- 期刊
混合场景下协作认知投入的多模态表征与分析路径研究
The Multimodal Characterization and Analysis Path of Collaborative Cognitive Engagement in Blended Scenario
摘要
随着研究者对学习成功背后因果机制的探寻，以及社会文化理论的深入人心，学习投入研究面临从“个体－行为”取向到“群体－认知”取向的转型，多模态学习分析也成为打开协作认知投入机制“黑箱”的钥匙。为破解这一问题，通过对不同视角的认知投入观点进行综合分析，深度阐述了混合场景下协作认知投入的概念，由此构建了涵盖激活系统、加工系统与反应系统的协作认知投入发生机制模型。进而选取文本、生理、语音、心理四种模态的学习过程数据，分别针对三个子系统构建了多模态表征框架与指标，以实现协作认知投入的量化评测。在多模态表征框架基础上，进一步明晰了混合场景下协作认知投入的分析路径，包括基于多模态特征融合的精准诊断、基于时间序列分析的演化规律、基于自我决定视角的动机归因以及基于可视化仪表盘的动态干预，由此构成了完整研究闭环，描绘出多模态数据支持的协作认知投入研究整体图景。
並列摘要
With the search for causal mechanisms behind learning success and the deepening of sociocultural theories, learning engagement research is facing a transition from an "individual-behavior" to a "group-cognitive" orientation, and multimodal learning analytics has become a key to unlock the "black box" of collaborative cognitive engagement. In this study, the concept of collaborative cognitive engagement in blended scenario was deeply elaborated through a comprehensive analysis of different perspectives on cognitive engagement, and an occurrence mechanism model of collaborative cognitive engagement covering the activation system, processing system, and reaction system was constructed. Then, the multimodal characterization framework and indexes were constructed for each system with four modalities: textual, physiological, phonological, and psychological, and the quantitative evaluation of collaborative cognitive engagement was realized. Based on the multimodal characterization framework, this study further clarified the analysis path of collaborative cognitive engagement, including accurate diagnosis based on multimodal feature fusion, evolutionary patterns based on time series analysis, motivational attribution based on self-determination perspective, and dynamic intervention based on visualization dashboard, thus forming a complete research loop and depicting the overall picture of collaborative cognitive engagement research under multimodal data.