| Project/Area Number |
23K16939
|
|---|
| Research Category |
Grant-in-Aid for Early-Career Scientists
|
| Allocation Type | Multi-year Fund |
|---|
| Review Section |
Basic Section 61030:Intelligent informatics-related
|
|---|
| Research Institution | University of Tsukuba |
|---|
Principal Investigator |
NGUYEN DAIHAI 筑波大学, システム情報系, 助教 (50968401)
|
|---|
| Project Period (FY) |
2023-04-01 – 2026-03-31
|
| Project Status |
Granted (Fiscal Year 2023)
|
| Budget Amount *help |
¥2,730,000 (Direct Cost: ¥2,100,000、Indirect Cost: ¥630,000)
Fiscal Year 2025: ¥520,000 (Direct Cost: ¥400,000、Indirect Cost: ¥120,000)
Fiscal Year 2024: ¥910,000 (Direct Cost: ¥700,000、Indirect Cost: ¥210,000)
Fiscal Year 2023: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000)
|
|---|
| Keywords | Generative models / Constrained domains / optimal transport / graph kernels / graph structured data |
|---|
| Outline of Research at the Start |
We design novel measures for graphs with the following advantages:
1) taking into account node features and global structures of graphs.
2) deriving valid kernels that can be used for kernel-based frameworks.
3) reducing the complexity of computing the pairwise distance or kernel matrices.
|
| Outline of Annual Research Achievements |
The research aims to develop optimal transport-based learning models for structured data, along with their related extensions and applications.
This year, our main focus has been on learning to generate structured data. Generating structured data, such as graphs, poses a significant challenge due to the constraints imposed on the generated samples. To address it, we formulated the problem of generating structured data as a distribution optimization problem on constrained domains. Then we developed a general framework based on the optimal transport theory to tackle this issue. We conducted an analysis of theoretical properties and convergence of the proposed, as well as experiments on synthetic and real world data sets to demonstrate its effectiveness in generating constrained samples.
|
| Current Status of Research Progress |
Current Status of Research Progress
1: Research has progressed more than it was originally planned.
Reason
We have achieved promising results in the topics of learning to generate structured data, with our research findings being published in top conferences and journals. Moving forward, we plan to further extend our research focus to include graph data and its applications.
|
| Strategy for Future Research Activity |
This year, our plan is to continue our efforts in developing optimal transport-based learning models tailored for graph-structured data, with a specific focus on their applications in Bioinformatics. Our research will encompass various domains within Bioinformatics, including molecular graphs, biological network data, and more.
|
