What if I take my model’s predictions of future observations, and hash them in order of posterior probability until I run out of time or get a hit?
What if I take my model’s predictions of future observations, and hash them in order of posterior probability until I run out of time or get a hit?