Evolving Internal Reinforcers for an Intrinsically Motivated Reinforcement-Learning Robot
| Venue: | 6TH IEEE INTERNATIONAL CONFERENCE ON DEVELOPMENT AND LEARNING (ICDL2007) (SUBMITTED |
| Citations: | 15 - 2 self |
BibTeX
@INPROCEEDINGS{Schembri_evolvinginternal,
author = {Massimiliano Schembri and Marco Mirolli and Gianluca Baldassarre},
title = { Evolving Internal Reinforcers for an Intrinsically Motivated Reinforcement-Learning Robot},
booktitle = {6TH IEEE INTERNATIONAL CONFERENCE ON DEVELOPMENT AND LEARNING (ICDL2007) (SUBMITTED},
year = {},
publisher = {}
}
Bookmark
 |
 |
 |
 |
 |
 |
OpenURL
Abstract
... (IMRL) has been proposed as a framework within which agents exploit “internal reinforcement” to acquire general-purpose building-block behaviors (“skills”) which can be later combined for solving several specific tasks. The architectures so far proposed within this framework are limited in that: (1) they use hardwired “salient events” to form and train skills, and this limits agents’ autonomy; (2) they are applicable only to problems with abstract states and actions, as grid-world problems. This paper proposes solutions to these problems in the form of a hierarchical reinforcement-learning architecture that: (1) exploits Evolutionary Robotics techniques so to allow the system to autonomously discover “salient events”; (2) uses neural networks so to allow the system to cope with continuous states and noisy environments. The viability of the proposed approach is demonstrated with a simulated robotic scenario.
