Saved in:
Bibliographic Details
Main Authors: Harper, Marc, Safyan, Joshua
Format: Preprint
Published: 2020
Subjects:
Online Access:https://arxiv.org/abs/2007.02449
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911247607267328
author Harper, Marc
Safyan, Joshua
author_facet Harper, Marc
Safyan, Joshua
contents We combine momentum from machine learning with evolutionary dynamics, where momentum can be viewed as a simple mechanism of intergenerational memory. Using information divergences as Lyapunov functions, we show that momentum accelerates the convergence of evolutionary dynamics including the replicator equation and Euclidean gradient descent on populations. When evolutionarily stable states are present, these methods prove convergence for small learning rates or small momentum, and yield an analytic determination of the relative decrease in time to converge that agrees well with computations. The main results apply even when the evolutionary dynamic is not a gradient flow. We also show that momentum can alter the convergence properties of these dynamics, for example by breaking the cycling associated to the rock-paper-scissors landscape, leading to either convergence to the ordinarily non-absorbing equilibrium, or divergence, depending on the value and mechanism of momentum.
format Preprint
id arxiv_https___arxiv_org_abs_2007_02449
institution arXiv
publishDate 2020
record_format arxiv
spellingShingle Momentum Accelerates Evolutionary Dynamics
Harper, Marc
Safyan, Joshua
Machine Learning
Information Theory
Dynamical Systems
We combine momentum from machine learning with evolutionary dynamics, where momentum can be viewed as a simple mechanism of intergenerational memory. Using information divergences as Lyapunov functions, we show that momentum accelerates the convergence of evolutionary dynamics including the replicator equation and Euclidean gradient descent on populations. When evolutionarily stable states are present, these methods prove convergence for small learning rates or small momentum, and yield an analytic determination of the relative decrease in time to converge that agrees well with computations. The main results apply even when the evolutionary dynamic is not a gradient flow. We also show that momentum can alter the convergence properties of these dynamics, for example by breaking the cycling associated to the rock-paper-scissors landscape, leading to either convergence to the ordinarily non-absorbing equilibrium, or divergence, depending on the value and mechanism of momentum.
title Momentum Accelerates Evolutionary Dynamics
topic Machine Learning
Information Theory
Dynamical Systems
url https://arxiv.org/abs/2007.02449