We need some easy to follow instructions on how to use the core Stan inside a user-written C++ program. See stan-dev/stan#3085 for example.
The instruction can simply guide through the task of compiling one of the models and running MCMC with the services. The biggest challenges are typically all the dependencies that we need to include in the C++
The current example on MDN from Edward tutorials needs small modifications to run on edward2. Documentation covering these modifications will be appreciated.
Be conservative about the changes necessary to the core of DAPPER.
ATM, this requires pretending variable-length state or obs are fixed-length. This enables using np.ndarrays without hassle, but includes overhead.
Suppose one chain is stuck on one mode; another on another mode. If those two chains may sample independently from each mode, the ESSs will be high when, really, they should be near zero since the samples don't represent anything like independently from the overall distribution. This is why multichain ESS makes sense and we should implement it. I feel like this will give a much better picture of t
This blogpost from Lyndon White mentions several antipatterns for Julia code: https://white.ucc.asn.au/2020/04/19/Julia-Antipatterns.html (thanks @bauglir for pointing this out). Some of the antipatterns mentioned here are also present in the FL code.
Add a description, image, and links to the bayesian-methods topic page so that developers can more easily learn about it.
To associate your repository with the bayesian-methods topic, visit your repo's landing page and select "manage topics."
The currently implemented version of the horseshoe distribution is not the parameterization that most ML papers use. This limits the ease of use of this as, for example, a prior in a tfp.layers.KLDivergenceAddLoss or in tfp.layers.DenseReparameterization. The regularized horseshoe would also be useful as an implemented distribution.
The alternative parameterization is shown here:
https://www.