Adaptive agents in changing environments, the role of modularity

Raffaele Calabretta, Juan Neirotti

Research output: Contribution to journalArticle

Abstract

We explored the role of modularity as a means to improve evolvability in populations of adaptive agents. We performed two sets of artificial life experiments. In the first, the adaptive agents were neural networks controlling the behavior of simulated garbage collecting robots, where modularity referred to the networks architectural organization and evolvability to the capacity of the population to adapt to environmental changes measured by the agents performance. In the second, the agents were programs that control the changes in network's synaptic weights (learning algorithms), the modules were emerged clusters of symbols with a well defined function and evolvability was measured through the level of symbol diversity across programs. We found that the presence of modularity (either imposed by construction or as an emergent property in a favorable environment) is strongly correlated to the presence of very fit agents adapting effectively to environmental changes. In the case of learning algorithms we also observed that character diversity and modularity are also strongly correlated quantities.

Original languageEnglish
Pages (from-to)257-274
Number of pages18
JournalNeural Processing Letters
Volume42
Issue number2
Early online date14 May 2014
DOIs
Publication statusPublished - 2015

Fingerprint

Garbage
Learning
Learning algorithms
Population
Weights and Measures
Robots
Neural networks
Experiments

Bibliographical note

*

Keywords

  • artificial life simulations
  • emergence of modularity
  • evolutionary robotics
  • evolvability

Cite this

@article{d3526721192d472bae10aaf933fd7b87,
title = "Adaptive agents in changing environments, the role of modularity",
abstract = "We explored the role of modularity as a means to improve evolvability in populations of adaptive agents. We performed two sets of artificial life experiments. In the first, the adaptive agents were neural networks controlling the behavior of simulated garbage collecting robots, where modularity referred to the networks architectural organization and evolvability to the capacity of the population to adapt to environmental changes measured by the agents performance. In the second, the agents were programs that control the changes in network's synaptic weights (learning algorithms), the modules were emerged clusters of symbols with a well defined function and evolvability was measured through the level of symbol diversity across programs. We found that the presence of modularity (either imposed by construction or as an emergent property in a favorable environment) is strongly correlated to the presence of very fit agents adapting effectively to environmental changes. In the case of learning algorithms we also observed that character diversity and modularity are also strongly correlated quantities.",
keywords = "artificial life simulations, emergence of modularity, evolutionary robotics, evolvability",
author = "Raffaele Calabretta and Juan Neirotti",
note = "*",
year = "2015",
doi = "10.1007/s11063-014-9355-8",
language = "English",
volume = "42",
pages = "257--274",
journal = "Neural Processing Letters",
issn = "1370-4621",
publisher = "Springer",
number = "2",

}

Adaptive agents in changing environments, the role of modularity. / Calabretta, Raffaele; Neirotti, Juan.

In: Neural Processing Letters, Vol. 42, No. 2, 2015, p. 257-274.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Adaptive agents in changing environments, the role of modularity

AU - Calabretta, Raffaele

AU - Neirotti, Juan

N1 - *

PY - 2015

Y1 - 2015

N2 - We explored the role of modularity as a means to improve evolvability in populations of adaptive agents. We performed two sets of artificial life experiments. In the first, the adaptive agents were neural networks controlling the behavior of simulated garbage collecting robots, where modularity referred to the networks architectural organization and evolvability to the capacity of the population to adapt to environmental changes measured by the agents performance. In the second, the agents were programs that control the changes in network's synaptic weights (learning algorithms), the modules were emerged clusters of symbols with a well defined function and evolvability was measured through the level of symbol diversity across programs. We found that the presence of modularity (either imposed by construction or as an emergent property in a favorable environment) is strongly correlated to the presence of very fit agents adapting effectively to environmental changes. In the case of learning algorithms we also observed that character diversity and modularity are also strongly correlated quantities.

AB - We explored the role of modularity as a means to improve evolvability in populations of adaptive agents. We performed two sets of artificial life experiments. In the first, the adaptive agents were neural networks controlling the behavior of simulated garbage collecting robots, where modularity referred to the networks architectural organization and evolvability to the capacity of the population to adapt to environmental changes measured by the agents performance. In the second, the agents were programs that control the changes in network's synaptic weights (learning algorithms), the modules were emerged clusters of symbols with a well defined function and evolvability was measured through the level of symbol diversity across programs. We found that the presence of modularity (either imposed by construction or as an emergent property in a favorable environment) is strongly correlated to the presence of very fit agents adapting effectively to environmental changes. In the case of learning algorithms we also observed that character diversity and modularity are also strongly correlated quantities.

KW - artificial life simulations

KW - emergence of modularity

KW - evolutionary robotics

KW - evolvability

UR - http://www.scopus.com/inward/record.url?scp=84940606881&partnerID=8YFLogxK

U2 - 10.1007/s11063-014-9355-8

DO - 10.1007/s11063-014-9355-8

M3 - Article

VL - 42

SP - 257

EP - 274

JO - Neural Processing Letters

JF - Neural Processing Letters

SN - 1370-4621

IS - 2

ER -