This repository contains the implementation of the paper: A Confidence Interval for the $\ell_2$ Expected Calibration Error.
All simulation results in our paper can be reproduced with the code in experiment/.
Reproducing our simulation by:
python run_simulation.py --setting simu1
To get results for all baselines, run:
bash run_simulation.sh