Information Enhancement for Interpreting Internal Representation in Neural Networks
| Project/Area Number |
21500221
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Sensitivity informatics/Soft computing
|
|---|
| Research Institution | Tokai University |
|---|
Principal Investigator |
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
TAKEUCHI Haruhiko 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 (00357401)
|
|---|
| Project Period (FY) |
2009 – 2011
|
| Project Status |
Completed (Fiscal Year 2011)
|
| Budget Amount *help |
¥4,420,000 (Direct Cost: ¥3,400,000、Indirect Cost: ¥1,020,000)
Fiscal Year 2011: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000)
Fiscal Year 2010: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000)
Fiscal Year 2009: ¥1,820,000 (Direct Cost: ¥1,400,000、Indirect Cost: ¥420,000)
|
|---|
| Keywords | ニューラルネットワーク / 自己組織化マップ / 特徴抽出 / 情報理論 / エントロピー / 相互情報量 / エンハンスメント / 内部表現 |
|---|
| Research Abstract |
In this study, we proposed a new method called "information enhancement" for interpreting the mechanism of neural networks. We tried to interpret all the components and all the possible combinations of the components by this information enhancement. Applied to the self-organizing maps, we found that clear class structure could be produced for interpretation. In addition, we found that the method was related to a new model for neural networks taking into account the social activities of neurons.
|
Report
(4 results)
Research Products
(55 results)
