| 在片学习及权值刷新神经网络硬件实现方法 |
| |
| 引用本文： | 庞维珍,李福成.在片学习及权值刷新神经网络硬件实现方法[J].天津大学学报(自然科学与工程技术版),1996,29(6):821-827. |
| |
| 作者姓名： | 庞维珍 李福成 |
| |
| 作者单位： | [1]电子工程系 [2]天津大学精密仪器与光电子工程学院 |
| |
| 摘 要： | 提出几种电流模式电路块，并用这些模拟电路块构成一个电流模型式ＢＰ人工神网络，该ＢＰ神经网络具有在片学习和在片权值存储能力。对所设计的模拟电路块进行了理论对构成的ＢＰ神经网络作了应用实例研究，并提供了ＰＳＰＩＣＥ仿真结果。
|
| 关 键 词： | 神经网络 电流模式电路 BP算法 模拟电路 |
HARDWARE IMPLEMENTATION OF NEURAL NETWORK WITH ON-CHIP LEANING AND WEIGHT REFRESHING |
| |
| Pang Weizhen, Ren Luyong,Wu Qinghua.HARDWARE IMPLEMENTATION OF NEURAL NETWORK WITH ON-CHIP LEANING AND WEIGHT REFRESHING[J].Journal of Tianjin University(Science and Technology),1996,29(6):821-827. |
| |
| Authors: | Pang Weizhen Ren Luyong Wu Qinghua |
| |
| Abstract: | n this paper,some Current-mode analog circuit blocks are presented and a Backward-Propagation(BP)artificial neural network is constructed from these blocks.The BP neural network has the ability of on-chip learning and on-chip weights storage.the theoretical analysis of the circuit blocks and al)plication examples of the BP neural network are given,and the PSPICE simulation results are also reported. |
| |
| Keywords: | artificial neural network current-mode circuit BP algorithm |
| 本文献已被 CNKI 维普 等数据库收录！ |
