| Project/Area Number |
20H01722
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (B)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Review Section |
Basic Section 09070:Educational technology-related
|
|---|
| Research Institution | Kyoto University |
|---|
Principal Investigator |
Flanagan Brendan 京都大学, 学術情報メディアセンター, 特定講師 (00807612)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
緒方 広明 京都大学, 学術情報メディアセンター, 教授 (30274260)
Majumdar Rwito 京都大学, 学術情報メディアセンター, 特定講師 (30823348)
|
|---|
| Project Period (FY) |
2020-04-01 – 2023-03-31
|
| Project Status |
Granted (Fiscal Year 2022)
|
| Budget Amount *help |
¥17,940,000 (Direct Cost: ¥13,800,000、Indirect Cost: ¥4,140,000)
Fiscal Year 2022: ¥4,290,000 (Direct Cost: ¥3,300,000、Indirect Cost: ¥990,000)
Fiscal Year 2021: ¥5,070,000 (Direct Cost: ¥3,900,000、Indirect Cost: ¥1,170,000)
Fiscal Year 2020: ¥8,580,000 (Direct Cost: ¥6,600,000、Indirect Cost: ¥1,980,000)
|
|---|
| Keywords | Knowledge Map / Knowledge Extraction / Learning Analytics / Educational Data Mining / Smart Learning Systems / Knowledge map / Smart learning systems / Knowledge extraction / Learning analytics / Educational Informatics |
|---|
| Outline of Research at the Start |
The study guidance code proposed by MEXT will standardize the representation of learning material knowledge structures, however a system to integrate teacher created materials with publisher contents and learning analytics systems is required to realize the full potential to support smart education and learning. This research investigates how meaningful analysis can be achieved by supporting the automated extraction, linking, management, and analysis of knowledge maps at scale.
|
| Outline of Annual Research Achievements |
Inital development in the previous year lead to the implementation and evaluation of several research sub-topics, and results were also disseminated in journals and international conferences, inlcuding a new knowledge tracing model based on the latest transformer deep learning model. The later is based on a BERT style transformer and outperformed the state-of-the-art deep knowledge tracing models at the time of presentation. Other research was also conducted to examine the explainability of more classic knowledge tracing models, such as: BKT by analyzing the internal parameters of the model and how they relate with the types of quizzes being recommended. An explainable group formation method was also proposed by applying a genetic algorithm to the creation of groups for study tasks based on the students current knowledge state as estimated by the knowledge map platform. A reading recommendation system was also designed based on the knowledge map platform preliminary evaluation was conducted in a school. The design of the system was presented as a poster paper at the leading conference on learning analytics, LAK.
|
| Current Status of Research Progress |
Current Status of Research Progress
3: Progress in research has been slightly delayed.
Reason
While research results were achieved and presented at conferences and in journals, the development and evaluation of the system was effected by ongoing Covid-19 restrictions. This caused unavoidable delays in the fundamental development and evaluation, which impacted the schedule for implementation and dissemination of research results.
|
| Strategy for Future Research Activity |
It is anticipated that the project will continue to show results in the sub-themes of areas, such as: recommendation, prediction, and increased usability of knowledge map systems for teachers and students alike. In particular, supporting teachers in the use and creation of knowledge maps using the knowledge map portal that is being developed in this project will continue to be a main focus and is expected to show promising results over the next year as Covid-19 restrictions are lifted and development increases. The use of knowledge maps in recommendation will also be a continued focus, with the concept design being expanded from vocabulary based knowledge maps for English reading, to grammar item based knowledge maps for both reading and quiz recommendation based on a students current knowledge state.
|
