Discrete Probabilistic Inverse Optimal Transport

Paper accepted to ICML 2022: https://proceedings.mlr.press/v162/chiu22b.html

Folders:
./SRC : source codes for performing PIOT simulations
./synthetic : contains python codes for running batch jobs, Jupyter Notebooks for simulations and plots, and pickle files for data in the paper (Note: several large size files are not included)

Required packages and version
----------------------------
conda
matplotlib 3.3.4
numpy 1.19.1
pandas 1.1.1
pickle
python 3.8.5
pytorch 1.4.0
scikit-learn 0.24.1
scipy 1.5.2
seaborn 0.11.1
tqdm 4.54.0

How to run batch jobs:

1. Check *.py files in ./synthetic/input_files/ and tune the parameters in it (slurm only)
2. Generate batch job files with "python FILE_NAME.py" inside ./synthetic/input_files/
3. Submit jobs by "sh ./synthetic/submit_all.sh" in ./synthetic