~~Hi, @cuishuhao , thanks for the code release.
I have some confusion about the details of the paper.
According to the inequality (5) in original paper, the two norms can bound each other. Thus, optimizing F-norm is equal to optimizing nuclear-norm..
Thanks~~
'bound each other' does not equal to 'the bigger, the bigger', such as 1<2<4<8(1,4 as a group,28 as the other group) then it might become 1.2<1.8<4.8<7.2. One group becomes bigger, another becomes smaller.
As in the paper, F-norm only denotes the discriminability, while nuclear norm denotes both the discriminability and diversity.
~~Hi, @cuishuhao , thanks for the code release. I have some confusion about the details of the paper. According to the inequality (5) in original paper, the two norms can bound each other. Thus, optimizing F-norm is equal to optimizing nuclear-norm.. Thanks~~