| Project/Area Number |
16K16129
|
|---|
| Research Category |
Grant-in-Aid for Young Scientists (B)
|
| Allocation Type | Multi-year Fund |
|---|
| Research Field |
Soft computing
|
|---|
| Research Institution | Fukuoka University |
|---|
Principal Investigator |
|
|---|
| Project Period (FY) |
2016-04-01 – 2018-03-31
|
| Project Status |
Completed (Fiscal Year 2017)
|
| Budget Amount *help |
¥4,030,000 (Direct Cost: ¥3,100,000、Indirect Cost: ¥930,000)
Fiscal Year 2017: ¥780,000 (Direct Cost: ¥600,000、Indirect Cost: ¥180,000)
Fiscal Year 2016: ¥3,250,000 (Direct Cost: ¥2,500,000、Indirect Cost: ¥750,000)
|
|---|
| Keywords | リザーバコンピューティング / レーザ / 時間遅延フィードバック / 時系列予測 / 相互結合システム / 機械学習 / reservoir computing / 相互結合 / 非線形システム / 応用光学 / 時間遅延ダイナミカルシステム / 半導体レーザ / 光情報処理 / 非線形ダイナミクス / ソフトコンピューティング / 画像、文章、音声等認識 |
|---|
| Outline of Final Research Achievements |
A novel information processing method based on delayed dynamical systems (DDSs) has been proposed, which is called reservoir computing (RC). In delay-based RC, a DDS is regarded as a recurrent neural network composed of many coupled nodes and virtual
nodes are considered in the feedback delay loop by time-multiplexing. In this study, we numerically and experimentally demonstrate RC based on mutually-delay coupled optoelectronic system. We consider N virtual nodes in each two systems, which results in 2N virtual nodes. To produce diifferent node states in each systems, we use feedback delay loops with different delay times. We applied a chaotic time-series prediction task to evaluate the processing performance of our system. It is found that our RC system can produce higher prediction accuracy than RC based on
a single optoelectronic system.
|
