二元神经网络模型中的Monte Carlo学习算法

THE MONTE CARLO LEARNING ALGORITHM IN BINARY MODEL OF NEURAL NETWORK

基于作者以前提出的一种单通道双极光学神经网络结构,利用自适应阈值,本文提出了一个新的Monte Carlo学习算法的判据。这个判据可以使Monte Carlo学习算法进一步改善神经网络的性能。

Based on the architecture of single-channel optical system for bipolar neural network with self-adaptive thresholds which the authors proposed previously, a new criterion for Monte Carlo learning algorithm is presented. With the new criterion, the network can be of higher storage capacity and error tolerance.