摘要
生成式人工智能赋能数学教育是当前研究的热点课题,但是大语言模型的逻辑推理能力有待提升,在数学解题上经常表现不佳.研究尝试将大语言模型与动态可视化软件整合,在基于自然语言环境的数学可视化方面做开拓性探索.把二者结合并在专业数据库上训练所生成的技术产品统称为AI数学助手,在介绍整合原理的基础上,以GPT3.5和GeoGebra为代表,给出实现AI数学助手的一个实例.AI数学助手在数学教育中的可能应用是自动生成图形化支架、明确解题/作图步骤、帮助解读GeoGebra指令等.AI数学助手会为人工智能时代下的教育公平发展、跨学科课程建设、数学教学设计和教师专业发展带来启示.
Empowering mathematical education with the new generation of generative artificial intelligence,represented by large language models,is a research topic that has garnered considerable attention.However,the low level of logical reasoning ability of large language models makes them perform poorly in mathematical problem-solving.This study attempts to integrate large language models with dynamic visualization software to make pioneering explorations in mathematical visualization under the natural language environment.We refer to the technology product created by integrating large language models with dynamic visualization software and training it on professional databases as the“AI Math Assistant”.First,we introduce its design principles.Then we implement an example of the AI Math Assistant by using GPT-3.5 and GeoGebra as representatives of large language models and dynamic visualization software.Finally,we discuss the potential applications of the AI Math Assistant in mathematics education and explore its implications for developing educational equity,designing mathematics instruction,constructing inter disciplinary curriculum and teachers’professional development in the era of artificial intelligence.
作者
邵铭宇
魏亮
李均强
苏洪雨
谭枫
SHAO Ming-yu;WEI Liang;LI Jun-qiang;SU Hong-yu;TAN Feng(School of Mathematical Science,South China Normal University,Guangdong Guangzhou 510631,China;Shenzhen Blue Swallow Education Technology Co.,Ltd.,Guangdong Shenzhen 518052,China)
出处
《数学教育学报》
北大核心
2025年第4期27-35,共9页
Journal of Mathematics Education
基金
广东省哲学社会科学规划2023年度学科共建项目——人工智能赋能的课堂教学评价研究:基于课堂教学特征AI分析平台(GD23XJY55)
2024广东省本科高校教学质量与教学改革工程项目——基于大学中学协同的师范生数智可视化教学技能培养模式的设计研究(粤教高函〔2024〕30)
国家自然科学基金面上项目——关于动力系统复杂性问题的研究(11871228)。
关键词
人工智能
动态可视化软件
大语言模型数学定制
数学教育
artificial intelligence
dynamic visualization software
mathematical customization of large language models
mathematics education
