認知型數位孿生導入教育決策：重構學校公共價值與責任

Integrating Cognitive Digital Twins Into Educational Decision-Making: Reframing Public Value and Accountability in Education

本研究回應當前教育治理面對複雜決策情境之挑戰，指出既有資料驅動決策模式多停留於績效監測與事後分析，缺乏能支援情境推演與前瞻判斷的認知機制。為此，本研究提出「認知型數位孿生」作為教育決策的新興概念，整合教育資料、專家知識與決策者認知，建構可進行情境模擬與多方案推演之虛擬決策環境。本研究進一步闡述認知型數位孿生運作機制如何透過資料整合、決策邏輯建模、模擬預測與回饋優化形成動態循環。相較於現有大型語言模型之黑箱與偏誤問題，認知型數位孿生強調知識結構與決策邏輯的整合，以提升決策之可解釋性與情境適切性。最後，本研究探討認知型數位孿生於教育治理與政策模擬之應用潛力，得以促進教育決策之智慧化與公共導向發展。

This study addresses the growing complexity of educational governance and argues that existing data-driven decision-making approaches remain largely limited to performance monitoring and retrospective analysis, lacking cognitive mechanisms to support scenario-based reasoning and forward-looking judgment. To address this gap, the study introduces the concept of the Cognitive Digital Twin as an emerging framework for educational decision-making. This framework integrates educational data, expert knowledge, and decision-maker cognition to construct a virtual environment capable of scenario simulation and multi-option analysis. Our study further explains how its operational mechanism forms a dynamic cycle through data integration, decision logic modeling, predictive simulation, and feedback-based refinement. In contrast to current large language models, which often suffer from opacity and bias, the Cognitive Digital Twin emphasizes the structured integration of knowledge and decision logic to enhance interpretability and contextual relevance. Finally, we discuss its potential applications in educational governance and policy simulation, highlighting its implications for more intelligent and publicly oriented decision-making in education.