Publication:
Learning to Avoid Risky Actions

Simple item page
dc.affiliation.dptoUC3M. Departamento de Ingeniería de Sistemas y Automáticaes
dc.affiliation.grupoinvUC3M. Grupo de Investigación: Laboratorio de Robótica (Robotics Lab)es
dc.contributor.authorMalfaz Vázquez, María Ángeles
dc.contributor.authorSalichs Sánchez-Caballero, Miguel
dc.date.accessioned2014-03-27T12:33:43Z
dc.date.available2014-03-27T12:33:43Z
dc.date.issued2011-12
dc.description.abstractWhen a reinforcement learning agent executes actions that can cause frequent damage to itself, it can learn, by using Q-learning, that these actions must not be executed again. However, there are other actions that do not cause damage frequently but only once in a while, for example, risky actions such as parachuting. These actions may imply punishment to the agent and, depending on its personality, it would be better to avoid them. Nevertheless, using the standard Q-learning algorithm, the agent is not able to learn to avoid them, because the result of these actions can be positive on average. In this article, an additional mechanism of Q-learning, inspired by the emotion of fear, is introduced in order to deal with those risky actions by considering the worst results. Moreover, there is a daring factor for adjusting the consideration of the risk. This mechanism is implemented on an autonomous agent living in a virtual environment. The results present the performance of the agent with different daring degrees.en
dc.description.sponsorshipThe funds provided by the Spanish Government through the project called “A New Approach to Social Robotics” (AROS), of MICINN (Ministry of Science and Innovation) and through the RoboCity2030-IICM project (S2009/DPI-1559), funded by Programas de Actividades I+D en la Comunidad de Madrid and cofunded by Structural Funds of the EU.en
dc.format.extent22
dc.format.mimetypeapplication/pdf
dc.identifier.bibliographicCitationCybernetics and Systems: An International Journal, 2011, vol. 42 (8), pp. 636-658en
dc.identifier.doi10.1080/01969722.2011.634681
dc.identifier.issn0196-9722 (print)
dc.identifier.issn1087-6553 (online)
dc.identifier.publicationfirstpage636
dc.identifier.publicationissue8
dc.identifier.publicationlastpage658
dc.identifier.publicationtitleCybernetics and Systems: An International Journalen
dc.identifier.publicationvolume42
dc.identifier.urihttps://hdl.handle.net/10016/18621
dc.identifier.uxxiAR/0000009819
dc.language.isoeng
dc.publisherTaylor & Francis Groupen
dc.relation.projectIDComunidad de Madrid. S2009/DPI-1559/ROBOCITY2030 IIes
dc.relation.publisherversionhttp://dx.doi.org/10.1080/01969722.2011.634681
dc.rights.accessRightsopen access
dc.subject.ecienciaRobótica e Informática Industriales
dc.subject.otherAutonomous agenten
dc.subject.otherDecision making systemen
dc.subject.otherFearen
dc.subject.otherReinforcement learningen
dc.subject.otherRisky actionsen
dc.titleLearning to Avoid Risky Actionsen
dc.typeresearch article*
dc.type.hasVersionAM*
dspace.entity.typePublication
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
learning_CSIJ_2011_ps.pdf
Size:
666.41 KB
Format:
Adobe Portable Document Format
Download
Collections
DISA - LR - Artículos de Revistas