Hej
Thank you for your excellent work.
I have a few questions.
How did you calculate the sample probability for the augmentation experiment (table II in the paper)?
Because I tried to re-engineer what you did but it seems that you didn't try to make the damage distribution uniform after damage injection.
Also, In fig1 of your paper, it appears that you have more damage types (D00, D10, D20 to be specific) on the Japan train+eval data set in your Jupyter notebook.

Thank you for your time.