Teaching material

The notebooks here were used to teach GPs in several iterations of this course at Jülich Research Center. Note that the latter resource might be outdated. Please consider the material here to be the up to date reference.

Contents

• 1D exact GP regression
• 2D exact GP regression
• 1D approximate GP regression using Sparse Stochastic Variational Inference

Generate and use notebooks

If you want to run these notebooks using your local Jupyter Lab / Notebook, then follow the steps below. We use jupytext to create notebooks from the Python scripts book/content/teaching/notebook_*.py in this repo.

$ git clone <this repo>
$ cd gp_playground/book/content/teaching/

# One-time setup of venv and ipy kernel. Use https://virtualenvwrapper.readthedocs.io or
#   $ python -m venv --system-site-packages bayes-ml-course-sys
#   $ source ./bayes-ml-course-sys/bin/activate
$ mkvirtualenv --system-site-packages bayes-ml-course-sys
$ pip install -r requirements.txt

# Install custom kernel, select that in Jupyter. --sys-prefix installs into the
# current venv, while --user would install into ~/.local/share/jupyter/kernels/
$ (bayes-ml-course-sys) python -m ipykernel install --name bayes-ml-course --sys-prefix

# This script converts all *.py to *.ipynb
$ (bayes-ml-course-sys) ./py-to-notebook.sh

# Start Jupyter
$ (bayes-ml-course-sys) jupyter-lab notebook_01_one_dim.ipynb