光学学报, 1996, 16 (6): 772, 网络出版: 2006-12-04
多状态、多阈值神经网络模型的光电混合实现
Opto-Electronics Hybrid Implementation of a Multi-State and Multi-Threshold Neural Network Model
神经网络 光互连 多值 权重 阈值 neural network optical interconnections multivalue interconnection weight matrx threshold
摘要
提出了一种采用高分辨率液晶电视(LCTV)实现Hopfield神经网络多值算法的光电系统。文章给出了平面多状态、多阈值的全互连Hopfield[1]神经网络模型,并采用该系统对颜色进行了联想和记忆的实验。初步的实验结果可以证实:此种高分辨率液晶电视神经网络系统是可行的。
Abstract
A new kind of optical neural network for associative colors using high resolution liquid crystal televisions (LCTV) is proposed and constructed. This new model adopts the multivalued neural algorithm based Hop field feedback associative memory model. The system design consideration and an experimental demonstration of the LCTV neural network are provided.
黄达诠, 黄海云. 多状态、多阈值神经网络模型的光电混合实现[J]. 光学学报, 1996, 16(6): 772. 黄达诠, 黄海云. Opto-Electronics Hybrid Implementation of a Multi-State and Multi-Threshold Neural Network Model[J]. Acta Optica Sinica, 1996, 16(6): 772.