How does random sampling from matrix valued distribution work in general?
In the univariate case we can draw from the uniform distribution over the interval from 0 to 1, then plug that value into the cdf of the distribution we want to sample from. Is there a similar procedure for matrix valued distributions?