论文部分内容阅读
多重训练编码和傀儡编码使存贮在Hopfield网格中的模式即使不是局部最小点也能被回忆出来,某一特定模式可以通过多重训练编码得以回忆,并讨论了最少训练次数,通过附加傀儡数据,傀儡编码可确保回忆Hopfield网络中所有的存贮模式.
Multiple training coding and puppet coding make the patterns stored in the Hopfield grid be recalled even if they are not local minima. A specific pattern can be recalled by multiple training codes and the minimum number of trainings is discussed. By adding the puppet data Puppet coding ensures that all storage modes in the Hopfield network are remembered.