William J. Wilkinson comments

Results 8 comments of


                                            William J. Wilkinson

question about the equation (64) in `Bayes-Newton` paper

Hi, sorry I never got around to this issue. You are correct, the pre-print does have an mistake in it. Equation 63 is actually incorrect, the likelihoods should not be...

Cannot run demo, possible incompatibility with latest Jax

Hi, Sorry for the slow response and apologies that you've been having issues with the package versions. This is indeed frustrating. I would love to update the package to work...

Double Precision Issues

Hi Harrison, That's great to hear that you've found the package useful, and thanks for sharing your paper - it looks great! I agree that `float64` is generally needed when...

set up on a new dataset and predict on new data points

Hi @andrewcztrack Glad to see you are trying things out. Everything looks OK except that you're using the cross-validation indices that we stored specifically for a different (smaller) data set....

set up on a new dataset and predict on new data points

That error is because the data does not divide evenly into 5 batches. You could truncate the data slightly to fix it. However, didn't you say that you wanted to...

Missing data for air_quality demo

Hi, You can find that file here: https://drive.google.com/file/d/1QcbzFJyAr95zY-B1r-0MkzcX7TReaF6C/view?usp=sharing

question about the equation (64) in `Bayes-Newton` paper

The sparse GP section of the paper has now been corrected in the latest arxiv version: https://arxiv.org/abs/2111.01721 I hope this clears things up.

Approximate Power-EP for Deep GPs with intermediate variables for density models/latent variable models

Hi thangbui, Firstly, thanks a lot for sharing this code, it's great. Does the repo contain code for a Deep GP latent variable model, or is this not fully implemented...