Learning with Opponent-Learning Awareness
Implements the LOLA (AAMAS'18) and LOLA-DiCE (ICML'18) algorithms.
Further resources:
- A pytorch implementation of LOLA-DiCE is available at https://github.com/alexis-jacq/LOLA_DiCE.
- A colab notebook with the nummerical evalution for DiCE is available at https://goo.gl/xkkGxN.
Installation
To run the code, you need to pip-install it as follows:
$ pip install -e .After installation, you can run different experiments using the run scripts provided in scripts/.
Use run_lola.py and run_tournament.py for running experiments from the AAMAS'18 paper.
Use run_lola_dice.py for reproducing experiments from the ICML'18 paper.
Check out notebooks/ for IPython notebooks with plots.
Note: this code is not tested on GPU, so there might be unexpected issues.
Disclaimer: This is a research code release that has not been tested beyond the use cases and experiments discussed in the original papers.
Contribution
Contributions to further enhance and improve the code are welcome.
Please email jakob.foerster at cs.ox.ac.uk and alshedivat at cs.cmu.edu with comments and suggestions.
Citations
LOLA:
@inproceedings{foerster2018lola,
title={Learning with opponent-learning awareness},
author={Foerster, Jakob and Chen, Richard Y and Al-Shedivat, Maruan and Whiteson, Shimon and Abbeel, Pieter and Mordatch, Igor},
booktitle={Proceedings of the 17th International Conference on Autonomous Agents and MultiAgent Systems},
pages={122--130},
year={2018},
organization={International Foundation for Autonomous Agents and Multiagent Systems}
}DiCE:
@inproceedings{foerster2018dice,
title={{D}i{CE}: The Infinitely Differentiable {M}onte {C}arlo Estimator},
author={Foerster, Jakob and Farquhar, Gregory and Al-Shedivat, Maruan and Rockt{\"a}schel, Tim and Xing, Eric and Whiteson, Shimon},
booktitle ={Proceedings of the 35th International Conference on Machine Learning},
pages={1524--1533},
year={2018},
volume={80},
series={Proceedings of Machine Learning Research},
address={Stockholmsmässan, Stockholm Sweden},
month={10--15 Jul},
publisher={PMLR},
pdf={http://proceedings.mlr.press/v80/foerster18a/foerster18a.pdf},
url={http://proceedings.mlr.press/v80/foerster18a.html},
}License
MIT