AI Paper Atlas

Research map / Multi-agent systems and games / Games and equilibria

Higher-Order Uncoupled Dynamics Do Not Lead to Nash Equilibrium - Except When They Do

Sarah Asad Toonsi, Jeff S Shamma

NeurIPS 2023 · Multi-agent systems and games › Games and equilibria · approach: Learning in Markov games · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
let a single player use higher-order gradient play (lambda=50, gamma=5) while others use standard gradient playJordan anti-coordination game · 3-player Jordan anti-coordination gameconvergence to mixed NEImproved
perturb payoffs with random normal noise (std 0.3 and 0.5) (ablation)higher-order gradient play dynamics · perturbed Jordan anti-coordination gamestabilization of unique mixed NE: std 0.3 and 0.5No clear effect
small diagonal perturbation delta_i (0.3877, 0.1446, 0.1352) (ablation)higher-order gradient play dynamics · diagonally perturbed Jordan gameconvergence to mixed NE: delta 0.3877, 0.1446, 0.1352No clear effect
larger diagonal perturbation delta_i (0.8831, 0.4259, 0.7546) (ablation)higher-order gradient play dynamics · diagonally perturbed Jordan gameconvergence to mixed NE: delta 0.8831, 0.4259, 0.7546Hurt
rescale payoff matrix with mu=5 (all players use higher-order play) (ablation)higher-order gradient play · rescaled Jordan anti-coordination gamestability: mu=1 stable -> mu=5 unstableHurt
rescale payoff matrix with mu=0.1 (ablation)higher-order gradient play · rescaled Jordan anti-coordination gamestability: stable for mu > 0.113; unstable at 0.1Hurt
use inherently unstable higher-order dynamics for one player (lambda=0.5, gamma=20, lambda_2=50, gamma_2=1)gradient play · identical-interest coordination gameconvergence to mixed NE (1/2,1/2)Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy