Code for reproducing the experiments found in our GLBIO oral presentation: "Towards Computing Attributions for Dimensionality Reduction Techniques"
To reproduce the results from the paper, you first want to install our package from TestPyPi:
pip install -i https://test.pypi.org/simple/ interpretable-tsne
Alternatively, you can install from source using the following git repo:
https://github.com/MattScicluna/interpretable_tsne
Dont forget to install the required packages:
pip install -r requirements.txt
If you want to run the jupyter notebooks, then you will want to set up a kernel for your jupyter notebook:
python -m ipykernel install --user --name <YOUR_ENV_NAME> --display-name "Python (<YOUR_ENV_NAME>)"
-
Download the datasets
bash create_experiment_files.sh
Note that all datasets (except for SARS-CoV-2) are available here
-
Define paths in
set_environment_variables_public.sh
-
Run experiments
bash sim_data_experiment.sh bash mnist_experiment.sh bash mnist_create_table.sh bash sars_cov_2_experiment.sh
-
Check out the Notebooks to generate the figures found in the paper!
We have a tutorial notebook available in notebooks/synthetic_data_tutorial.ipynb