【作 者】：
翟雪松，张丽洁，夏亮亮，徐 鑫，朱 强
【关 键 词 】：
生成式人工智能； 逆向工程； 人机协作学习； 复杂问题解决能力； 编程教育
【栏 目】：
课程与教学
【中文摘要】：
大模型为学习者提供跨模态的学习资源，同时也为创新人机协同教学模式提出了更高要求。研究引入了逆向工程教学思维，分析了其在流程与机理上与生成式人工智能的相契互补性，并基于自主开发的逆向工程编程学习平台，开展了探索性编程教学实验。通过LDA主题词抽取和人机协作感知因子分析，研究挖掘出该模式下人机协作五类行为和情感取向。此外，问卷结果显示学习者在此教学模式下表现出较高的感知偶然性、人机协作感知以及持续学习意愿，但人机信任度处于中位水平。结合主题词分析，研究也提出未来人机协作的优化方向：通过逆向工程引领人机协作，降维拆解复杂问题；构建多智能体生态，提高多人—多机群体协作效能；塑造新型人机劳动关系，发展新智生产力。研究为未来人工智能协作学习提供了理论依据和数据参考，也提出了未来研究进一步改进的思路和方法。
【英文摘要】：
Large models provide learners with cross-modal learning resources, but also puts forward higher requirements for innovative human-computer collaborative teaching models. This study introduced reverse engineering teaching thinking, analyzed its compatibility and complementarity with generative artificial intelligence (GAI) in terms of process and mechanism, and carried out exploratory programming teaching experiment based on the self-developed reverse engineering programming learning platform. Through LDA topic word extraction and an analysis of human-computer collaboration perception factors, five types of behaviors and emotional orientations of human-computer collaboration under this model are explored. In addition, the questionnaire results indicate that learners exhibit a higher perceived contingency, human-computer collaboration perception and willingness to continue learning in this teaching mode, but the human-computer trust is at a medium level. Combined with the analysis of topic words, this study also proposes the optimization direction of human-computer collaboration in the future: leading human-computer collaboration through reverse engineering to decompose complex problems by dimensionality reduction; constructing a multi-agent system to enhance the efficiency in multi-person and multi-agent cooperation; shaping a new human-computer labor relation to develop new intelligent productivity. This study provides theoretical foundation and data reference for future artificial intelligence collaborative learning, and also puts forward ideas and methods for further improvement in future research.