摘要
生成式人工智能技术的发展使得机器的文本生成、人机对话和逻辑推理能力得到极大提升,也催生了对话式人机协同学习这一新型学习方式。对话式人机协同学习旨在通过人机之间的多轮对话和双向反馈,激发学生与机器的思维碰撞,引发学习者的认知冲突和自主建构,逐步达成对所学内容的共同认识,探索对复杂问题的最优解,实现学生智慧和机器智能的共同增长。在此基础上,该文系统阐述了对话式人机协同学习的一般过程,围绕学习目标的生成性、学习内容的适应性、学习主体的互惠性、人机角色的互换性讨论了对话式人机协同学习的核心特征,并从认知辅助式学习、思维启发式学习、自由探索式学习、人机论辩式学习四个方面探讨了对话式人机协同学习的典型模式。未来对话式人机协同学习的开展需要进一步推动大模型技术研发、加强提示语工程建设、健全自适应反馈机制、验证对话式学习成效、提高学生智慧学习力。
The development of generative artificial intelligence technology has greatly improved the text generation,human-machine dialogue,and logical reasoning abilities of machines,and has also given rise to a new type of learning method called conversational human-machine collaborative learning.Dialogue based human-machine collaborative learning aims to stimulate the collision of thinking between students and machines through multiple rounds of dialogue and bidirectional feedback,trigger cognitive conflicts and autonomous construction among learners,gradually achieve a common understanding of the learned content,explore the optimal solutions to complex problems,and achieve the joint growth of student intelligence and machine intelligence.On this basis,the article systematically elaborates on the general process of conversational human-machine collaborative learning,discusses the core characteristics of conversational human-machine collaborative learning around the generation of learning objectives,adaptability of learning content,reciprocity of learning subjects,and interchangeability of human-machine roles,and explores typical models of conversational human-machine collaborative learning from four aspects:cognitive assisted learning,thinking heuristic learning,free exploration learning,and human-machine argumentative learning.The development of future dialogue based human-machine collaborative learning needs to further promote the development of large model technology,strengthen the construction of prompt language engineering,improve adaptive feedback mechanisms,verify the effectiveness of dialogue based learning,and improve students’intelligent learning ability.
作者
王一岩
刘淇
郑永和
Wang Yiyan;Liu Qi;Zheng Yonghe(Research Institute of Science Education,Beijing Normal University,Beijing 100875;School of Computer Science and Technology,University of Science and Technology of China,Hefei 230027,Anhui)
出处
《中国电化教育》
CSSCI
北大核心
2024年第11期21-27,共7页
China Educational Technology
基金
国家重点研发计划“文化科技与现代服务业”重点专项“面向终身学习的个性化‘数字教师’智能体技术研究与应用”课题三“面向终身学习的自适应教育关键技术”(课题编号:2021YFF0901003)研究成果。
关键词
人机协同学习
对话式学习
生成式人工智能
论辩式学习
human-machine collaborative learning
dialogue based learning
Generative Artificial Intelligence
argumentative learning
作者简介
王一岩:博士,研究方向为智能教育、人机协同教育、教育信息科学与技术;刘淇:教授,博士,博士生导师,研究方向为大数据、人工智能、智能教育;通讯作者:郑永和:教授,硕士,博士生导师,院长,研究方向为教育信息科学与技术、科技与教育政策、科学教育。
