You are the first teacher I saw that actually explains things properly. I can tell you do know what you are talking about, unlike many others on youtube that just pretend to know. Congratulations, you are excellent and thank you for teaching us, you really helped me understand this.
watched with joy! your positive vibe makes the whole thing lot easier and more fun to watch and follow! I didn't even realize how quickliy almost a 20-min long video passed! thanks a bunch
Great energy & explanation, nice calligraphy, amazing blackboard (our markers don't have ink half the time in university). It would be nice to mention what is the rank of matrix and do we need to always manually calculate it for us to reason about the matrix itself.
Was obvious it was rank 3 at a glance; C2 = C1 * 2 and C4 = C1 + C3. If you think about it, geometrically they're pointing in the same direction and so are redundant. Linearly dependent as you say. I need a nice way to do this in a function. I could calculate the determinant / matrix of minors (which gives the inverse matrix which multiplied by the original matrix gives the identity). But then all you have is the identity, you've lost the scalar information... I want to reduce the rank of a neural net to make it more efficient. Say your matrix was my NN. In reality it only exists in 3d space. We've got vectors going the same way. So I want to express it elegantly. So I can represent that as a 3*3 identity matrix * a 3d vector surely? The matrix is implicit so I can reduce the whole matrix to a 3d vector. Is that right??? Surely not... If it is, how do I arrive at that vector? Also, the cap... I sung between liking it and strongly disliking it at various stages in the video. I'll watch some more to make up my mind about it.
I love ur work inrly do support u and all but as a question can't we just find the row echelon(withe the third rule) and just count the non zero rows to get the rank? It's the same and much time efficient and ones again I rly appreciate ur work and keep it going fam🤍🤍
one small correction 14:40 you are talking about switching the columns not the rows but fair enough its visually clear what you are doing so no big deal.
i want to download some videos so i will watch when i go back home coz i dont have wifi at home bt they cant be downloaded please let all your videos to be downloadable
1) Eliminate the second column because it is just double the first column. 2) Eliminate the first column because it is a linear combination of the 3rd and 4th columns. 3) The remaining 3rd, 4th, and 5th columns have at least one 3 by 3 sub-matrix with nonzero determinant, so the rank of A is 3.
Legit reason why videos exist, I find his video very helpful, and if I do struggle to understand then just simply rewatch it all over until I understand
