Models are written to be compatible with the Pykeen package.
Download BetaE data from here and
put it in the top-level directory, keeping the folder name KG_data.
To recreate the experiment data, first run each of naive_experiments.py, se_experiments.py, transe_experiments.py,
translational_experiments.py, se_orthogonal_experiments.py, and translational_orthogonal_experiments.py.
The results of the above runner scripts will create data in an opinionated subdirectory structure within the data directory.
This data can then be read in and used to recreate the paper figures.
See format_results.ipynb for the figure creation code.