-
For future flexibility we might want to make all priors submodels that produce vectors of the right length and that have a default mode that wraps the input prior from a user. This would make it easie…
-
The complete implementation will probably have to wait for a new configuration system, but we can at least start thinking about how this could be done and try to find libraries that could support us i…
-
Nico has devised a method for allowing family priors to be given with a skewed dirichlet that is unique per feature. The skew is toward persistence of one state or another of the feature, deformed by …
-
Hi,
I have encountered an error while trying to sample from the prior of a Turing model with an `LKJCholesky` prior. The error seems to stem from `PDMats` calling `oneunit(Any)` at some point.
T…
-
It would be great if nimble could use Jeffreys-type improper distributions for normal and multivariate normal models. Nimble could exploit the conjugacy of these priors.
For example, if , the Jeff…
-
In the recent version of BAMM, the option validateEventConfiguration is set to true (enabled). This is a rather unfortunate default option because it seems to invalidate the MCMC. More specifically, w…
-
This is a question/feature request for policy gradient based methods (e.g. A2C). Is it possible to specify a prior for the policy before training?
For instance, if I have 3 possible discrete Actio…
-
From Alexei:
While looking at the LPhyBEAST code with Sebastian Hoehna we noticed that there is a bug for RealParameters in the DefaultOperatorSchedule. If the parameter comes from a Normal or Lapl…
-
This is some pseudo-code that could make `StationaryRandomFields.jl` compatible with Comrade.
Comrade expects its priors to be a `Distributions.jl` object. Since we want a general multivariate dist…
-
Hi, I wrote a small function that does a job similar to `prior_from_idata`, but can also handle log-normal distributions. @ricardoV94 suggested it could find its place in `pymc-experimental`, or at le…