Robust population coding in free-energy-based reinforcement learning

Makoto Otsuka; Junichiro Yoshimoto; Kenji Doya

doi:10.1007/978-3-540-87536-9_39

Robust population coding in free-energy-based reinforcement learning

Makoto Otsuka, Junichiro Yoshimoto, Kenji Doya

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

1 Citation (Scopus)

Abstract

We investigate the properties of free-energy-based reinforcement learning using a new experimental platform called the digit floor task. The simulation results showed the robustness of the reinforcement learning method against noise applied in both the training and testing phases. In addition, reward-dependent and reward-invariant representations were found in the distributed activation patterns of hidden units. The representations coded in a distributed fashion persisted even when the number of hidden nodes were varied.

Original language	English
Title of host publication	Artificial Neural Networks - ICANN 2008 - 18th International Conference, Proceedings
Pages	377-386
Number of pages	10
Edition	PART 1
DOIs	https://doi.org/10.1007/978-3-540-87536-9_39
Publication status	Published - 2008
Externally published	Yes
Event	18th International Conference on Artificial Neural Networks, ICANN 2008 - Prague, Czech Republic Duration: 03-09-2008 → 06-09-2008

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Number	PART 1
Volume	5163 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	18th International Conference on Artificial Neural Networks, ICANN 2008
Country/Territory	Czech Republic
City	Prague
Period	03-09-08 → 06-09-08

All Science Journal Classification (ASJC) codes

Theoretical Computer Science
General Computer Science

Access to Document

10.1007/978-3-540-87536-9_39

Cite this

Otsuka, M., Yoshimoto, J., & Doya, K. (2008). Robust population coding in free-energy-based reinforcement learning. In Artificial Neural Networks - ICANN 2008 - 18th International Conference, Proceedings (PART 1 ed., pp. 377-386). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 5163 LNCS, No. PART 1). https://doi.org/10.1007/978-3-540-87536-9_39

Otsuka, Makoto ; Yoshimoto, Junichiro ; Doya, Kenji. / Robust population coding in free-energy-based reinforcement learning. Artificial Neural Networks - ICANN 2008 - 18th International Conference, Proceedings. PART 1. ed. 2008. pp. 377-386 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 1).

@inproceedings{92f48a547723410ebbf2ab582702b30e,

title = "Robust population coding in free-energy-based reinforcement learning",

abstract = "We investigate the properties of free-energy-based reinforcement learning using a new experimental platform called the digit floor task. The simulation results showed the robustness of the reinforcement learning method against noise applied in both the training and testing phases. In addition, reward-dependent and reward-invariant representations were found in the distributed activation patterns of hidden units. The representations coded in a distributed fashion persisted even when the number of hidden nodes were varied.",

author = "Makoto Otsuka and Junichiro Yoshimoto and Kenji Doya",

year = "2008",

doi = "10.1007/978-3-540-87536-9_39",

language = "English",

isbn = "3540875352",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

number = "PART 1",

pages = "377--386",

booktitle = "Artificial Neural Networks - ICANN 2008 - 18th International Conference, Proceedings",

edition = "PART 1",

note = "18th International Conference on Artificial Neural Networks, ICANN 2008 ; Conference date: 03-09-2008 Through 06-09-2008",

}

Otsuka, M, Yoshimoto, J & Doya, K 2008, Robust population coding in free-energy-based reinforcement learning. in Artificial Neural Networks - ICANN 2008 - 18th International Conference, Proceedings. PART 1 edn, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), no. PART 1, vol. 5163 LNCS, pp. 377-386, 18th International Conference on Artificial Neural Networks, ICANN 2008, Prague, Czech Republic, 03-09-08. https://doi.org/10.1007/978-3-540-87536-9_39

Robust population coding in free-energy-based reinforcement learning. / Otsuka, Makoto; Yoshimoto, Junichiro; Doya, Kenji.
Artificial Neural Networks - ICANN 2008 - 18th International Conference, Proceedings. PART 1. ed. 2008. p. 377-386 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 5163 LNCS, No. PART 1).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - Robust population coding in free-energy-based reinforcement learning

AU - Otsuka, Makoto

AU - Yoshimoto, Junichiro

AU - Doya, Kenji

PY - 2008

Y1 - 2008

N2 - We investigate the properties of free-energy-based reinforcement learning using a new experimental platform called the digit floor task. The simulation results showed the robustness of the reinforcement learning method against noise applied in both the training and testing phases. In addition, reward-dependent and reward-invariant representations were found in the distributed activation patterns of hidden units. The representations coded in a distributed fashion persisted even when the number of hidden nodes were varied.

AB - We investigate the properties of free-energy-based reinforcement learning using a new experimental platform called the digit floor task. The simulation results showed the robustness of the reinforcement learning method against noise applied in both the training and testing phases. In addition, reward-dependent and reward-invariant representations were found in the distributed activation patterns of hidden units. The representations coded in a distributed fashion persisted even when the number of hidden nodes were varied.

UR - http://www.scopus.com/inward/record.url?scp=58849087267&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=58849087267&partnerID=8YFLogxK

U2 - 10.1007/978-3-540-87536-9_39

DO - 10.1007/978-3-540-87536-9_39

M3 - Conference contribution

AN - SCOPUS:58849087267

SN - 3540875352

SN - 9783540875352

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 377

EP - 386

BT - Artificial Neural Networks - ICANN 2008 - 18th International Conference, Proceedings

T2 - 18th International Conference on Artificial Neural Networks, ICANN 2008

Y2 - 3 September 2008 through 6 September 2008

ER -

Otsuka M, Yoshimoto J, Doya K. Robust population coding in free-energy-based reinforcement learning. In Artificial Neural Networks - ICANN 2008 - 18th International Conference, Proceedings. PART 1 ed. 2008. p. 377-386. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 1). doi: 10.1007/978-3-540-87536-9_39

Robust population coding in free-energy-based reinforcement learning

Abstract

Publication series

Conference

All Science Journal Classification (ASJC) codes

Access to Document

Other files and links

Fingerprint

Cite this