Why does regularization use L1, L2 not others?

Created
TagsRegularization

They can cover the majority of the applications and a linear combination of L1 and L2 approximates any norm to second order at the origin.