This directory contains a Jupyter Notebook that was presented during the following events:
To run the notebook (surrogates_overview.ipynb) you need to install a
collection of Python dependencies listed in the requirements.txt file
(included in this directory) by executing pip install -r requirements.txt.
Alternatively, you can run it via Binder or Colab by clicking the buttons
listed above.
The scripts directory contains a collection of Python modules to help with
building and visualising interactive surrogate explainers of image and tabular
data.
It also implements a simple image classifier based on PyTorch and
a number of interactive iPyWidgets for no-code experiments with these
explainers.
This Jupyter Notebook can also be rendered as a reveal.js slide show through RISE. To this end, it needs to be run from within a Jupyter Notebook environment (not Jupyter Lab). Next:
- execute all cells; and
- launch RISE presentation by clicking the bar chart icon
(
) shown in the
Jupyter Notebook toolbar.
