Open ghost opened 2 years ago
Hi, Is it possible to define the fitness function as a Wasserstein distance between all transformed samples and some predefined distribution?
In the optimization process I only care about the resulting distribution of the y=f(x)
Thank you!
Hi, Is it possible to define the fitness function as a Wasserstein distance between all transformed samples and some predefined distribution?
In the optimization process I only care about the resulting distribution of the y=f(x)
Thank you!