How is a neural network structured?

Question

Neural networks

I don't quite understand how a neural network works. Let's say there is a logical OR operation with input data 1 and 0. After it adjusts all its weights, we will get 1 at the output. But what if the input is 0 and 1, and not 1 and 0 turns out the output will always be 0, what is wrong. Maybe I didn't understand something?
2) with inputs of 1 and 0, how do you prevent weight 5 from going to infinity and weight 6 from going to 0?
After all, this is the easiest option to get the correct answer (in our case)

Perhaps this comes from a misunderstanding of question 1, but it is never too late to clarify.

Answer 1

wiki

Answer 2

The boolean ИЛИdoesn't need a hidden "h" layer - just one is enough.
From i ₁ and i ₂ through the weights w ₁ , w ₂ and the offset b (three parameters in total), immediately to the activator function and to the exit:
b + w ₁ * i ₁ + w ₂ * i ₂
For ИЛИthe weight b = - 1, w ₁ = 2, w ₂ = 2
The activator function must make 0 from values ​​<= 0, and make 1 from values ​​> 0.

Answer 3

I already once advised you. I repeat. Still, try to STUDY the subject of your interest on a regular basis, and not from articles on the web, even on Habré, and not snatch individual facts. And you keep grabbing. By the way, an example with logical operations for learning neural networks, launched somewhere by someone and now walking on the web, from my point of view, does nothing but harm and confuse the student. And from a practical point of view, it makes no sense at all.
There are a lot of good books out there now. Including for beginners. For example
Andrew Trask, Grokay deep learning. ,
Tariq Rashid. Building a Neural Network
Francois Chollet Deep Learning with Python
These are the books to start with.
Then you don't have to run to the forum for every sneeze.