2022 Fiscal Year Final Research Report
Minimax Optimal Functional Estimation on Large-Scale Discrete Distributions
| Project/Area Number |
20K19750
|
|---|
| Research Category |
Grant-in-Aid for Early-Career Scientists
|
| Allocation Type | Multi-year Fund |
|---|
| Review Section |
Basic Section 60030:Statistical science-related
|
|---|
| Research Institution | University of Tsukuba |
|---|
Principal Investigator |
Fukuchi Kazuto 筑波大学, システム情報系, 助教 (30838090)
|
|---|
| Project Period (FY) |
2020-04-01 – 2023-03-31
|
| Keywords | minimax optimality / functional estimation |
|---|
| Outline of Final Research Achievements |
We are committed to advancing the techniques for scrutinizing the minimax optimality of estimation problems. Unveiling the minimax optimality of such problems holds substantial value, as it elucidates the most efficient method for tackling estimation problems. As a consequence of this development, we provide insightful characterizations of minimax optimality in the context of privacy-constrained and fairness-constrained estimation problems.
|
| Free Research Field |
machine learning
|
| Academic Significance and Societal Importance of the Research Achievements |
minimax最適性の理解は推定問題の本質的な難しさを示唆してくれるため，推定がうまくいかない状況を避けたり，実験計画を立てたりする状況で活用できる．本研究では，汎関数推定問題のminimax最適性を解析する技術を開発する中で得られた結果を使って，特に最近社会的要請の強いプライバシー，公平性制約が課された推定問題におけるminimax最適性の特性を明らかにした．
|
