联想学习记忆的神经网络模型

联想学习记忆的神经网络模型 AN ASSOCIATION-LEARNING-MEMORY NEURAL NETWORK MODEL 梁明理 Liang Mingli Liu Jun 1 first-author 刘军 1 武汉大学物理系 武汉大学物理系 Department of Physics,Wuhan University Department of Physics,Wuhan University 本文提出了联想学习记忆的神经网络模型,讨论了该模型的主要特性。对400个神经元、30个随机图样组成的网络系统给出并分析了计算机模拟计算结果。讨论了在不同的学习因子λ和联想因子 g 的情况下网络的演化行为,研究了网络在学习新态时联想演化对学习效果的影响以及初态噪声 P 4 和联想噪声 P a 对已存记忆态恢复行为的影响。由此得出了一些结论。 This paper presents a new neural network model and discusses main pro- perties of association and learning of the model.Calculations of computer simulation to a system with 400 neurons and 30 random patterns are given and analysed.The evolution behaviors of the network under different learning factor λ and association factor g are discussed.The effects of association on learning while the network learning a new pattern,and the effects of initial noise P i and association noise P a on the retrival of a learned pattern are also studied.Some conclusions are obtained. 神经元 网络 图样 联想演化 突触联系强度 学习 记忆 neuron network pattern association evolution synaptic strength learning memory 1990-01-01 2021-04-01 1