PRIVACY PRESERVING COLLABORATIVE FILTERING BASED ON INFORMATION GEOMETRY
| Project/Area Number |
22500142
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Intelligent informatics
|
|---|
| Research Institution | National Institute of Advanced Industrial Science and Technology |
|---|
Principal Investigator |
AKAHO Shotaro 独立行政法人産業技術総合研究所, ヒューマンライフテクノロジー研究部門, 研究グループ長 (40356340)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
KAMISHIMA Toshihiro 独立行政法人産業技術総合研究所, ヒューマンライフテクノロジー研究部門, 主任研究員 (50356820)
FUJIKI Jun 福岡大学, 理学部, 准教授 (10357907)
|
|---|
| Project Period (FY) |
2010 – 2012
|
| Project Status |
Completed (Fiscal Year 2012)
|
| Budget Amount *help |
¥4,290,000 (Direct Cost: ¥3,300,000、Indirect Cost: ¥990,000)
Fiscal Year 2012: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000)
Fiscal Year 2011: ¥1,430,000 (Direct Cost: ¥1,100,000、Indirect Cost: ¥330,000)
Fiscal Year 2010: ¥1,560,000 (Direct Cost: ¥1,200,000、Indirect Cost: ¥360,000)
|
|---|
| Keywords | 知識発見とデータマイニング / 情報幾何 / 推薦システム / 機械学習 / 個人情報保護 / 多変量解析 |
|---|
| Research Abstract |
In this project, the collaborative filtering preserving privacy was treated as a statistical inference in the information geometric framework. We studied the robust criterion for the privacy preservation and obtained the optimal condition. We further showed the relationship between beta divergence criterion of the information geometry and the H infinity filter in the field of robust control, which shows the possibility to apply the collaborative filtering to time series data. For the collaborative filtering, we developed the simultaneous optimization procedure of dimension reduction and clustering from Bayesian framework. Moreover, we found out a fairness-aware learning framework based on privacy preservation and investigated its fundamental property.
|
Report
(4 results)
Research Products
(22 results)
