Cargando…
Spurious normativity enhances learning of compliance and enforcement behavior in artificial agents
How do societies learn and maintain social norms? Here we use multiagent reinforcement learning to investigate the learning dynamics of enforcement and compliance behaviors. Artificial agents populate a foraging environment and need to learn to avoid a poisonous berry. Agents learn to avoid eating p...
Autores principales: | , , , , , |
---|---|
Formato: | Online Artículo Texto |
Lenguaje: | English |
Publicado: |
National Academy of Sciences
2022
|
Materias: | |
Acceso en línea: | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8784148/ https://www.ncbi.nlm.nih.gov/pubmed/35022231 http://dx.doi.org/10.1073/pnas.2106028118 |
_version_ | 1784638670911832064 |
---|---|
author | Köster, Raphael Hadfield-Menell, Dylan Everett, Richard Weidinger, Laura Hadfield, Gillian K. Leibo, Joel Z. |
author_facet | Köster, Raphael Hadfield-Menell, Dylan Everett, Richard Weidinger, Laura Hadfield, Gillian K. Leibo, Joel Z. |
author_sort | Köster, Raphael |
collection | PubMed |
description | How do societies learn and maintain social norms? Here we use multiagent reinforcement learning to investigate the learning dynamics of enforcement and compliance behaviors. Artificial agents populate a foraging environment and need to learn to avoid a poisonous berry. Agents learn to avoid eating poisonous berries better when doing so is taboo, meaning the behavior is punished by other agents. The taboo helps overcome a credit assignment problem in discovering delayed health effects. Critically, introducing an additional taboo, which results in punishment for eating a harmless berry, further improves overall returns. This “silly rule” counterintuitively has a positive effect because it gives agents more practice in learning rule enforcement. By probing what individual agents have learned, we demonstrate that normative behavior relies on a sequence of learned skills. Learning rule compliance builds upon prior learning of rule enforcement by other agents. Our results highlight the benefit of employing a multiagent reinforcement learning computational model focused on learning to implement complex actions. |
format | Online Article Text |
id | pubmed-8784148 |
institution | National Center for Biotechnology Information |
language | English |
publishDate | 2022 |
publisher | National Academy of Sciences |
record_format | MEDLINE/PubMed |
spelling | pubmed-87841482022-02-01 Spurious normativity enhances learning of compliance and enforcement behavior in artificial agents Köster, Raphael Hadfield-Menell, Dylan Everett, Richard Weidinger, Laura Hadfield, Gillian K. Leibo, Joel Z. Proc Natl Acad Sci U S A Social Sciences How do societies learn and maintain social norms? Here we use multiagent reinforcement learning to investigate the learning dynamics of enforcement and compliance behaviors. Artificial agents populate a foraging environment and need to learn to avoid a poisonous berry. Agents learn to avoid eating poisonous berries better when doing so is taboo, meaning the behavior is punished by other agents. The taboo helps overcome a credit assignment problem in discovering delayed health effects. Critically, introducing an additional taboo, which results in punishment for eating a harmless berry, further improves overall returns. This “silly rule” counterintuitively has a positive effect because it gives agents more practice in learning rule enforcement. By probing what individual agents have learned, we demonstrate that normative behavior relies on a sequence of learned skills. Learning rule compliance builds upon prior learning of rule enforcement by other agents. Our results highlight the benefit of employing a multiagent reinforcement learning computational model focused on learning to implement complex actions. National Academy of Sciences 2022-01-12 2022-01-18 /pmc/articles/PMC8784148/ /pubmed/35022231 http://dx.doi.org/10.1073/pnas.2106028118 Text en Copyright © 2022 the Author(s). Published by PNAS. https://creativecommons.org/licenses/by-nc-nd/4.0/This open access article is distributed under Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND) (https://creativecommons.org/licenses/by-nc-nd/4.0/) . |
spellingShingle | Social Sciences Köster, Raphael Hadfield-Menell, Dylan Everett, Richard Weidinger, Laura Hadfield, Gillian K. Leibo, Joel Z. Spurious normativity enhances learning of compliance and enforcement behavior in artificial agents |
title | Spurious normativity enhances learning of compliance and enforcement behavior in artificial agents |
title_full | Spurious normativity enhances learning of compliance and enforcement behavior in artificial agents |
title_fullStr | Spurious normativity enhances learning of compliance and enforcement behavior in artificial agents |
title_full_unstemmed | Spurious normativity enhances learning of compliance and enforcement behavior in artificial agents |
title_short | Spurious normativity enhances learning of compliance and enforcement behavior in artificial agents |
title_sort | spurious normativity enhances learning of compliance and enforcement behavior in artificial agents |
topic | Social Sciences |
url | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8784148/ https://www.ncbi.nlm.nih.gov/pubmed/35022231 http://dx.doi.org/10.1073/pnas.2106028118 |
work_keys_str_mv | AT kosterraphael spuriousnormativityenhanceslearningofcomplianceandenforcementbehaviorinartificialagents AT hadfieldmenelldylan spuriousnormativityenhanceslearningofcomplianceandenforcementbehaviorinartificialagents AT everettrichard spuriousnormativityenhanceslearningofcomplianceandenforcementbehaviorinartificialagents AT weidingerlaura spuriousnormativityenhanceslearningofcomplianceandenforcementbehaviorinartificialagents AT hadfieldgilliank spuriousnormativityenhanceslearningofcomplianceandenforcementbehaviorinartificialagents AT leibojoelz spuriousnormativityenhanceslearningofcomplianceandenforcementbehaviorinartificialagents |