Open anandmishra22 opened 6 years ago
It is similar with NCA, but not the same.
It is my own loss under experiment, you can ignore it . It is based on the assumption that the distribution of any class is Mixture Gaussian.
It is similar with NCA, but not the same.
It is my own loss under experiment, you can ignore it . It is based on the assumption that the distribution of any class is Mixture Gaussian.