Generalizing Data-Driven Technologies to Improve Individualized STEM Instruction by Intelligent Tutors

智能导师是使用人工智能等技术为学习者提供个性化指导和反馈的计算机系统。他们已经被证明在提高学生学习方面非常有效，很大程度上是因为他们有能力在学生学习时为他们提供个性化的、适应性的支持。本项目旨在推进智能家教的建设技术。为此，它将开发一个构建智能导师的框架，使其更容易创建适合个体学习者的个性化学习体验。由此产生的智能辅导系统将使用数据来确定系统应该采取什么行动，系统应该在什么时候采取行动，并解释为什么这些行动应该改善学习。这种新的智能辅导系统通过为复杂的STEM主题提供方便、低成本、个性化的指导，有可能改变STEM的学习环境。该项目将开发一个可推广的数据驱动框架，以在三个STEM领域（逻辑、概率和编程）和干预类型（工作示例、错误示例和帕森斯问题）中引入广泛的教学干预和强大而灵活的教学决策政策。该系统将被设计为允许在广泛的stem相关领域使用智能导师。该系统将使用先进的机器学习和数据挖掘技术来生成教学干预、混合主动教学政策和人在循环解释。这项研究将推进数据驱动的混合主动决策生成的知识，平衡学生的代理意识和他们在关键决策点对有效教学干预的需求。通过一系列的实证研究，将新干预措施与现有的辅导系统进行比较，以确定对学习成果、代理、个性化和有效互动的影响，从而评估最终系统的有效性。本项目由美国国家科学基金改进本科STEM教育计划：教育与人力资源资助。IUSE: EHR计划支持研究和开发项目，以提高所有学生STEM教育的有效性。该项目处于“参与学生学习”轨道，通过该轨道，该项目支持有前途的实践和工具的创建、探索和实施。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

Intelligent tutors are computer systems that use technologies such as artificial intelligence to provide learners with personalized instruction and feedback. They have been shown to be highly effective at improving students' learning, largely because of their ability to provide students with individualized, adaptive support as they learn. This project aims to advance the technology for building intelligent tutors. To do so, it will develop a framework for building intelligent tutors that makes it easier to create personalized learning experiences that adapt to individual learners. The resulting intelligent tutoring system will use data to determine what actions the system should take, when the system should act, and explain why these actions should lead to improved learning. This new intelligent tutoring system has the potential to transform STEM learning environments by providing accessible, low-cost, individualized instruction for complex STEM topics. This project will develop a generalizable data-driven framework to induce a wide range of instructional interventions and robust, yet flexible pedagogical decision-making policies across three STEM fields (logic, probability, and programming) and types of intervention (worked examples, buggy examples, and Parsons problems). The system will be designed to allow use of the intelligent tutor in a wide range of STEM-related domains. The system will use advanced machine learning and data mining techniques to generate instructional interventions, mixed-initiative pedagogical policies, and human-in-the-loop explanations. This research will advance knowledge about data-driven generation of mixed-initiative decision making that balances a student’s sense of agency with their need for effective instructional interventions at critical decision points. The efficacy of the resulting system will be evaluated via a series of empirical studies comparing the new interventions with existing tutoring systems to determine the impact on learning outcomes, agency, personalization, and effective interactions. This project is supported by the NSF Improving Undergraduate STEM Education Program: Education and Human Resources. The IUSE: EHR program supports research and development projects to improve the effectiveness of STEM education for all students. This project is in the Engaged Student Learning track, through which the program supports the creation, exploration, and implementation of promising practices and tools.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.