Week 3 "Explanation for Vectorized Implementation"

Hello, in the “Explanation for Vectorized Implementation” lecture in Week 3 around 4:06, I am curious how the W^[1]X^(1), W^[1]X^[2], and so on lead to having column vectors. Can someone explain me? Best regards.

Hello @Juheon_Chu,

Let’s take one step back.

W^[1] is a matrix with a shape of (h, k), where h represents the number of nodes/units in the first layer, and k is the number of features in X.

X^(1) is a vector or a column matrix of shape (k, 1), where 1 represents that it is one sample.

So, a matrix multiplication of (h, k) and (k, 1) gives us (h, 1) which is another vector or column matrix.

If it is still confusing, let me know and also it would be helpful that you share your version of understanding so we can figure out where to focus on.


1 Like

Oh…that’s clear! Thank you so much!

1 Like