Development of AI methodology with super-computed data for estimating earthquake disaster
Project/Area Number |
18K18873
|
Research Category |
Grant-in-Aid for Challenging Research (Exploratory)
|
Allocation Type | Multi-year Fund |
Review Section |
Medium-sized Section 22:Civil engineering and related fields
|
Research Institution | The University of Tokyo |
Principal Investigator |
|
Project Period (FY) |
2018-06-29 – 2022-03-31
|
Project Status |
Completed (Fiscal Year 2021)
|
Budget Amount *help |
¥5,980,000 (Direct Cost: ¥4,600,000、Indirect Cost: ¥1,380,000)
Fiscal Year 2020: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000)
Fiscal Year 2019: ¥3,770,000 (Direct Cost: ¥2,900,000、Indirect Cost: ¥870,000)
Fiscal Year 2018: ¥910,000 (Direct Cost: ¥700,000、Indirect Cost: ¥210,000)
|
Keywords | 人工知能 / 地震被害推定 / 大規模解析 |
Outline of Final Research Achievements |
In order to mitigate earthquake damage, many attempts have been made to improve the reliability of earthquake damage estimation. In this research, we developed an efficient artificial intelligence method for earthquake damage estimation by learning from large-scale numerical simulation results (supercomputed data) and physical considerations, and demonstrated its effectiveness through application examples. The reduction of the analysis cost is expected to contribute to fast and immediate earthquake damage estimation and to the plan for earthquake disaster prevention that takes into account many scenarios. We also show that artificial intelligence based on physical considerations can reduce the cost of analysis by shifting the analysis from conventional equation-based modeling to operations that are more suitable for recent computer architectures. The results show the potential of combining equation-based modeling and data-science methods.
|
Academic Significance and Societal Importance of the Research Achievements |
このような解析コスト削減の実現は,高速な即時地震被害想定や,多数シナリオを考慮した事前震災対策用想定への貢献が期待される.また,物理的考察を踏まえた人工知能化により,従来のequation-based modelingによる解析を,昨今の高演算密度に適した計算機アーキテクチャに親和性の高い演算へシフトすることで解析コストの低減を実現できることを示す等,equation-based modelingとdata-science的な手法の融合の可能性を示す結果となっている.
|
Report
(5 results)
Research Products
(29 results)
-
-
-
-
-
-
-
-
-
-
[Journal Article] 416-PFLOPS Fast Scalable Implicit Solver on Low-Ordered Unstructured Finite Elements Accelerated by 1.10-ExaFLOPS Kernel with Reformulated AI-Like Algorithm: For Equation-Based Earthquake Modeling2019
Author(s)
Tsuyoshi Ichimura*, Kohei Fujita, Takuma Yamaguchi, Akira Naruse, Jack C. Wells, Christopher J. Zimmer, Tjerk P. Straatsma, Takane Hori, Simone Puel, Thorsten W. Becker, Muneo Hori, Naonori Ueda
-
Journal Title
SC19: The International Conference for High Performance Computing, Networking, Storage, and Analysis
Volume: -
Related Report
Peer Reviewed / Int'l Joint Research
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-