運用對話機器人於軟體工程課程之教學實踐研究

Abstract

軟體工程(Software Engineering)課程主要教導學生如何以系統性與規範化的流程與工具去開發和維護軟體，課程知識涵蓋層面相當廣泛，根據過去幾年軟體工程課程的教學經驗，我們發現了課程廣度、有限的助教資源、有限的教學互動、個人學習弱項等幾項教學議題，在此情況下，學生不易通盤整合課程所學，且受限於固定的教學時空(上課時段與教室)與有限的老師與助教資源，造成學生無法或不便提問與討論，老師或助教也很難深入了解每位同學的學習情況，導致無法全面性的滿足學生的學習需求。因此，我們希望透過對話機器人之建置與實施，可以針對上述問題找出合適的解決方案，本研究分析Chatbot的三個特性：互動性(Communicative)、知識性(Informative)、調適性(Adaptive)，基於這些特性建置一個新的教學Chabot平台：TABot，TABot平台將結合Chatbot引擎Rasa、課程知識地圖技術、雲端試算表服務與Discord平台，讓學生可藉由詢問TABot直接獲得重要資訊，如教學素材、範例程式碼、練習題、參考資源、課堂成績等，並可提供學生量身訂製之教學素材與考題，且提供團隊協同開發輔助功能，希望可提升學習主動性、降低學習門檻、改善學習弱項，最後提升整體之學習成效。 The Software Engineering course mainly teaches students how to develop and maintain software with systematic and disciplined processes and tools. This course covers a wide range of required knowledge. Based on the teaching experience of software engineering courses in the past few years, we have discovered several issues, such as the course's diverse knowledge, the limited resources of teachers and teaching assistants, the limited opportunity for interactions, and personal learning weaknesses. Teaching effectiveness is often limited by fixed class hours and classrooms; it is still unable to meet students' needs thoroughly. Teachers are also hard to understand the learning status of students. Therefore, we expected that we could find suitable solutions for the above issues by establishing and implementing chatbots. This study analyzes the three characteristics of the Chatbot: Communicative, Informative, and Adaptive. Based on these characteristics, we have built a new Chabot platform that supports teaching, called TABot. The TABot platform is integrated with the popular Chatbot engine, Rasa, course knowledge map, cloud spreadsheet services, and Discord platform. It allows students to obtain important information, such as teaching materials, sample documents, sample code, exercise labs, reference resources, and grades. It also provides personalized teaching materials and quizzes based on study history, and supports collaborative project development. It is expected that it can help students improve their learning motivation, lower the learning thresholds, improve their learning weaknesses, and improve overall study effectiveness.