Multicalibration at Web Scale - Replication Materials for Benchmark Experiments

This repository contains replication materials for the benchmark experiments for the paper "Multicalibration at Web Scale"

How to run

The was tested with python 3.12, you need to install the dependencies in the requirements.txt file. You can do this by running the following command in the root directory of the repository:

pip install -r requirements.txt

To generate the results, run:

python run_experiments.py

By default, results will be store in results/.

Acknowledgement

The code is is based on Hansen et al.'s code for their paper "When is Multicalibration Post-Processing necessary" (https://github.com/dutchhansen/empirical-multicalibration). We thank the authors for making their code available under the MIT License (see LICENSE file).