发表机构
University of Jyväskylä; Faculty of Information Technology（于韦斯屈莱大学; 信息技术学院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本文提出一个学习分析框架，利用量子电路模拟器的响应日志识别瓶颈任务和错误模式，支持课程重新设计，并在492名学习者数据上验证。
AI 中文摘要
交互式量子电路模拟器越来越多地被用于支持来自不同学科背景的学习者的入门量子计算教育。这些环境还可以生成所提交电路的详细响应日志，但此类数据很少被转化为教师可用于课程重新设计和开发的证据。我们提出了一个用于量子电路模拟器任务的学习分析框架，并报告了其在TIM学习环境中提供的入门量子计算课程的匿名日志上的任务级分支。该任务分支通过参与度、完成度、首次尝试准确性、有效尝试次数、会话化活动时间以及常见正确和错误构建的观察答案空间报告来表征练习。数据集涵盖2024-2025年队列的492名学习者、47,533次有效提交和21个必修量子电路练习。该框架识别出瓶颈任务、广泛的解决方案空间以及仅从最终分数无法看到的重复错误答案模式。其最独特的输出是观察答案空间的纯文本表示，旨在使开放式电路提交对教师直接可解释。我们展示了响应日志如何揭示可操作的任务设计问题并支持具体的课程重新设计。学习者级分支仅作为正在进行的研究扩展保留，并非本文的核心实证贡献。
英文摘要
Interactive quantum circuit simulators are increasingly used to support introductory quantum computing education for learners from varied disciplinary backgrounds. These environments can also generate detailed response logs of submitted circuits, but such data are rarely transformed into evidence that instructors can use for course redesign and development. We present a learning analytics framework for quantum circuit simulator tasks and report its task-level branch on anonymised logs from an introductory quantum computing course delivered in the TIM learning environment. The task branch characterises exercises through participation, completion, first-attempt accuracy, valid-attempt counts, sessionised active time, and observed answer-space reporting of common correct and incorrect constructions. The dataset covers the 2024-2025 cohort of 492 learners, 47,533 valid submissions, and 21 obligatory quantum circuit exercises. The framework identifies bottleneck tasks, broad solution spaces, and recurring wrong-answer patterns that are not visible from final scores alone. Its most distinctive output is a plain-text representation of the observed answer space, designed to make open-ended circuit submissions directly interpretable to instructors. We show how response logs can reveal actionable task-design issues and support concrete course redesign. A learner-level branch is retained only as an ongoing research extension and is not a central empirical contribution of the present paper.
Comments10 pages, 8 figures. Accepted and presented as a technical paper at the Quantum Science and Engineering Education Conference (QSEEC 2026), held in conjunction with IEEE Quantum Week 2026
