【作 者】：
公 雪， 马正婧， 谢中取， 刘秉新
【关 键 词 】：
人工智能； 机器人； 交互设计； 教育数字化； 人机协同； 个性化学习
【栏 目】：
学习环境与资源
【中文摘要】：
虚拟机器人作为通用人工智能新型载体，以其显著的多模态交互优势而受到重视，以支持学习者自主学习的方式赋能教育数字化转型。为提升其在针对不同课程的智能协同学习精准性，文章提出了基于MTV模型的虚拟机器人智能指导交互系统，通过定量研究方法，旨在分析虚拟机器人交互的耦合关系。文章以K12阶段不同类型课程的学习者学习数据为个案，结合虚拟机器人交互系统设计中用户逻辑、介质逻辑以及过程逻辑三层交互逻辑，对学习者与课程交互、学习者与系统交互以及学习者与虚拟机器人交互这三类交互形式进行分析，创建学习者情感偏好标签。通过结合ChatterBot机器学习算法挖掘出K12阶段五类学习者在学习取向、学习风格及学习需求三个维度的状态特征，提升中国式教育环境中集成虚拟机器人的教育交互精准度。文章发现，虚拟机器人智能指导交互系统在个性化学习路径指导、多维度课程资源推荐方面优势明显，为虚拟机器人发挥“智慧学伴”同伴效应及智能学习跟踪等提供实践层面参考。
【英文摘要】：
Virtual Robots, as a novel carrier of general artificial intelligence, have been emphasized for their significant multi-modal interaction advantages, empowering the digital transformation in education by supporting learners' autonomous learning. In order to improve its precision in intelligent collaborative learning for various courses, this study proposed a virtual robot intelligent guidance interaction system based on the MTV model, which aimed to analyze the loosely coupled relationship of virtual robot interactions through quantitative research methods. This study used the learning data of learners in different types of courses in K12, analyzed three types of interactions, namely learner-course interaction, learner-system interaction, and learner-virtual robot interaction together with the user logic, medium logic, and process logic in the design of the virtual robot interaction system, and constructed learners emotional preference labels. Combined with the ChatterBot machine learning algorithm, the state characteristics of five types of learners in K12 in the three dimensions of learning orientation, learning style and learning needs were excavated so as to enhance the precision of educational interaction with integrated virtual robots in Chinese educational environments. This study finds that the virtual robot intelligent guidance interaction system has significant advantages in personalized learning path guidance and multidimensional curriculum resource recommendation, providing practical references for the virtual robot to leverage the peer effect of "smart learning companion" and intelligent learning tracking.