Toward Personalizing Quantum Computing Education: An Evolutionary LLM-Powered Approach

TL;DR

The paper addresses the difficulty of delivering personalized, context-aware quantum computing education by proposing an Intelligent Teaching Assistant System (ITAS) that combines a knowledge-graph–augmented architecture with two specialized LLM agents: a Teaching Agent and a Lesson Planning Agent. A central Knowledge Graph enables persistent memory, cross-component context, and learning-path reasoning, while a user-facing tag system provides explicit control to mitigate LLM hallucinations and over-reliance. The approach offers dynamic, sub-lesson adaptation and data-driven insights from rich interaction data, aiming to bridge the gap between theory and practice in quantum education. If validated, this framework could enable scalable, personalized quantum instruction with robust planning and explainable, context-aware tutoring that adapts to diverse learner backgrounds and progress.

Abstract

Quantum computing education faces significant challenges due to its complexity and the limitations of current tools; this paper introduces a novel Intelligent Teaching Assistant for quantum computing education and details its evolutionary design process. The system combines a knowledge-graph-augmented architecture with two specialized Large Language Model (LLM) agents: a Teaching Agent for dynamic interaction, and a Lesson Planning Agent for lesson plan generation. The system is designed to adapt to individual student needs, with interactions meticulously tracked and stored in a knowledge graph. This graph represents student actions, learning resources, and relationships, aiming to enable reasoning about effective learning pathways. We describe the implementation of the system, highlighting the challenges encountered and the solutions implemented, including introducing a dual-agent architecture where tasks are separated, all coordinated through a central knowledge graph that maintains system awareness, and a user-facing tag system intended to mitigate LLM hallucination and improve user control. Preliminary results illustrate the system's potential to capture rich interaction data, dynamically adapt lesson plans based on student feedback via a tag system in simulation, and facilitate context-aware tutoring through the integrated knowledge graph, though systematic evaluation is required.

Figures (17)

• Figure 1: Lack of Context: Copilot gives a generic answer about classical communication, not specific to the quantum teleportation lesson.
• Figure 2: Missing Context: Copilot asks for code it should already have access to, highlighting system orthogonality.
• Figure 3: Static Learning Path: Copilot provides a generic explanation, not a personalized response to the student's confusion.
• Figure 4: LLM Dependence: Copilot provides a near-complete solution, potentially hindering deep learning.
• Figure 5: Contextual Awareness: The agent has access to a detailed and accurate lesson plan.