, in collaboration with engineering faculty at Northwestern University and at Oxford University who are willing to use them with their students. The first concerns thermodynamic cycles, an idealization used in the conceptual design of power plants, propulsion systems, refrigeration systems, and heat pumps. To master the design and analysis of thermodynamic cycles requires a deep understanding of a substantial body of thermodynamic. The second concerns feedback controllers. The concepts of feedback and control theory permeate modern science and engineering. While mathematical analyses are required to design optimal controllers, we believe that many of the important concepts of feedback could be grasped by high school students, given appropriate scaffolding. ***