Inversion sampling is so much more complicated than rejection sampling...
dtorresr
I can see the initial calculation for the inverses being complicated, but once we have the final equations, is it really that computationally expensive that we would sometimes have to resort to rejection sampling instead?
L100magikarp
Computing the inverse of just sampling from a circle was pretty complicated. I imagine that computing the inverse of a multi-dimensional function that is even slightly more intricate could quickly become extremely difficult if not computationally intractable.
Inversion sampling is so much more complicated than rejection sampling...
I can see the initial calculation for the inverses being complicated, but once we have the final equations, is it really that computationally expensive that we would sometimes have to resort to rejection sampling instead?
Computing the inverse of just sampling from a circle was pretty complicated. I imagine that computing the inverse of a multi-dimensional function that is even slightly more intricate could quickly become extremely difficult if not computationally intractable.