Collaborative Research: An Explainable AI Environment for Adaptive Learning and Practice of Quantum Computing

Quantum computing has the potential to revolutionize fields from drug discovery and materials science to cybersecurity and AI. Realizing this potential creates an urgent need to prepare a quantum-ready workforce. Yet quantum computing is difficult to learn and teach: unlike most STEM subjects, where instruction aims to replace misconceptions with scientific knowledge, it requires students to reorganize their established knowledge of classical computing, mathematics, and physics, which often conflicts with the behavior of quantum systems. To support students through this reorganization, the project will extend theories of learning science to quantum computing and build upon recent advances in generative AI to develop an explainable AI-based learning environment. The environment will generate tailored and verified practice materials on demand, provide fine-grained guidance and feedback as students study quantum concepts and write quantum programs, and make its strategies transparent, enabling instructors to adjust them as needed. It will be piloted in quantum computing courses, providing approximately 120 students with personalized learning while giving instructors and teaching assistants insight into the strategies that most effectively support this reorganization. The comprehensive environment will be released openly for institutions with limited quantum expertise to adopt. This project will help prepare a quantum-ready workforce and strengthen the nation's long-term leadership in this field. The project will investigate how students reorganize classical knowledge into quantum reasoning and will use that understanding to build an AI environment for practicing quantum concepts and programming with tailored support. It will be built around three components. A conceptual component will support understanding of quantum principles, using deep reinforcement learning, guided by conceptual change theory, to focus on the moments most important for learning, with a vision-language model that aligns multiple representations of quantum computing: mathematics, circuits, visualizations, and natural language. A strategic component, grounded in self-regulated problem-solving theory, will scaffold quantum programming practice, adapting its support through natural language feedback and generating explanations and code verified by tools such as quantum circuit equivalence checkers. An explainability component will identify the model components responsible for particular support strategies, making the system's decisions transparent and controllable. The environment will be evaluated through classroom deployments that compare it with standard instruction. Data collection will include pre- and post-assessments, along with interaction logs, observations, and interviews. Statistical, qualitative, and network-based analyses will assess the effectiveness of the environment and examine students' conceptual and strategic reasoning patterns. The project will advance learning science theories for quantum computing and contribute new methods for adaptive and transparent AI for learning. 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.