How to define a likelihood of a mixture model?

[tumble-weed]

Hi, I want to create a mixture of 2 gaussian processes G1 and G2, where the mixing ratio comes from a gaussian process with a bernoulli likelihood B. it involves 3 gaussian processes, G1,G2 and B , where the posterior will be G1 B + (1-B) G2 (similar but simpler than the idea in this paper: https://papers.nips.cc/paper/1900-mixtures-of-gaussian-processes.pdf). i am not sure if this is simple to do in gpytorch, any help would be appreciated

[jacobrgardner]

@tumble-weed I think this should be straightforward if you use our Pyro interface, which should let you define this model and do inference pretty easily. We have examples of our Pyro integration here: https://github.com/cornellius-gp/gpytorch/tree/master/examples/07_Pyro_Integration

[tumble-weed]

thanks, i will take a look