Cognitive-Aware Generative AI System for Intelligent Code Debugging and Programming Instruction

This study proposes a generative AI-based system for programming error detection and heuristic correction, aiming to address instruction challenges in programming courses and promote a cognitive science-oriented transformation in programming education. Technically, the system adopts a three-dimensional collaborative mechanism - static analysis, dynamic tracing, and large model inference - to overcome traditional one-dimensional detection methods and accurately identify students' cognitive blind spots. Pedagogically, it pioneers a chain-of-thought heuristic strategy that guides students through a progressive hint sequence of "error attribution, logic reconstruction, and optimization validation,"fostering a complete cognitive process from "requirement analysis"to "algorithm design"and "code implementation."Equipped with a dynamic cognitive support engine, the system generates personalized feedback strategies based on learner profiling, achieving a shift from mechanical error correction to cognitive restructuring. Experimental validation demonstrates that the framework effectively promotes knowledge internalization and skill enhancement. Future developments will extend to the creation of visualization training modules for complex programming paradigms such as recursion and concurrency, continuously advancing the intelligent transformation of programming education.