So picking better samples (more efficient sampler) is just decreasing the runtime since we don't need as many samples? Both methods will still get you the right answer right?
gloose
Specifically, how do we define cost? Is it strictly about the number of samples required to get the same variance, or does it also take computation time per sample and other possible costs into account?
So picking better samples (more efficient sampler) is just decreasing the runtime since we don't need as many samples? Both methods will still get you the right answer right?
Specifically, how do we define cost? Is it strictly about the number of samples required to get the same variance, or does it also take computation time per sample and other possible costs into account?