How does the number of hidden layers affect the result of the neural network?

Question

Neural networks

The whole question fit in the topic))

Answer 1

The neural network becomes flexible, but at the expense of learning speed.
Able to highlight more complex entities, but at the cost of possible retraining.

Answer 2

Affects.
More layers - directly increases the complexity of the architecture, and increases the number of possible variations of interactions. What is considered the best type to fit the solution.
P.S.
there is an understanding that each new layer in the neural network must REDUCE the entropy with which it works, otherwise it is simply pointless to increase the number of layers, respectively, it is also impossible to increase the number of layers indefinitely (as well as infinitely reduce entropy)