Empowering Learning-by-Teaching through Reinforcement Learning and Explainable AI

This project advances transformative, high-impact STEM education by developing low-cost, scalable, generalizable, and effective learning technologies that equip students for lifelong learning and meaningful collaboration with both humans and AI. Learning-by-teaching is a powerful instructional intervention with significant potential to enhance student learning. This project will design innovative, AI-powered online learning environments that promote the learning-by-teaching approach by enabling students to learn how to teach AI agents. Beyond enhancing academic performance, these AI-powered learning-by-teaching environments aim to develop essential cognitive and collaborative competencies, preparing students to succeed in a future increasingly shaped by AI technologies and human-AI partnerships. This project will design, implement, and empirically evaluate a novel integration of a Teachable Agent, reinforcement learning, and explainable AI to support students' acquisition of both procedural and conceptual knowledge within two online learning environments grounded in the learning-by-teaching paradigm. These environments aim not only to promote learning by teaching but also to help students learn how to teach across two STEM domains and learner populations: middle school algebra and college-level probability. The project team will (1) enhance the Teachable Agent framework with programming-by-demonstration to support integrated learning of both conceptual and procedural knowledge; (2) implement a dual-loop reinforcement learning framework that derives effective pedagogical strategies from expert student demonstrations and aligns them with learner preferences through human feedback; and (3) integrate cognitive and learning theories with generative and explainable AI to produce transparent, pedagogically meaningful explanations. This project is funded by the Research on Innovative Technologies for Enhanced Learning (RITEL) program that supports early-stage exploratory research in emerging technologies for teaching and learning. 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.