This post builds on our recent introduction to multi-level modeling with tfprobability, the R wrapper to TensorFlow Probability. We show how to pool not just mean values (“intercepts”), but also relationships (“slopes”), thus enabling models to learn from data in an even broader way. Again, we use an example from Richard McElreath’s “Statistical Rethinking”; the terminology as well as the way we present this topic are largely owed to this book.
This looks really interesting. Do you have any idea what the performance is like compared to fitting this model in rstan?
Thanks! Well, you could always test and report your results here 
Ultimately, I expect TFP to run fast in many cases, but it is still at a release << 1.0 (0.7, as of today) , a lot of features are still being added, and there's the adaptation to TF 2 which requires pretty substantial modifications, so as of today not everything will have been optimized for performance.
This topic was automatically closed after 42 days. New replies are no longer allowed.