研究課題/領域番号 |
22KJ1914
|
補助金の研究課題番号 |
22J15869 (2022)
|
研究種目 |
特別研究員奨励費
|
配分区分 | 基金 (2023) 補助金 (2022) |
応募区分 | 国内 |
審査区分 |
小区分62030:学習支援システム関連
|
研究機関 | 京都大学 |
研究代表者 |
OCHEJA Patrick 京都大学, 学術情報メディアセンター, 特別研究員(PD)
|
研究期間 (年度) |
2023-03-08 – 2025-03-31
|
研究課題ステータス |
交付 (2023年度)
|
配分額 *注記 |
2,200千円 (直接経費: 2,200千円)
2023年度: 1,300千円 (直接経費: 1,000千円、間接経費: 300千円)
2022年度: 1,200千円 (直接経費: 1,200千円)
|
キーワード | blockchain / lifelong learning / user model / learning log / BOLL / personalized feedback / learning analytics / generative AI / learning analytic / assessment |
研究開始時の研究の概要 |
This research will focus on advancing the emerging field of blockchain in education by developing a novel blockchain system (blockchain of learning logs - Boll) for connecting and transferring academic records across institutions. We will use learning analytics to validate relevance of connected distributed learner data by measuring the effect of learning behaviour across different schools on learning outcome, and building a connected lifelong learner knowledge model using decentralized data.
|
研究実績の概要 |
Our research on integrating distributed user models via blockchain included a usability study of the Blockchain of Learning Logs (BOLL) system. This evaluation engaged students and teachers to assess the system's user-friendliness, control features, interface consistency, and practical relevance. Based on the findings, we improved the system's design and developed a new user interface to facilitate easier deployment in educational settings. The results were presented at the 31st International Conference on Computers in Education in Matsue, Japan.
Further, in collaboration with the Migalabs team at the Barcelona SuperComputing Center, we developed custom tools for collecting and analyzing data on the Ethereum blockchain. Our analysis focused on large blocks to explore potential scalability challenges, particularly with upcoming technology like rollups and blob transactions. These findings will be shared at the 6th Blockchain and Internet of Things Conference in Fukuoka, Japan.
We also initiated research on using blockchain-connected data and generative artificial intelligence (AI) to provide personalized learning guidance. Preliminary assessments indicate that this approach could automate the process of delivering personalized feedback, significantly reducing teacher workload. Our initial results suggest that the integration of rich contextual data from various educational environments can enhance the efficacy of large language models in personalizing responses, thereby improving the learning experience.
|
現在までの達成度 (区分) |
現在までの達成度 (区分)
2: おおむね順調に進展している
理由
Our research has made significant progress, aligning well with our outlined objectives. This past year, we focused on developing systems for studies in diverse learning environments and engaged with external blockchain experts to enhance its educational applications. This collaboration has advanced satisfactorily and promises to significantly contribute to our research goals in future phases.
Additionally, we have started integrating large language models (LLMs) to personalize learning experiences. This integration uses lifelong learning logs stored on the blockchain, providing a rich source of contextual information for learners. This development is a major step towards more effectively tailoring educational experiences through generative AI.
|
今後の研究の推進方策 |
We plan to continue our research by evaluating the implementation of our blockchain system in various educational settings. We will focus on deploying this system in schools to enhance teaching and learning activities. Additionally, we will explore how data from the blockchain can be used to build user models and provide contextual data to LLMs for personalized feedback.
In summary, our research will focus on: 1. Deploy our educational blockchain system in select schools to connect lifelong learning data and study its impact on educational processes. 2. Evaluate how blockchain-connected data can improve the use of AI systems in education. 3. Conduct a comprehensive analysis of the effects of using connected lifelong learning data and the resulting user models in various learning contexts.
|