科學學習強調學生主動建構知識與理解科學概念，電腦模擬因具備視覺化與
高互動性的特質，已廣泛應用於支持學生在科學活動中進行觀察與推理。隨著教學逐漸轉向線上環境，教師難以即時掌握學生的理解狀況與學習歷程，導致學生在操作模擬時常面臨概念迷失與學習瓶頸。為解決此問題，本研究建置一套結合大型語言模型與模擬平台的智慧助教系統，透過提示詞設計與多模態互動，協助學生在科學學習歷程中獲得即時與個別化的學習支援。
本研究建構的智慧助教，搭配CoSci電腦模擬平台，打造具備思維鏈推理能
力與結構化提示詞框架的教學對話模組。系統可根據學生的操作、模擬畫面與對話內容進行即時判斷與教學引導，並在特定情境下啟動動作模組與評分模組，以回應學生的學習需求。研究以臺灣北部高中一年級學生為對象，共22人，進行一日內的科學學習活動，並設計具外在動機機制（積分條）與無動機設計兩種版本，比較其對學生表現與互動歷程的影響。
研究結果顯示，缺乏外在動機機制的學生反而展現出更顯著的學習成效，智
慧助教能有效促進其主動提問、批判性思考與知識整合等深層學習行為。此外，智慧助教的回應品質與推理能力亦受提示詞設計影響，具情境整合與概念導向的提示詞更能提升其教學判斷的精準性與互動表現，進而有助於學生的知識建構與理解深化。分析也發現，學生在互動過程中展現的知識處理行為與智慧助教所推論的理解，部分與其學習成果具有顯著關聯。最後根據系統現有限制提出技術改進建議，作為未來智慧助教應用於科學教育之參考。

Science learning emphasizes students’ active knowledge construction and conceptual understanding. Due to their visual and interactive nature, computer simulations have been widely used to support students' observation and reasoning in scientific activities. However, as instruction increasingly shifts to online environments, it becomes difficult for teachers to monitor students’ understanding and learning processes in real time, often leading to conceptual confusion and learning bottlenecks during simulation-based learning. To address this issue, this study developed an intelligent tutoring system that integrates a large language model with a simulation platform. By incorporating prompt design and multimodal interaction, the system provides students with immediate and individualized learning support throughout their science learning process.
The intelligent tutor, built upon the CoSci simulation platform, features a dialogue module equipped with Chain-of-Thought reasoning and a structured prompt framework. The system responds to students’ simulation operations, screen visuals, and dialogues to deliver real-time instructional guidance. When appropriate, it also activates action and assessment modules to address students’ learning needs. This study was conducted with 22 first-year senior high school students in northern Taiwan who participated in a one-day science learning activity. Two interface versions were compared—one with an external motivational mechanism (a score bar) and one without—to examine their impact on students’ performance and interaction behaviors.
The results revealed that students without the external motivational mechanism demonstrated significantly better learning outcomes. The intelligent tutor effectively fostered deep learning behaviors such as student-initiated questioning, critical thinking, and knowledge integration. Furthermore, the quality of the tutor’s responses and reasoning was influenced by prompt design; prompts that integrated contextual information and conceptual guidance enhanced the accuracy of instructional judgments and improved interaction quality, thereby supporting students’ knowledge construction and conceptual understanding. Analysis also indicated that students’ knowledge-processing behaviors and the tutor’s inferred understanding were partially correlated with their learning outcomes. Finally, this study proposes technical improvements to address current system limitations and offers recommendations for future applications of intelligent tutors in science education.

Batir, S. B., & Akçay, H. (2022). Effects of multimodal representations on students' science learning. International Society for Technology, Education, and Science.
Chang, C. J., Liu, C. C., Wu, Y. T., Chang, M. H., Chiang, S. F., Chiu, B. C., Wen, C. T., Hwang, F. K., Chao, P. Y., Lai, C. H., Wu, S. W., & Chang, C. K. (2016). Students' perceptions on problem solving with collaborative computer simulation. In 24th International Conference on Computers in Education, ICCE 2016 (pp. 166-168). Asia-Pacific Society for Computers in Education.
Chi, M. T., & Wylie, R. (2014). The ICAP framework: Linking cognitive engagement to active learning outcomes. Educational psychologist, 49(4), 219-243.
Cohn, C., Hutchins, N., Le, T., & Biswas, G. (2024, March). A chain-of-thought prompting approach with llms for evaluating students’ formative assessment responses in science. In Proceedings of the AAAI Conference on Artificial Intelligence (Vol. 38, No. 21, pp. 23182-23190).
Collins, A., Brown, J. S., & Newman, S. E. (1987). Cognitive apprenticeship: Teaching the craft of reading, writing, and mathtematics. Center for the Study of Reading Technical Report; no. 403.
Cress, U., & Kimmerle, J. (2023). Co-constructing knowledge with generative AI tools: Reflections from a CSCL perspective. International Journal of Computer-Supported Collaborative Learning, 18(4), 607-614.
Dai, C. P., & Ke, F. (2022). Educational applications of artificial intelligence in simulation-based learning: A systematic mapping review. Computers and Education: Artificial Intelligence, 3, 100087.

Dai, C. P., Ke, F., Pan, Y., Moon, J., & Liu, Z. (2024). Effects of artificial intelligence-powered virtual agents on learning outcomes in computer-based simulations: A meta-analysis. Educational Psychology Review, 36(1), 31.
De Jong, T., Linn, M. C., & Zacharia, Z. C. (2013). Physical and virtual laboratories in science and engineering education. Science, 340(6130), 305-308.
Feng, G., Zhang, B., Gu, Y., Ye, H., He, D., & Wang, L. (2023). Towards revealing the mystery behind chain of thought: A theoretical perspective. Advances in Neural Information Processing Systems, 36, 70757-70798.
Gerard, L. F., Ryoo, K., McElhaney, K. W., Liu, O. L., Rafferty, A. N., & Linn, M. C. (2016). Automated guidance for student inquiry. Journal of Educational Psychology, 108(1), 60.
Graesser, A. C., Chipman, P., Haynes, B. C., & Olney, A. (2005). AutoTutor: An intelligent tutoring system with mixed-initiative dialogue. IEEE Transactions on Education, 48(4), 612-618.
Hattori, H., Kato, A., & Yoshizoe, M. (2024, December). Integrating large language models into agent models for multi-agent simulations: Preliminary report. In 2024 Winter Simulation Conference (WSC) (pp. 230-241). IEEE.
Jacobson, M. J., Taylor, C. E., & Richards, D. (2016). Computational scientific inquiry with virtual worlds and agent-based models: New ways of doing science to learn science. Interactive Learning Environments, 24(8), 2080-2108.
Kruglanski, A. W., Friedman, I., & Zeevi, G. (1971). The effects of extrinsic incentive on some qualitative aspects of task performance 1. Journal of personality, 39(4), 606-617.

Lee, A. V. Y., Teo, C. L., & Tan, S. C. (2024). Prompt engineering for knowledge creation: Using chain-of-thought to support students’ improvable ideas. AI, 5(3), 1446-1461.
Lee, G. G., Latif, E., Wu, X., Liu, N., & Zhai, X. (2024). Applying large language models and chain-of-thought for automatic scoring. Computers and Education: Artificial Intelligence, 6, 100213.
Lee, J., Chen, S. S., & Liang, P. P. (2025, April). Interactive sketchpad: A multimodal tutoring system for collaborative, visual problem-solving. In Proceedings of the Extended Abstracts of the CHI Conference on Human Factors in Computing Systems (pp. 1-14).
Merrill, W., & Sabharwal, A. (2023). The expressive power of transformers with chain of thought. arXiv preprint arXiv:2310.07923.
Munawar, S., Khalil Toor, S., Aslam, M., & Aimeur, E. (2019). Paca-its: A multi-agent system for intelligent virtual laboratory courses. Applied Sciences, 9(23), 5084.
National Research Council. (2000). Inquiry and the National Science Education Standards: A guide for teaching and learning. National Academies Press.
Ortiz, A. M. L., & Aliazas, J. V. C. (2021). Multimodal representations strategies in teaching science towards enhancing scientific inquiry skills among grade 4. IOER International Multidisciplinary Research Journal, 3(3), 107-118.
Qin, F., Hao, Z., Yu, J., Liu, Z., & Zhang, Y. (2025). AI instructional agent improves student's perceived learner control and learning outcome: Empirical evidence from a randomized controlled trial. arXiv preprint arXiv:2505.22526.
Reiff, R., Harwood, W. S., & Phillipson, T. (2002). A scientific method based upon research scientists' conceptions of scientific inquiry.

Ryan, R. M., & Deci, E. L. (2000). Intrinsic and extrinsic motivations: Classic definitions and new directions. Contemporary educational psychology, 25(1), 54-67.
Swaak, J., Van Joolingen, W. R., & De Jong, T. (1998). Supporting simulation-based learning; the effects of model progression and assignments on definitional and intuitive knowledge. Learning and instruction, 8(3), 235-252.
Trey, L., & Khan, S. (2008). How science students can learn about unobservable phenomena using computer-based analogies. Computers & Education, 51(2), 519-529.
Triona, L. M., & Klahr, D. (2003). Point and click or grab and heft: Comparing the influence of physical and virtual instructional materials on elementary school students' ability to design experiments. Cognition and Instruction, 21(2), 149-173.
Vanzo, A., Chowdhury, S. P., & Sachan, M. (2024). GPT-4 as a homework tutor can improve student engagement and learning outcomes. arXiv preprint arXiv:2409.15981.
Wang, L., Ma, C., Feng, X., Zhang, Z., Yang, H., Zhang, J., Chen, Z., Tang, J., Chen, X., Lin, Y., Zhao, W. X., Wei, Z., & Wen, J. (2024). A survey on large language model based autonomous agents. Frontiers of Computer Science, 18(6), 186345.
Wei, J., Wang, X., Schuurmans, D., Bosma, M., Ichter, B., Xia, F., Chi, E., Le, Q., & Zhou, D. (2022). Chain-of-thought prompting elicits reasoning in large language models. Advances in neural information processing systems, 35, 24824-24837.
Wiesner, T. F., & Lan, W. (2004). Comparison of student learning in physical and simulated unit operations experiments. Journal of Engineering Education, 93(3), 195-204.
Wooldridge, M., & Jennings, N. R. (1995). Intelligent agents: Theory and practice. The knowledge engineering review, 10(2), 115-152.
Wu, Y., Zhang, P., Xiong, W., Oguz, B., Gee, J. C., & Nie, Y. (2023). The role of chain-of-thought in complex vision-language reasoning task. arXiv preprint arXiv:2311.09193.
Xi, Z., Chen, W., Guo, X., He, W., Ding, Y., Hong, B., Zhang, M., Wang, J., Jin, S., Zhou, E., Zheng, R., Fan, X., Wang, X., Xiong, L., Zhou, Y., Wang, W., Jiang, C., Zou, Y., Liu, X., … Gui, T. (2023). The rise and potential of large language model based agents: A survey. arXiv preprint arXiv:2309.07864.
Zacharia, Z. C., & Constantinou, C. P. (2008). Comparing the influence of physical and virtual manipulatives in the context of the Physics by Inquiry curriculum: The case of undergraduate students’ conceptual understanding of heat and temperature. American Journal of Physics, 76(4), 425-430.
朱翊瑄（2024）。生成式人工智慧對於線上科學探究之影響。桃園市國立中央大 學資訊工程學系碩士論文。
許志仲（2024）。大型語言模型驅動之智慧學習助教對網路科學探究之成效評估。桃園市國立中央大學資訊工程學系碩士論文。
鄭宇辰（2025）。融合思考鏈機制的主動型生成式智慧教學助手之設計與其對科學學習成效之分析。桃園市國立中央大學資訊工程學系碩士論文。