Probably Approximately Wrong

An infrequent blog, by Nicola Branchini

Recent posts

Nov 8, 2022
Simple Monte Carlo is independent of dimension. Or is it ? It is essentially the status quo to claim that the error (Mean Squared Eror = MSE) of Monte Carlo integration is “independent of the dimension” of the variable being integrated, or some equivalent variant of this statement.…
May 26, 2022
Combining independent and unbiased estimators I came across this 1.25 page paper by Don Rubin and Sanford Weisberg (Rubin & Weisberg) in Biometrika from 1975.…
Feb 6, 2020
Sequential Monte Carlo and Improved Auxiliary Particle Filters In this post, my aims are: Introduce Bayesian inference in state space models Introduce approximate inference using importance sampling, in state space models.…