Scaling MAP-Elites to Deep Neuroevolution - Archive ouverte HAL Access content directly
Conference Papers Proceedings of the 2020 Genetic and Evolutionary Computation Conference Year : 2020

Scaling MAP-Elites to Deep Neuroevolution

(1) , (2) , (2) , (3)
1
2
3
Cédric Colas
Vashisht Madhavan
  • Function : Author
Joost Huizinga
  • Function : Author
Jeff Clune
  • Function : Author

Abstract

Quality-Diversity (QD) algorithms, and MAP-Elites (ME) in particular, have proven very useful for a broad range of applications including enabling real robots to recover quickly from joint damage, solving strongly deceptive maze tasks or evolving robot morphologies to discover new gaits. However, present implementations of ME and other QD algorithms seem to be limited to low-dimensional controllers with far fewer parameters than modern deep neural network models. In this paper, we propose to leverage the efficiency of Evolution Strategies (ES) to scale MAP-Elites to high-dimensional controllers parameterized by large neural networks. We design and evaluate a new hybrid algorithm called MAP-Elites with Evolution Strategies (ME-ES) for post-damage recovery in a difficult highdimensional control task where traditional ME fails. Additionally, we show that ME-ES performs efficient exploration, on par with state-of-the-art exploration algorithms in high-dimensional control tasks with strongly deceptive rewards.
Fichier principal
Vignette du fichier
2003.01825(1).pdf (5.44 Mo) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-03099878 , version 1 (06-01-2021)

Identifiers

Cite

Cédric Colas, Vashisht Madhavan, Joost Huizinga, Jeff Clune. Scaling MAP-Elites to Deep Neuroevolution. GECCO 2020 - Genetic and Evolutionary Computation Conference, Jul 2020, Cancun / Virtual, Mexico. ⟨10.1145/3377930.3390217⟩. ⟨hal-03099878⟩

Collections

ENSTA INRIA INRIA2
54 View
151 Download

Altmetric

Share

Gmail Facebook Twitter LinkedIn More