Hopfield 网——擅长于联想记忆与解迷路 实现H网联想记忆的关键，是使被记忆的模式样本对应网络能量函数的极小值。 设有M个N维记忆模式，通过对网络N个神经元之间连接权 wij 与N个输出阈值θjjj的设计，使的: 这M个记忆模式所对应的网络状态正好是网络能量函数的M个极小值。 比较困难，目前还没有一个适应任意形式的记忆模式的有效、通用的设计方法。 H网的算法 1）学习模式——决定权重 想要记忆

-Hopfield network- specializes in the associative memory lost Xie H net associative memory key is to make a very small value of the memory mode, the sample corresponds to the network energy function. M N-dimensional memory model, the right wij and N output threshold θjjj connection between the network of N neurons, so that: the M memory mode corresponding to the network status is just the network energy function M a very small value. More difficult, there is no one to adapt to any form of memory mode is active, universal design. H network algorithm 1) learning mode- determines the weight you want memory