Reproduce “Convergence of Time-Averaged Mean Field Gradient Descent Dynamics for Continuous Multi-Player Zero-Sum Games” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Yulong Lu, Pierre Monmarché
The authors introduce mean-field dynamics in which each player responds to an exponentially discounted memory of other players’ strategy distributions. Memory decays at rate α, while players move at uniform speed. For compact torus games with smooth pairwise antisymmetric interactions, sufficiently small α yields exponential convergence to the unique entropy-regularized equilibrium. This continuous-time guarantee does not cover finite-particle implementations.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.