数字神经网络的优势

The Superiority of Digital Neural Networks

摘要: 本文首先讨论了以M-P 神经元为标准的神经网络的特点及局限性。其次介绍了一种新的可用数字设备实现的概率逻辑神经元网络,比较了两者的特性。并指出前者在训练方法上和在硬件实施上的难点,强调了后者的易实现和训练方法的简单性。最后对后者网络状态的演变可用随机自动机理论进行分析的特点,显示出后者对于通用连接机分析的更一般理论。

Abstract: In this paper,first,the characteristics and limitations of neural networks with standardM-P neuron model are discussed.Secondly,a novel probabilistic logic neural network (PLNN) imple-mented by digital device such as RAM is introduced.Finally,the simplicity of learning algorithm andhardware-easily-implemented property of PLNN are emphasized,which shows that PLNN possesses agreat potential in parallel processing and machine inference