《电路智导》AI学伴的创建研究

Research on the Creation of an AI Learning Companion for Circuit Intelligence Guide

摘要: 针对电气信息类专业电路课程教学中存在的教学课时紧张、内容有难度等教学挑战，以及通用人工智能模型在特定专业领域中的适用性与针对性有限，本研究设计了一款AI学伴智能体，作为电路课程的自主学习方法辅助工具。该AI学伴依托Coze平台构建智能体，结合工作流等相关工具，整合了结构化、专业性较强的电路知识库(涵盖基本原理、典型电路分析设计、常见问题解答、练习题与示例)。本文按照“检索增强生成(RAG)”框架，在大语言模型的生成能力基础上嵌入可检索的外部知识库，从而提升《电路智导》智能体的准确性与针对性。同时，本文系统讨论了技术选型背后的机理、RAG工作原理、知识库组织形式对检索效果的影响，以及低代码平台的优势与局限性，进一步阐述了本研究的定位和贡献。通过整合知识库检索与多模态数据(文字、图片链接)生成能力，该学伴能够提供精准答疑、高质量习题生成及丰富的学习资源支持。该学伴显著提升了学生的自主学习效率、理解深度与课堂参与度，并减轻了教师负担，为AI学伴领域的专业化应用提供了有价值的实践案例。

Abstract: To address the teaching challenges in circuit courses for electrical and information-related majors, such as limited instructional hours and difficult content, as well as the limited applicability and specificity of general-purpose AI models in specialized domains, this study designs an AI learning companion agent as an auxiliary tool for self-directed learning in circuit courses. This AI companion is developed on the Coze platform, leveraging workflow-related tools to integrate a structured and highly specialized circuit knowledge base (covering fundamental principles, typical circuit analysis and design, frequently asked questions, exercises, and examples). Following the Retrieval-Augmented Generation (RAG) framework, this study enhances the accuracy and relevance of the “Circuit Guide” intelligent agent by embedding an externally retrievable knowledge base into the generative capabilities of a large language model (LLM). Additionally, the paper systematically discusses the rationale behind the technical choices, the working principles of RAG, the impact of knowledge base organization on retrieval effectiveness, and the advantages and limitations of low-code platforms, and further elaborates on the positioning and contributions of this research. By integrating knowledge base retrieval with multimodal data generation (text and image links), the AI companion provides precise Q&A support, high-quality exercise generation, and comprehensive learning resources. The proposed AI companion significantly improves students’ self-learning efficiency, depth of understanding, and classroom engagement, while also reducing instructors’ workload, and offers a valuable practical case for the specialized application of AI learning companions.