|
To show the general behavior of a MLP, we have further more applied a small simulation for download:
mlp.zip 138 KB
A multilayer-network with random activation-functions and weights.
In the program above you can see a fully connected feedforward multilayer network.
Here you can change the inputs yourself, and then see how the different neurons react.
You can try guessing which activation-function they have, and how great an influence the different neurons have on each other.
Some neurons has a huge influence on others. P.t. there is no sourcecode, since this is an old program.
|