摘要
建立了神经网络的非平衡态热力学理论,给出了神经网络熵产生率,稳定条件及应用实例.为神经网络的研究增加了一种新方法,为人工神经网络的建模与应用提供了有益的信息.
The non-equilibrium thermodynamic theory of artificial neural networks is introduced in this paper. The concepts of entropy production rate, network stability condition and its applications are given. All of them are the foundations of non-equilibrium thermodynamic theory of artificial neural networks, which will be a new approach in the study of artificial neural networks. The results are beneficial to the modeling and applications of artificial networks.
出处
《甘肃科学学报》
2005年第1期77-80,共4页
Journal of Gansu Sciences
基金
甘肃省自然科学基金(3ZS041-A25-024)
关键词
神经网络
非平衡态
热力学理论
Neural networks
non-equilibrium
thermodynamic theory
