Running HMC for mixture Gaussian distribution


#1

Hey; I have a multidimensional mixture model which is

0.1*Normal(mean_1,cov_1) + 0.2*Normal(mean_2,cov_2) + 0.3*Normal(mean_3,cov_3) + 0.4*Normal(mean_4,cov_4)

Can anyone gives any hint on how to run HMC on this posterior distribution ? Thanks


#2

Have you taken a look at the Pyro Gaussian Mixture Model tutorial? Our HMC implementation supports enumeration of discrete variables, so you should be able to modify the code there to work with HMC.