README.md

Nash Fixed Point Networks (N-FPNs)

This repo provides the code for the paper Learning to Predict Equilibria via Fixed Point Networks (preprint available here), which was joint work by Howard Heaton*, Daniel McKenzie*, Qiuewei Li, Samy Wu Fung, Stanley Osher, and Wotao Yin.

Please cite as

@article{Heaton2021Learning,
    title={Learning to Predict Equilibria via Fixed Point Networks},
    author={Heaton, Howard and McKenzie, Daniel and Li, Qiuwei and Fung, Samy Wu and Osher, Stanley and Yin, Wotao},
    journal={arXiv preprint arXiv:2106.00906},
    year={2021}}

Code for reproducing the experiments of Section 5.1 are in matrix-mames. Code to reproduce the experiments of Section 5.2 are in traffic-routing. The Jupyter notebooks N_FPN_Rock_Paper_Scissors and N_FPN_Toy_Traffic.ipynb contain two tutorial-style examples of using N-FPNs.