0
$\begingroup$

My problem with FFNN is that I do not understand in which use cases this network makes sense. Does anyone have an example where this is used? Once I read on the internet that it could be used for YES/NO Questions. For example the inputs are attributes of a human and after going through all the hidden nodes there will be an answer like 1 or 0 if that human is married or some sort of stuff. My Question now is: How do you define those hidden nodes. I mean every Innput Node is connected to every hidden node in the first layer. Why do I need more than one hidden node. I think I just need a good example where I understand how to define hidden nodes.

$\endgroup$
  • $\begingroup$ So, what is your question: 1. when to use FFNNs or 2. how to determine the number of hidden nodes? $\endgroup$ – nbro Apr 8 at 10:08
  • $\begingroup$ More like when to use it to understand how to determine the number of hidden nodes :D $\endgroup$ – Justin Holze Apr 8 at 11:48
0
$\begingroup$

We can add multiple hidden layers which contain neurons in a Feed Forward neural network.

For a neuron in the hidden layer,

$ output = act( w . x + b )$

Where $act$ is the activation function.

Here, the value of $x$ could be the input features ( if the neuron lies in the first hidden layer ) or it could be the activations from the previous layer ( if the neuron lies in an intermediate hidden layer ).

With some more comfort, we can write the above equation as,

$ output = act(\sum_{i=0}^{N} w_i x_1 + b ) $

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.