Matrices multiplications


Just wanted to share sth about matrices multiplications, because I think the schema that is shown in the 11th video is pretty complicated compared to what I learned in school.

(I’m reacting to this picture)

Here is how I learned it in school: when you multiply matrices A and B, imagine B “elevating”. This way, the result placed below is placed such a way that each resulting number is at the intersection of the adequate row & column.

I find it way easier to keep track of the rows and column with that in mind.

It works with matrix (dot) vector, vector (dot) vector too indeed, and helps getting intuition of the shape required for the dot product to work.

Hope it helps.

Hello @Julien, thank you for your suggestion. We really appreciate that!

I will pass it to our Curriculum Developer team!


1 Like

How we were taught matrix multiplication in high school does little to provide the intuition for what much of machine learning in based on - moving input vectors in space (in lay terms), and formally linear transformations and change in basis. This course gets into change in basis a bit, but the following videos provide a good visual perspective on the concept:

Linear Transformations and Matrices

Matrix Multiplication