/mechanism-learning

Source code for reproducing the experimental results reported in T. Osogami et al., "Learning Efficient Truthful Mechanisms for Trading Networks", IJCAI-23.

Primary LanguageJupyter NotebookApache License 2.0Apache-2.0

Mechanism-Learning

Scope

Python code for reproducing the experimental results reported in Takayuki Osogami, Segev Wasserkrug, and Elisheva Shamash, "Learning Efficient Truthful Mechanisms for Trading Networks", in Proceedings of the 32nd International Joint Conference on Artificial Intelligence (IJCAI-23).

Usage

To install the package:

conda env create -n mechlearn -f environment.yml
conda activate mechlearn
pip install .

To run all of the experiments:

cd exp
./exp.sh

To draw figures in the paper, run the jupyter notebooks under notebooks.

Authors