upvote
In ML, "best" is what minimises the loss function, so in a certain (very basic) sense it does.
reply