pyprobml
pyprobml copied to clipboard
probml
Python code for "Probabilistic Machine learning" book by Kevin Murphy
Extend https://github.com/probml/pyprobml/issues/767 by adding rank-1 Gaussian VI to the comparison. The method is described in sec 9.3.5. There is some JAX code in [vb_gauss_lowrank](https://github.com/probml/pyprobml/blob/master/scripts/vb_gauss_lowrank.py). See also [this scirpt](https://github.com/probml/pyprobml/blob/master/scripts/vb_gauss_lowrank_labour_force_demo.py) which reproduces...
# Instructions * [Follow these guidelines](https://github.com/probml/pyprobml/blob/master/notebooks/README.md) ## Chapter 18: Gaussian processes | Figure | Script | Notebook | PR | Author | | :- | :- | :- | :-...
# Instructions * [Follow these guidelines](https://github.com/probml/pyprobml/blob/master/notebooks/README.md) ## Chapter 17: Bayesian neural networks | Figure | Script | Notebook | PR | Author | | :- | :- | :- |...
## Description - Re-implemented ADF for logistic regression. - ADF now uses unscented sigma-point Gaussian quadrature in its update step, as suggested by the [original paper](https://ieeexplore.ieee.org/document/4383733). - Standardized the dataset...
## Description This PR updates the SNGP demo notebook to self contain implementations of SNGP specific layers for `Spectral Normalization` and `Random Features GP`. The implementation is a simplified port...
As part of the SNGP demo (#819, #983 and #1033), implement the SNGP layers instead of using these from [edward2](https://github.com/google/edward2) library.
### When executing this code: try: from probml_utils import latexify, savefig, is_latexify_enabled except ModuleNotFoundError: %pip install -qq git+https://github.com/probml/probml-utils.git from probml_utils import latexify, savefig, is_latexify_enabled ### Error Obtained: Installing build dependencies...
https://github.com/probml/pyprobml/blob/master/notebooks/book1/13/mixexpDemoOneToMany.ipynb The random initialization of parameters and weights occurs within the E-Step (instead of before it), so it's not really making use of the fit parameters from the M-Step in...
On p.25 of book 1 (in the latest available online version dated back June 2023), in Section 1.5.4.2, it is stated that we often normalize each row of the TF-IDF...
