基于多模态交互信息的协作学习投入分析

Analysis of Collaborative Learning Engagement based on Multimodal Interaction Information

协作学习投入面临多维、动态数据采集与多元数据融合分析的挑战,而多模态交互分析能够提取并融合动态、多维度的投入信息,对表征协作学习交互过程具有重要价值。基于此,文章构建了基于多模态交互信息的协作学习投入分析指导框架,并从投入水平的整体性表征、投入状态的多维性分析和投入信息的时序性挖掘等角度,阐述了协作学习投入分析应考虑的关键问题。此外,文章还提出了包含多模态信息采集、多模态特征提取、多模态特征融合分析等三大模块构成的多模态交互信息分析路径。研究结论显示,多模态交互信息能实现对协作交互过程全方位、细粒度地记录,进而对小组协作学习投入的状态进行动态表征与持续分析。未来研究可能的方向在于基于智能语音处理技术的交互会话分析、基于仪表盘的投入信息可视化呈现以及基于智能代理的协作投入适应性支持。

Collaborative learning engagement is faced with the challenge of multi-dimensional and dynamic data collection and multi-dimensional data fusion analysis,while multimodal interaction analysis can extract and fuse dynamic and multi-dimensional engagement information,which is of great value for characterizing the interactive process of collaborative learning.Then,this paper constructed a framework for collaborative learning engagement analysis based on multimodal interaction information,and expounded the key issues that should be considered in collaborative learning engagement analysis from the perspectives of overall characterization of engagement level,multidimensional analysis of engagement state and temporal mining of engagement information.In addition,this paper also proposed a multimodal interaction information analysis path,which consisted of three modules:multimodal information collection,multimodal feature extraction,and multimodal information fusion analysis.It was concluded that the multimode interaction information can record the collaborative interaction process in an all-round and fine-grained way,and further dynamically characterize and continuously analyze the state of group collaborative learning engagement.Future research may focus on interactive session analysis based on intelligent speech processing technology,dashboard-based visual presentation of engagement information,and collaborative engagement adaptive support based on intelligent agent.