基于大模型的教学智能体构建与应用研究

Design and Application of Pedagogical Agent with Foundation Model

随着生成式人工智能的快速发展,基于大模型的智能体已经逐步具备了多模态感知、检索增强生成、推理与规划、交互与进化等能力。该研究提出基于大模型的教学智能体的基本概念与框架,以“大模型”为技术核心,重点构建“教育任务设定”“教育任务规划”“教育能力实现与拓展”“教育内容记忆与反思”“交互协作与动态进化”多个功能模块,支持与多类型对象交互并实现动态进化,涵盖人机交互、多智能体交互以及环境交互。基于所提出的框架,研究以项目式学习任务为应用场景,阐述了教学智能体作为“助教智能体”和“同伴智能体”,在个性化驱动问题提出、项目方案共同设计、项目作品协作完成、项目作品多角色评价多个环节的作用及相关支撑技术。最后,研究进一步探讨了教学智能体的发展方向与未来展望。

With the rapid development of generative artificial intelligence,agent with foundation model has gradually acquired the capabilities of multimodal perception,retrieval and augmentation generation,reasoning and planning,interaction,and evolution.In this study,we propose the basic concept and framework of pedagogical agent with foundational model,with foundation model as the core,focusing on the construction of“educational task setting”,“educational task planning”,“educational capability realization and expansion”,“educational content memory and reflection”,“interactive collaboration and dynamic evolution”.It also supports interaction with multiple types of objects and dynamic evolution,covering human-computer interaction,multi-agent interaction,and environment interaction.Based on the proposed framework,this study takes the project-based learning task as an application scenario,and describes how the proposed pedagogical agent can be used as“assistant agent”and“peer agent”in personalization-driven problem posing,co-design of the project plan,collaborative completion of the project work,and multi-role evaluation of the project work.Finally,this study further discusses the future development and outlook of pedagogical agent.