For parametric families, is it possible to access the gradient of the log likelihood at the maximum?
For parametric families, is it possible to access the gradient of the log likelihood at the maximum?