Reinforcement Learning on Slow Features of High-Dimensional Input Streams

Robert Legenstein, Niko Wilbert and Laurenz Wiskott

PLOS Computational Biology, 2010, vol. 6, issue 8, 1-13

Abstract: Humans and animals are able to learn complex behaviors based on a massive stream of sensory information from different modalities. Early animal studies have identified learning mechanisms that are based on reward and punishment such that animals tend to avoid actions that lead to punishment whereas rewarded actions are reinforced. However, most algorithms for reward-based learning are only applicable if the dimensionality of the state-space is sufficiently small or its structure is sufficiently simple. Therefore, the question arises how the problem of learning on high-dimensional data is solved in the brain. In this article, we propose a biologically plausible generic two-stage learning system that can directly be applied to raw high-dimensional input streams. The system is composed of a hierarchical slow feature analysis (SFA) network for preprocessing and a simple neural network on top that is trained based on rewards. We demonstrate by computer simulations that this generic architecture is able to learn quite demanding reinforcement learning tasks on high-dimensional visual input streams in a time that is comparable to the time needed when an explicit highly informative low-dimensional state-space representation is given instead of the high-dimensional visual input. The learning speed of the proposed architecture in a task similar to the Morris water maze task is comparable to that found in experimental studies with rats. This study thus supports the hypothesis that slowness learning is one important unsupervised learning principle utilized in the brain to form efficient state representations for behavioral learning.Author Summary: Humans and animals are able to learn complex behaviors based on a massive stream of sensory information from different modalities. Early animal studies have identified learning mechanisms that are based on reward and punishment such that animals tend to avoid actions that lead to punishment whereas rewarded actions are reinforced. It is an open question how sensory information is processed by the brain in order to learn and perform rewarding behaviors. In this article, we propose a learning system that combines the autonomous extraction of important information from the sensory input with reward-based learning. The extraction of salient information is learned by exploiting the temporal continuity of real-world stimuli. A subsequent neural circuit then learns rewarding behaviors based on this representation of the sensory input. We demonstrate in two control tasks that this system is capable of learning complex behaviors on raw visual input.

Date: 2010
References: View references in EconPapers View complete reference list from CitEc
Citations: View citations in EconPapers (3)

Downloads: (external link)
https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1000894 (text/html)
https://journals.plos.org/ploscompbiol/article/fil ... 00894&type=printable (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:plo:pcbi00:1000894

DOI: 10.1371/journal.pcbi.1000894

Access Statistics for this article

More articles in PLOS Computational Biology from Public Library of Science
Bibliographic data for series maintained by ploscompbiol ().