Alexey Radul
Alexey Radul
As Taylor noted in code review yesterday, Puma is full of accidental use of C++ copy constructors in places where passing objects (including asymptotically large objects like vectors and sets)...
Minor discrepancy between Lite and Puma: In Lite, the inference action `(bogo_possibilize scope block)` runs for one transition (as is typically desired), but in Puma it crashes demanding a fourth...
The model in question is ``` [assume a (normal 0 10)] ;; 20 times [observe (normal a 1) (normal 5 1)] ; Generate a random data set ;; Expect the...
Certainly it is if all the drg nodes are deterministic or principal (and the likes of #462 and #470 are fixed), but what about likelihood-free consequences or brush? Or down-stream...
By analogy with HMC and Slice (both of which were implemented considerably later). #208 seems relevant.
The original example for #415, namely `test/inference_language/test_enumerative_gibbs.py:testOccasionalRejection` actually shouldn't ever reject the Gibbs transitions, but does because the Particle class computes the correction wrong (and differently in Lite and Puma).
For example, here's a way to make a 2-D clustering model with a fixed number of clusters: ``` [assume c1_mu (mvn (array 0 0) (id_matrix 2))] [assume c1_sig (inv_wishart ...)]...
I recently reviewed the semantics intended for the `madeSpLogDensityOfDataBound` method. There was a comment in the source to the effect that the bound was supposed to account for all possible...
A multivariate Gaussian conditioned on values for some of its dimensions is again a multivariate Gaussian (on the remaining dimensions), with analytically known mean and covariance. We even have the...
[This is a variant of #462]. Specifically, consider single-site Gibbs on a CRP choice that produced a singleton pre-Gibbs. The current Gibbs implementation will - Ask that CRP what the...