Generative Adversarial Nets for High-dimensional Mean Field Games

Overview

There are two generative models in the data generation business right now: Generative Adversarial Nets (GAN) and Variational Autoencoder (VAE). These two models have different take on how the models are trained. GAN is rooted in game theory, its objective is to find the Nash Equilibrium between discriminator net and generator net. On the other hand, VAE is rooted in bayesian inference, i.e. it wants to model the underlying probability distribution of data so that it could sample new data from that distribution.

In this work, we use GAN to solve high-dimensional mean field games (MFGs), which are beyond reach with existing solution methods. In particular, the MFG is represented by two partial differential equations. The generator and discriminator are used to approximate them, respectively.

A glimpse of experiment

Convergence of training at different dimensions


...

Diffusion of mean field

Each point represents a player and different color represents different time slots.
...

Code

Publication