Users can access the the used dataset in our paper through google driver as shown in the following link.
https://drive.google.com/drive/folders/1idb8NpiCaAzHLaklw6KhTxuJnIQYiM9W?usp=drive_link
Users can train MILLION framework through the following code:
python main.py
For the hyper-parameters, users can manually set in config.py file.
Users can follow the jupyter notebooks, i.e., interpolation_plot.ipynb and improvement_plot.ipynb, to conduct risk control experiments.