dpsimpson

Default prior on the negative binomial is very informative

11

#### Summary: The default prior for the over-dispersion parameter of the negative binomial likelihood puts a lot of prior mass on large amounts of over-dispersion #### Description: This is found...

SBC histogram

6

There is a code for making the histogram for the SBC paper in the sbc_histogram branch.

## Comments on first 5 chapters Chapter 1: - Expand out the writing but good - Good intro to BDA. Maybe need to expand out bayes a bit later (Chapter...

Bug: Test and training data the same in examples/Customizing_embeddings.ipynb

1

I'm pretty sure there's a typo in the `optimize_matrix` routine in block 10 of the customizing_embeddings notebook. The code ``` e1_train, e2_train, s_train = tensors_from_dataframe( df[df["dataset"] == "train"], "text_1_embedding", "text_2_embedding",...

[WIP] Sparse cholesky

8

## Summary Decided to dive in on the sparse cholesky. This is a WIP. - [ ] Implement prim sparse cholesky - [ ] Implement prim sparse triangular solves -...

Bug: `arena_matrix` traits aren't specialized properly for sparse matrices

3

## Description This is one of those things where I can't tell if it's a bug or a user error, but I'm trying to use an `arena_matrix` type and I'm...