W and b in neural network algo

AI Discussions
1

how to know w’s and b’s .In the code they generated automatically.

2

Hi @Akash_das2110
Welcome to the community!

There are many ways to print the weights of each layer like

model.layers[..].get_weights()[..]

also, I think this post will be have a good answer with examples about how to get weights

Cheers,
Abdelrahman

3

thanks @AbdElRhaman_Fakhry but what mechanism going behind it…I want to know that.

4

I mean how this process choose the correct w and b that minimize the cost function?

5

@Akash_das2110
The model initialize the weights W with random normal values, initialize the values of b with zeros, and the gradient descent algorithms start to change the values of weights according to get prediction(y_predict = W*X+b) to get a close values to the output values y to minimize the cost(the far between the prediction y_predict and the real values y ) also I thinks this topic will help you

Note if you want to learn more about gradient descent algorithm you could take Deep learning specialization that will show you how the gradient descent mechanism, and the types of the gradient descent algorithm in addition to many other techniques

please feel free to ask any questions,
Cheers,
Abdelrahman

6

ohh…Now its clear…thanks for your time.

7

Hello @Akash_das2110!

Besides the sources mentioned by @AbdElRhaman_Fakhry, this article will help you a lot to understand how to choose correct w, b, and minimize cost. Also, consider taking the MLS and then DLS courses.

Best,
Saif.

1 Like
8

thanks mate…the visual representation is really helpful.

9

Hey @Akash_das2110! I am glad you found it helpful. Stay connected, as part 2 is coming this Saturday.

1 Like
10

Ok nice …Currently I am looking for summer internship , can you please guide me for this?

11

hey @AbdElRhaman_Fakhry Currently I am looking for summer internship , can you please guide me for this?