论文部分内容阅读
本文提出了符合多态Ｉｓｉｎｇ模型神经网络动力学规则的生成泛函，并从该泛函导出了零温度时严格的三态、四态Ｉｓｉｎｇ神经网络的并行动力学方程．所获得的无自耦网络的动力学方程与发表的结果符合得很好．此方法的优点是在导出多态Ｉｓｉｎｇ神经网络的严格动力学方程的同时能求出其它多个序参量．
In this paper, we present a generator functional that conforms to the dynamical rules of the neural network of the polymorphic Ising model and derive the parallel dynamics equations of the strict tri-state and quaternary Ising neural networks at zero temperature. The kinetic equations obtained for the autotuning network are in good agreement with the published results. The advantage of this method is that it can derive many other order parameters while deriving the strict dynamic equations of the multistep Ising neural network.