An Intelligent Assistant to Support Teachers and Students in Simulation-Based Science Learning

本项目将开发一个基于人工智能的对话框架（iLab），在基于模拟的科学实验背景下，为中学理科学生和教师创造基于对话的互动实验室体验。该框架的一个关键组成部分是一个智能对话代理（SimPal），它将与教师进行对话，询问他们与模拟实验相关的教学目标，并使用计算表示存储这些目标。然后，代理将使用此表示来促进和调解与学生进行的交互式对话（由最先进的大型语言模型提供支持），以增强他们的学习体验。代理会主动询问学生的反思问题，为学生提供实时定制的反馈，跟踪学生的进度，然后分析学生的反应并反馈给老师。与现有的智能辅导系统和教学对话代理不同，该框架将与任何领域的任何现成的第三方模拟一起工作，并由任何教师或美国学区使用，这是该项目的一个独特之处。此外，教师将作为合作伙伴开发和部署这项技术。因此，该项目有望为学生的学习做出独特的贡献，从而在美国学校中具有广泛的影响力和吸引力。在这个项目下开发的框架将作为老师和每个学生之间的桥梁，在大教室里，老师不可能亲自支持每个学生。它将使学生参与到一场有意义的对话中，从学生的回答和教师的教学目标中获得信息，同时适应学生和教师，从而引导学生对潜在的科学概念和原理有更深的理解。评估研究将在威斯康辛州的多个教室进行，教师将把这项技术整合到常规课程中。该项目将为智能支持指明一条道路，旨在改善科学课堂学习面临的四大挑战：(1)随时随地获得高质量的学习材料；(2)适应性，确保不同背景、经历和能力的学生根据教师的目标获得定制的反馈；(3)可重复性，学生可以在短时间内进行多次实验，并获得实时反馈；(4)可扩展性，以减轻运行面向查询的科学调查的负担。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。

This project will develop an artificial intelligence-based conversational framework (iLab) to create dialog-based interactive laboratory experiences for middle school science students and teachers in the context of simulation-based science experiments. A key component of the framework is an intelligent conversational agent (SimPal) that will engage with teachers in a dialog to solicit their instructional goals associated with simulation experiments and store them using a computational representation. The agent will then use this representation to facilitate and mediate an interactive dialog (powered by state-of-the-art large language models) with students as they run experiments to enhance their learning experience. The agent will proactively ask students reflection questions, provide them with real-time customized feedback, track students' progress, and then analyze their responses and report back to the teacher. Unlike existing intelligent tutoring systems and pedagogical conversational agents, the framework will work with any off-the-shelf third-party simulations in any domain and be used by any teacher or U.S. school district, a unique feature of this project. Further, teachers will work as partners in developing and deploying this technology. As such, the project is expected to make unique contributions to benefit student learning and, thereby, have a broad reach and appeal in U.S. schools.The framework developed under this project will serve as a bridge between the teacher and each student in large classrooms where it is not possible for the teacher to support each student personally. It will engage the student in a meaningful dialog, informed both by the responses from the student and by the teacher’s instructional goals while adapting to both the student and the teacher, thereby guiding the student toward a deeper understanding of the underlying scientific concepts and principles. Evaluation studies will be conducted in multiple Wisconsin classrooms where teachers will integrate the technology into the regular curriculum. The project will chart a path toward intelligent support designed to ameliorate four major challenges to learning in the science classroom: (1) access to high-quality learning materials at any time and place; (2) adaptability to ensure that students with a variety of backgrounds, experiences, and abilities receive custom feedback based on the teacher’s goals; (3) repeatability so students can run several experiments in a short time and receive real-time feedback; and (4) scalability to reduce the burden of running inquiry-oriented science investigations.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.