David Augustin

Let me retreat from this issue for the moment. I got stuck with a number of other things for the next days.

I think this looks really good and fits very nicely into the pints interface @Rebecca-Rumney ! A little bit unrelated to the API, I am wondering whether it is actually...

I agree, I like the likelihood as you proposed! I guess I was more wondering whether we would want to store the 3 dimensional tensor in the MCMC controller in...

Sum of PDFs with disjoint parameter sets should be easy enough. Somewhat analogous to 'pints.SumOfIndependentLogPDFs', but conditioning on disjoint parameter sets and and introducing some rule for the mapping, say...

But having varying psi messes up the intended functionality for Likelihoods a bit in p(psi|theta).

So would it be acceptable to alter the values in the likelihood p(psi|theta), even if usually the "data" is not meant to change. So a quick solution would be to...

@MichaelClerx @ben18785 @martinjrobins @chonlei Additionally to the really cool HierarchicalLogPrior that @simonmarchant is implementing, I've been thinking about how one could compose hierarchical posteriors that are not constrained to Normally...

I realised that for gradient based sampling or optimisation, we would need to be able to compute the partials of the top-level likelihoods with respect to both, it's inputs \theta...

> Hi David. Did you use parameter transforms for NUTS? We’ve found that NUTS runs much faster when using them (which is what Stan does)... > […](#) Hi Ben :)...

I am using Gaussian priors truncated at 0. The boundary at zero should not be a problem because of the log-transforms no? I am able to infer the posteriors though...