| Presentation | 1999/3/19 An Approach to Reliable Pattern Recognition with Randomized ANNs and its Alications Tohru NAKAGAWA, Keiichi HORIKAWA, Kazuya OHISHI, Hiroyuki KOTERA, Hajime KITAGAWA, |
|---|---|
| PDF Download Page | PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | This paper presents an architecture of artificial neural networks, abbreviated ANN, for implementing highly reliable pattern recognition. The architecture includes at least one randomized multiple ANN (abbreviated RANN) and, if necessary, more RANNs to be connected with each other in a pipelined manner. In a learning phase, each ANN which is an element of a RANN should be trained under the same learning condition except for the randomized initial state of each ANN. In a recall phase, the outputs of each ANN should be unified with the others using a simple operation such as logical AND. In the case of coin identification for a Japanese vending machine, our RANN has a 0% misclassification rate. Moreover, in the case of a handwritten character recognition, the proposed pipelined architecture can reduce the misclassification ratio from 9% to 3%. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Pattern Recognition / Generalization / Reliability / Randomized ANN / Coin Identification. |
| Paper # | NC98-155 |
| Date of Issue |
| Conference Information | |
| Committee | NC |
|---|---|
| Conference Date | 1999/3/19(1days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | |
| Chair | |
| Vice Chair | |
| Secretary | |
| Assistant | |
| Paper Information | |
| Registration To | Neurocomputing (NC) |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | An Approach to Reliable Pattern Recognition with Randomized ANNs and its Alications |
| Sub Title (in English) | |
| Keyword(1) | Pattern Recognition |
| Keyword(2) | Generalization |
| Keyword(3) | Reliability |
| Keyword(4) | Randomized ANN |
| Keyword(5) | Coin Identification. |
| 1st Author's Name | Tohru NAKAGAWA |
| 1st Author's Affiliation | Department of Information & Control Engineering Toyota Technological Institute() |
| 2nd Author's Name | Keiichi HORIKAWA |
| 2nd Author's Affiliation | Daicel Chemical Industries, LTD. |
| 3rd Author's Name | Kazuya OHISHI |
| 3rd Author's Affiliation | Department of Information & Control Engineering Toyota Technological Institute |
| 4th Author's Name | Hiroyuki KOTERA |
| 4th Author's Affiliation | Department of Information & Control Engineering Toyota Technological Institute |
| 5th Author's Name | Hajime KITAGAWA |
| 5th Author's Affiliation | Department of Information & Control Engineering Toyota Technological Institute |
| Date | 1999/3/19 |
| Paper # | NC98-155 |
| Volume (vol) | vol.98 |
| Number (no) | 674 |
| Page | pp.pp.- |
| #Pages | 8 |
| Date of Issue | |