詳細書目資料

7
0
0
0

結合YOLO與RAG架構之音樂學習互動教具與評量系統開發 = Development of an Interactive Music Learning Aid and Assessment System associated with YOLO and the RAG Framework / 李遠平.

館藏資訊

摘要註

本研究針對基礎音樂教育中所面臨的問題，例如：學生在理解學習內容上的困難、教師無法掌握學生的學習狀況、無法根據學生學習狀況安排適當課程內容等等。因此，提出一套整合人工智慧(Artificial Intelligence, AI)與物聯網(Internet of Things, IoT)技術的互動式數位教具與智慧評量系統來解決上述的問題。系統效仿STEAM教育模式為基礎，讓使用者親自使用學習系統來達到教學學習的效果。系統內共整合了三個子系統：互動教學學習系統、IoT教學訊息系統與RAG(Retrieval-Augmented Generation)評量系統。教學學習系統部署於Raspberry Pi 4 Model B上，採用YOLOv7 進行視覺辨識音樂符號，透過本研究之模型訓練結果，F1值曲線圖顯示信心閥值在0.566時，辨識準確度達0.99，在此基礎之下，搭配直覺式操作介面提升使用互動性；IoT系統負責學習資料傳輸與訊息橋接；評量系統則透過結構化資料檢索與大語言模型(Large Language Model, LLM)生成學習回饋，實現即時評量與個別化建議。實測系統運作結果顯示，在教學學習系統上，系統會即時給予學生相對應的聲音回饋，讓學生在學習時能夠及時將視覺與聲音聯繫在一起；而教師則可透過評量系統即時了解學生的學習狀況，並依此規劃適合學生學習狀況的課程內容。在未來規劃上，希望能夠實際運用在相關的教學場域上，以進一步驗證此套系統的實用性。. This study aims at the basic music education problems, such as students' difficulty in understanding learning content, teachers' inability to grasp students' learning status, and inability to arrange appropriate course content according to students' learning status. Therefore, an interactive digital teaching aid and smart assessment system integrating artificial intelligence (AI) and Internet of Things (IoT) technologies is proposed to solve the above problems. The system is based on the STEAM education model, allowing users to use the learning system personally to achieve teaching results. The system integrates three subsystems: interactive teaching and learning system, IoT teaching information system and RAG (Retrieval-Augmented Generation) assessment system. The teaching and learning system is deployed on a Raspberry Pi 4 Model B, using YOLOv7 for visual recognition of music symbols. The F1 value curve of the model training results of this study shows that when the confidence threshold is 0.566, the recognition accuracy is as high as 0.99. On this basis, the intuitive operation interface is used to improve the interactivity of use; the IoT system is responsible for the transmission of learning data and information bridging; the assessment system generates learning feedback through structured data retrieval and large language model (LLM), realizing real-time assessment and individualized suggestions. The results of the actual system operation show that in the teaching and learning system, the system will give students corresponding sound feedback in real time, allowing students to connect vision and sound in time when learning; and teachers can understand students' learning status in real time through the assessment system, and plan course content suitable for students' learning status accordingly. In future planning, it is hoped that it can be actually used in relevant teaching fields to further verify the practicality of this system..

回到最上