made `makemoons_comparison.ipynb` standalone notebook

probml / pyprobml

Python code for "Probabilistic Machine learning" book by Kevin Murphy

MIT License

6.45k stars 1.52k forks source link

made `makemoons_comparison.ipynb` standalone notebook #1079

Closed AnandShegde closed 2 years ago

AnandShegde commented 2 years ago

Description

Created notebook comparing the following techniques on Moons dataset.

MLP
Deep Ensemble
MC Dropout
Bootstrap
NUTS
GP
SNGP
VI

Figure Number

Figures

Issue

Steps

Checklist:

[x] Performed a self-review of the code
[x] Tested on Google Colab.

Potential problems/Important remarks

review-notebook-app[bot] commented 2 years ago

Check out this pull request on

See visual diffs & provide feedback on Jupyter Notebooks.

Powered by ReviewNB

murphyk commented 2 years ago

LGTM. One concern is that SNGP uses DeepResNet(tf.keras.Model) which is a different model to cmlp, so we're conflating two things here - the architecture and the estimation method. But changing this may be hard, so I will merge for now.