Feedback Nerual Network(一）：Concept

It requires a certain period of working time, then can reach Stable Status, so it is a Dynamic Network、

 

 

 

Two types of Feedback NN according to the value of output is discrete or continuous
DHNN: Discrete Hopfield NN
CHNN: Continuous Hopfield NN

 

If system is stable, then it will converge from any initial state to a Stable State
If system is unstable, then the self-oscillation with limited amplitude or ultimate-loop 极限環 will be occurred, because of the output is 1 & -1 or 1 & 0 binary states.

 

Structure & Work Mode：

 

For each node of Discrete Hopfield neuron:it is similiar with feedforward network,unless a feedback input.

 

 

 

 

 

 Two Work Modes:
1Asynchronous Mode
Each time only one node of neurons adjusts his status, all the other nodes keep their status unchanged .

The order of adjustment can be selected dynamically or following a pre-setting order ( Here k stands for kth step ).

 

 2Synchronous Mode
All the nodes of neurons are adjusted simultaneously.