| Project/Area Number |
09650416
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
情報通信工学
|
|---|
| Research Institution | HIROSHIMA UNIVERSITY |
|---|
Principal Investigator |
AE Tadashi Hiroshima University, Faculty of Eng., Professor, 工学部, 教授 (50005386)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
SAKAI Keiichi Hiroshima University, Faculty of Eng., Reseach Assoc., 工学部, 助手 (90274117)
|
|---|
| Project Period (FY) |
1997 – 1998
|
| Project Status |
Completed (Fiscal Year 1998)
|
| Budget Amount *help |
¥3,300,000 (Direct Cost: ¥3,300,000)
Fiscal Year 1998: ¥1,500,000 (Direct Cost: ¥1,500,000)
Fiscal Year 1997: ¥1,800,000 (Direct Cost: ¥1,800,000)
|
|---|
| Keywords | Brainware / Brain Functional Processing / Optical Interconnection / Inductive Learning |
|---|
| Research Abstract |
A brain-computing scheme using optical interconnection is discussed, where the optical interconnection plays an important role for high-speed matching procedure. The scheme has two levels ; Level 1 (Component Level) and Level 2 (Structure Level). At Level : 1 an. elementary pattern recognition is performed as. in the conventional pattern recognition, while the syntax pattern recognition is done at the structure level.
Level 2 is applied to both space and time. The structure of space is essentially the same as the syntax of a configuration (figure, sentence, etc.), while the structure of time is applied to a time-variant scheme of both component level and structure level The former is similar to the conventional sequential system, but the vector consisting of real values is used instead of discrete values. The latter shows a new scheme, ie, a time-variant structure. We can discuss an anticipatory system using this time-variant structure. For anticipatory system we have proposed a new learning algorithm AST and verified its effectiveness by simulation. The learning algorithm should be a constructive inductive learning, and the structure must be obtained by learning procedure. In other words, the structure of automatic truth maintenance or the structure of minimal systematic representation must be constructed by learning. The rule-based representation is used for the structure construction of automatic truth maintenance. The machine representation is also used for the structure construction of minimal systematic representation. The representation depends on the application, but it must be extended to numerical case front symbolic case.
As a result, we could develop a new type of brainware system using optical interconnection.
|
