🤖 AI Summary
Existing approaches to Computational Thinking Problems (CTPs) in education lack systematic methods for assessment and generation.
Method: This paper proposes the first dual-track evaluation framework integrating formal semantic modeling with educational cognitive dimensions. It combines program static analysis, knowledge graph modeling, Cognitive Task Decomposition (CTD), and rule-guided templated synthesis to enable automated CTP diagnosis, difficulty calibration, and targeted generation.
Contribution/Results: The framework establishes an interpretable and scalable paradigm for CTP analysis and design, overcoming the limitations of experience-driven item authoring. Empirical validation on K–12 programming problem banks demonstrates a 42% improvement in problem coverage, a 3.1× increase in teacher design efficiency, and significantly enhanced transfer of students’ computational thinking skills.