| Project/Area Number |
17K12743
|
|---|
| Research Category |
Grant-in-Aid for Young Scientists (B)
|
| Allocation Type | Multi-year Fund |
|---|
| Research Field |
Intelligent informatics
|
|---|
| Research Institution | National Institute of Informatics |
|---|
Principal Investigator |
Konishi Takuya 国立情報学研究所, ビッグデータ数理国際研究センター, 特任研究員 (20760169)
|
|---|
| Project Period (FY) |
2017-04-01 – 2020-03-31
|
| Project Status |
Completed (Fiscal Year 2019)
|
| Budget Amount *help |
¥2,470,000 (Direct Cost: ¥1,900,000、Indirect Cost: ¥570,000)
Fiscal Year 2018: ¥1,170,000 (Direct Cost: ¥900,000、Indirect Cost: ¥270,000)
Fiscal Year 2017: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000)
|
|---|
| Keywords | ベイズ学習 / 変分ベイズ法 / 機械学習 |
|---|
| Outline of Final Research Achievements |
Latent variable models are probabilistic models to represent the hidden features and relations behind data. Variational Bayesian inference provides a learning algorithm for latent variable models and is recognized as an active research topic in machine learning and statistics. We focus on the theoretical aspects of variational Bayesian inference and analyze the asymptotic behavior of learning algorithms of latent variable models that have not considered in the literature.
|
| Academic Significance and Societal Importance of the Research Achievements |
ベイズ推定は機械学習や統計学を中心に自然科学の様々な分野で応用されている．その一方で，潜在変数モデルのような複雑な確率モデルに対しては計算量的に厳密な推定が困難なことが知られている．変分ベイズ法はこの問題を解決する有力な近似手法の一つであり，本研究で行った変分ベイズ法の理論的な解析によって学習アルゴリズムの基礎的な理解が深まるとともに，応用範囲が広がることが期待できる．
|
