ConceptEngineering
Low rank approximation
Getting an approximation of a full-rank matrix that is low rank to allow fast multiplications and efficient computations.
Know first
Named as a prerequisite by the lectures that define this. The number is how many say so. A grey one has no page yet.
Singular value decomposition2Convolutional neural networks1Matrix decomposition1Matrix norm1Matrix rank1Rank1SVD1
Opens up
Concepts that name this one as a prerequisite.
Lecturers
2Each in their own terms, from the easiest level.
Stephen Boyd
Finding a matrix of lower rank that approximates a given matrix by minimizing a specific norm of the difference.Song Han
Low rank approximation breaks a convolution or fully connected layer into smaller matrices or layers to reduce computational complexity.
Connections
7The lectures that define it or assume it, and the concepts next to it.