基于深度学习的cMOOC参与者交互水平自动评估研究

Research on Automatic Evaluation of Interaction Level of cMOOCParticipants Based on Deep Learning

交互是联通主义学习的核心和取得成功的关键。现有交互水平评估主要依靠人工编码实现,存在时间滞后性和耗时耗力等局限,难以实时反映交互水平状况。因此,本研究以国内首门联通型慕课“互联网+教育:理论与实践的对话Ⅱ”为案例,运用人工智能领域中的深度学习算法,通过模型构建、模型检验和模型应用三个阶段实现了对联通慕课参与者交互水平的自动化评估。研究发现,在联通主义学习中:1)学习者成为交互主体,多元主体驱动课程交互;2)寻径和意会是决定联通主义学习成效的关键。本研究一方面检验和发展了原有联通主义交互规律研究,另一方面实现了联通主义学习情境中参与者交互水平的自动化评估,为联通主义学习评价和课程评价提供了工具支撑,探索了人工智能技术在改进教育研究和实践方面的价值和可能性。

Interaction is the core and key to the success of connectionist learning.The existing interaction level evaluation mainly depends on manual coding,which has limitations such as time lag,time-consuming and labor-consuming,so it is difficult to reflect the interaction level in real time.Therefore,this study takes the first domestic cMOOC“Internet+Education:Dialogue between Theory and Practice II”as a case,and uses the deep learning algorithm in the field of artificial intelligence to realize the automatic evaluation of the interaction level of cMOOC participants through three stages:model construction,model checking and model application.At the same time,it is found that in connectionism learning:1)learners become interactive subjects,and multiple subjects drive curriculum interaction;2)pathfinding and understanding are the key to determine the effectiveness of connectionism learning.On the one hand,this study tests and develops the original research on the interaction law of connectionism,on the other hand,it realizes the automatic evaluation of the interaction level of participants in the learning situation of connectionism,which provides a tool support for the evaluation of connectionism learning and curriculum evaluation.This paper explores the value and possibility of artificial intelligence technology in improving educational research and practice.