| Two Criteria for Learning in Feedforward Neural Networks |
| |
| 作者姓名： | 彭汉川 甘强 韦钰 |
| |
| 作者单位： | Peng Hanchuan Gan Qiang Wei Yu 1，2 (Department of Biomedical Engineering，Southeast University，Nanjing 210018) 1(National Laboratory of Pattern Recognition，Beijing 100080) |
| |
| 摘 要： | ＴｗｏＣｒｉｔｅｒｉａｆｏｒＬｅａｒｎｉｎｇｉｎＦｅｅｄｆｏｒｗａｒｄＮｅｕｒａｌＮｅｔｗｏｒｋｓＰｅｎｇＨａｎｃｈｕａｎ（彭汉川）１，２ＧａｎＱｉａｎｇ（甘强）１ＷｅｉＹｕ（韦钰）１１，２（ＤｅｐａｒｔｍｅｎｔｏｆＢｉｏｍｅｄｉｃａｌＥｎｇｉｎｅ...
|
Two Criteria for Learning in Feedforward Neural Networks |
| |
| Peng Hanchuan,Gan Qiang,Wei Yu.Two Criteria for Learning in Feedforward Neural Networks[J].Journal of Southeast University(English Edition),1997(2). |
| |
| Authors: | Peng Hanchuan Gan Qiang Wei Yu |
| |
| Institution: | Peng Hanchuan Gan Qiang Wei Yu 1,2 |
| |
| Abstract: | Two criteria for improving the generalization of learning in feedforward neural networks are proposed. Based on the criteria, a multi channel weight sharing neural network is designed and applied to unconstrained handwritten character recognition. Its recognition performance is much better than that of the standard BP network. |
| |
| Keywords: | feedforward neural networks generalization |
| 本文献已被 CNKI 等数据库收录！ |
