Using reinforcement learning to guide the development of self-organised feature maps for visual orienting

Kevin Brohan; Kevin Gurney; Piotr Dudek

doi:10.1007/978-3-642-15822-3_23

Using reinforcement learning to guide the development of self-organised feature maps for visual orienting

Kevin Brohan, Kevin Gurney, Piotr Dudek

Research output: Chapter in Book/Conference proceeding › Conference contribution › peer-review

Abstract

We present a biologically inspired neural network model of visual orienting (using saccadic eye movements) in which targets are preferentially selected according to their reward value. Internal representations of visual features that guide saccades are developed in a self-organised map whose plasticity is modulated under reward. In this way, only those features relevant for acquiring rewarding targets are generated. As well as guiding the formation of feature representations, rewarding stimuli are stored in a working memory and bias future saccade generation. In addition, a reward prediction error is used to initiate retraining of the self-organised map to generate more efficient representations of the features when necessary. © 2010 Springer-Verlag Berlin Heidelberg.

Original language	English
Title of host publication	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)\|Lect. Notes Comput. Sci.
Publisher	Springer Nature
Pages	180-189
Number of pages	9
Volume	6353
ISBN (Print)	3642158218, 9783642158216
DOIs	https://doi.org/10.1007/978-3-642-15822-3_23
Publication status	Published - 2010
Event	20th International Conference on Artificial Neural Networks, ICANN 2010 - Thessaloniki Duration: 1 Jul 2010 → … http://dl.acm.org/citation.cfm?id=1889001.1889028

Conference

Conference	20th International Conference on Artificial Neural Networks, ICANN 2010
City	Thessaloniki
Period	1/07/10 → …
Internet address	http://dl.acm.org/citation.cfm?id=1889001.1889028

Access to Document

10.1007/978-3-642-15822-3_23

Cite this

Brohan, K., Gurney, K., & Dudek, P. (2010). Using reinforcement learning to guide the development of self-organised feature maps for visual orienting. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)|Lect. Notes Comput. Sci. (Vol. 6353, pp. 180-189). Springer Nature. https://doi.org/10.1007/978-3-642-15822-3_23

@inproceedings{edc41706ff574381b7b93fe1403cd4ec,

title = "Using reinforcement learning to guide the development of self-organised feature maps for visual orienting",

abstract = "We present a biologically inspired neural network model of visual orienting (using saccadic eye movements) in which targets are preferentially selected according to their reward value. Internal representations of visual features that guide saccades are developed in a self-organised map whose plasticity is modulated under reward. In this way, only those features relevant for acquiring rewarding targets are generated. As well as guiding the formation of feature representations, rewarding stimuli are stored in a working memory and bias future saccade generation. In addition, a reward prediction error is used to initiate retraining of the self-organised map to generate more efficient representations of the features when necessary. {\textcopyright} 2010 Springer-Verlag Berlin Heidelberg.",

author = "Kevin Brohan and Kevin Gurney and Piotr Dudek",

year = "2010",

doi = "10.1007/978-3-642-15822-3_23",

language = "English",

isbn = "3642158218",

volume = "6353",

pages = "180--189",

booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)|Lect. Notes Comput. Sci.",

publisher = "Springer Nature",

address = "United States",

note = "20th International Conference on Artificial Neural Networks, ICANN 2010 ; Conference date: 01-07-2010",

url = "http://dl.acm.org/citation.cfm?id=1889001.1889028",

}

Brohan, K, Gurney, K & Dudek, P 2010, Using reinforcement learning to guide the development of self-organised feature maps for visual orienting. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)|Lect. Notes Comput. Sci.. vol. 6353, Springer Nature, pp. 180-189, 20th International Conference on Artificial Neural Networks, ICANN 2010, Thessaloniki, 1/07/10. https://doi.org/10.1007/978-3-642-15822-3_23

Using reinforcement learning to guide the development of self-organised feature maps for visual orienting. / Brohan, Kevin; Gurney, Kevin; Dudek, Piotr.
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)|Lect. Notes Comput. Sci.. Vol. 6353 Springer Nature, 2010. p. 180-189.

Research output: Chapter in Book/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Using reinforcement learning to guide the development of self-organised feature maps for visual orienting

AU - Brohan, Kevin

AU - Gurney, Kevin

AU - Dudek, Piotr

PY - 2010

Y1 - 2010

N2 - We present a biologically inspired neural network model of visual orienting (using saccadic eye movements) in which targets are preferentially selected according to their reward value. Internal representations of visual features that guide saccades are developed in a self-organised map whose plasticity is modulated under reward. In this way, only those features relevant for acquiring rewarding targets are generated. As well as guiding the formation of feature representations, rewarding stimuli are stored in a working memory and bias future saccade generation. In addition, a reward prediction error is used to initiate retraining of the self-organised map to generate more efficient representations of the features when necessary. © 2010 Springer-Verlag Berlin Heidelberg.

AB - We present a biologically inspired neural network model of visual orienting (using saccadic eye movements) in which targets are preferentially selected according to their reward value. Internal representations of visual features that guide saccades are developed in a self-organised map whose plasticity is modulated under reward. In this way, only those features relevant for acquiring rewarding targets are generated. As well as guiding the formation of feature representations, rewarding stimuli are stored in a working memory and bias future saccade generation. In addition, a reward prediction error is used to initiate retraining of the self-organised map to generate more efficient representations of the features when necessary. © 2010 Springer-Verlag Berlin Heidelberg.

U2 - 10.1007/978-3-642-15822-3_23

DO - 10.1007/978-3-642-15822-3_23

M3 - Conference contribution

SN - 3642158218

SN - 9783642158216

VL - 6353

SP - 180

EP - 189

BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)|Lect. Notes Comput. Sci.

PB - Springer Nature

T2 - 20th International Conference on Artificial Neural Networks, ICANN 2010

Y2 - 1 July 2010

ER -

Brohan K, Gurney K, Dudek P. Using reinforcement learning to guide the development of self-organised feature maps for visual orienting. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)|Lect. Notes Comput. Sci.. Vol. 6353. Springer Nature. 2010. p. 180-189 doi: 10.1007/978-3-642-15822-3_23

Using reinforcement learning to guide the development of self-organised feature maps for visual orienting

Abstract

Conference

Access to Document

Fingerprint

Cite this