Ali Ghodsi, Lec 9: SPCA, Nystrom Approximation, NMF

[@kentjefferson3906]

This is the most beautiful explanation of Nystrom methods I've ever seen. This guy is just THAT GOOOOD !!

[@RajivSambasivan]

He has a real gift for teaching and expressing really complex ideas in a very accessible manner. All of his lectures are simply fantastic.

[@wenweilin4794]

This series of lectures is amazing. Please keep uploading. Thanks you professor.

[@liyangxie1920]

I am not sure if I misunderstand something or not: in 33:13, in order for (\Sigma)^{1/2} to have inverse, rank(\Sigma)=m, (means each diagonal of \Sigma is nonzero), and of course we have rank(\Sigma)=rank(A) \leq rank(K), so we have m \leq rank(K), but not m \geq rank(K). am I right?

[@mouadet-tali4089]

@55:23 start of NMF