How do multiple hidden layers in a neural network improve its ability to learn?

Question

Most neural networks bring high accuracy with only one hidden layer, so what is the purpose of several hidden layers?

Answer 1

To answer your question, you first need to find out the reason why the term "deep learning" was coined almost ten years ago. Deep learning is nothing more than a neural network with several hidden layers. The term deeply roughly refers to how our brain transmits sensory inputs (especially the eyes and cortex of vision) through different layers of neurons to draw a conclusion. However, about ten years ago, researchers were not able to train neural networks with more than 1 or two hidden layers due to various problems, such as fading, exploding gradients , getting stuck at local minima, and less effective optimization methods (compared to what is used in present) and some other problems. In 2006 and 2007, several researchers 1 and 2showed some new methods to better train neural networks in more hidden layers, and then the era of deep learning has begun since then.

, , (). , , , X Y. , , , , . . , (, 99%)? :

• ,
• , , , ?
• . , ? !
• , ?
• , , , , .

( ), . . , , . , , :

, ( ) , , , , , , , , .... , , .