基於生成式人工智慧與擴增實境之科學實驗學習平台建置與成效分析;Design and Effectiveness Analysis of a Science Experiment Learning Platform Based on Generative Artificial Intelligence and Augmented Reality

NCUIR > College of Electrical Engineering & Computer Science > Graduate Institute of Computer Science and Information Engineering > Electronic Thesis & Dissertation > Item 987654321/98385

Please use this identifier to cite or link to this item: https://ir.lib.ncu.edu.tw/handle/987654321/98385

Title:	基於生成式人工智慧與擴增實境之科學實驗學習平台建置與成效分析;Design and Effectiveness Analysis of a Science Experiment Learning Platform Based on Generative Artificial Intelligence and Augmented Reality
Authors:	王維彤;Wang, Wei-Tung
Contributors:	資訊工程學系
Keywords:	生成式人工智慧;擴增實境;科學實驗;認知負荷;內在動機;自我效能;Generative Artificial Intelligence;Augmented Reality;Science Experiment;Cognitive Load;Intrinsic Motivation;Self-Efficacy
Date:	2025-07-28
Issue Date:	2025-10-17 12:42:57 (UTC+8)
Publisher:	國立中央大學
Abstract:	隨著擴增實境與人工智慧技術的快速發展，教育領域開始探索其在實驗教學中的應用潛力，特別是在提升學生探究能力與概念理解方面展現出顯著成效。本研究旨在建置一套融合生成式人工智慧與擴增實境技術的科學實驗學習系統，並探討其對中學生科學學習成效之影響。系統以微軟HoloLens 2頭戴式裝置為基礎，搭配實體光學儀器操作，讓學生得以在真實空間中觀察隨儀器移動的虛擬光線路徑變化，並透過生成式 AI 教學機器人即時獲得學習引導與回饋。系統設計亦融合 Kolb 經驗學習理論，依據「具體經驗」、「反思觀察」、「抽象概念化」與「應用驗證」四個階段，引導學生進行完整的學習歷程。為驗證本系統之有效性，本研究採準實驗設計，邀請某國中七年級學生共 57 人參與，分為AI 引導學習組（28 人）與單人自主學習組（29 人）。兩組皆接受相同課程與教學內容，單人自主學習組在不具人工智慧輔助的擴增實境環境中進行光學實驗學習；AI 引導學習組則使用本系統輔助進行光學實驗與概念學習。資料收集包含學習成效前後測、學習動機與自我效能問卷、以及學生行為紀錄分析，以瞭解系統在不同層面的影響。研究結果顯示，AI 引導學習組學生在整體學習成效上有顯著提升，顯著優於自我學習組。然而，在內在動機與自我效能方面，雖有提升趨勢，但未達統計顯著水準。進一步分析發現，AI 引導學習組學生在抽象概念理解與探究操作策略方面的進步幅度明顯。此外，行為分析亦發現，AI 引導學習組學生展現出更多主動操作與反思性思考行為，顯示 AI 引導能有效促進學生由具體經驗轉化為高階概念理解，並應用於問題解決情境中，顯現出本系統在教學互動性與學習成效提升上的應用潛力。;With the rapid advancement of augmented reality (AR) and artificial intelligence (AI) technologies, the education sector has begun exploring their potential applications in experimental teaching, particularly in enhancing students′ inquiry skills and conceptual understanding. This study aims to develop a science experiment learning system that integrates generative AI and AR technologies, and to investigate its impact on middle school students′ science learning outcomes. The system is based on the Microsoft HoloLens 2 head-mounted device, combined with hands-on manipulation of physical optical instruments, enabling students to observe and manipulate virtual light paths in a real-world environment. Through interactions with a generative AI chatbot, students receive real-time guidance and feedback. The system is also designed in accordance with Kolb′s experiential learning theory, guiding students through the four stages of "concrete experience," "reflective observation," "abstract conceptualization," and "active experimentation." To evaluate the effectiveness of the system, a quasi-experimental design was adopted. A total of 57 seventh-grade students from a junior high school participated in the study, divided into an AI-guided learning group (28 students) and an individual self-learning group (29 students). Both groups received the same curriculum and instructional content. The individual self-learning group conducted optical experiments using an AR environment without AI support, while the AI-guided group used the proposed system to engage in optical experiments and conceptual learning. Data collection included pre- and post-tests of learning performance, questionnaires on learning motivation and self-efficacy, and behavioral log analyses to examine the system′s impact on multiple dimensions. The results indicated that the AI-guided learning group demonstrated a statistically significant improvement in overall learning performance, outperforming the self-learning group. However, while there was an upward trend in intrinsic motivation and self-efficacy, these differences did not reach statistical significance. Further analysis revealed that the AI-guided group showed notable gains in abstract conceptual understanding and inquiry-based operational strategies. Additionally, behavioral analysis showed that students in the AI-guided group exhibited more proactive manipulation and reflective thinking behaviors. These findings suggest that AI guidance effectively facilitates the transformation from concrete experiences to higher-order conceptual understanding and application in problem-solving contexts, highlighting the system′s potential in enhancing instructional interactivity and learning effectiveness.
Appears in Collections:	[Graduate Institute of Computer Science and Information Engineering] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	7	View/Open

社群 sharing

Loading...