Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • harvard1
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Towards a framework for reinforcement learning with artificial neural networks
KTH, Superseded Departments, Numerical Analysis and Computer Science, NADA.
2004 (English)Licentiate thesis, comprehensive summary (Other scientific)
Place, publisher, year, edition, pages
Stockholm: Numerisk analys och datalogi , 2004.
Series
Trita-NA, ISSN 0348-2952 ; 0437
Keyword [en]
aritficial neural networks, reinforcement learning, BCPNN, modularity, genetic algorithms
National Category
Computer Sciences
Identifiers
URN: urn:nbn:se:kth:diva-500ISBN: 91-7283-928-7 (print)OAI: oai:DiVA.org:kth-500DiVA: diva2:14236
Presentation
2004-12-10, E32, KTH, Lindstedsvägen 3, Stockholm, 13:00
Available from: 2005-11-23 Created: 2005-11-23 Last updated: 2018-01-13
List of papers
1. Reinforcement Learning Based on a Bayesian Confidence Propagating Neural Network
Open this publication in new window or tab >>Reinforcement Learning Based on a Bayesian Confidence Propagating Neural Network
2003 (English)Conference paper, Published paper (Refereed)
Abstract [en]

We present a system capable of reinforcement learning (RL) based on the Bayesian confidence propagating neural network (BCPNN). The system is called BCPNNRL and its architecture is somewhat motivated by parallels to biology. We analyze the systems properties and we benchmark it against a simple Monte Carlo (MC) based RL algorithm, pursuit RL methods, and the Associative Reward Penalty (AR-P) algorithm. The system is used to solve the n-armed bandit problem, pattern association, and path finding in a maze.

National Category
Computer Sciences
Identifiers
urn:nbn:se:kth:diva-8797 (URN)
Conference
2003, April 10-11, SAIS-SSLS Joint Workshop, Center for Applied Autonomous Sensor Systems, Örebro, Sweden
Available from: 2005-11-23 Created: 2005-11-23 Last updated: 2018-01-13Bibliographically approved
2. Reinforcement learning in a noisy fine grid environment
Open this publication in new window or tab >>Reinforcement learning in a noisy fine grid environment
2004 (English)Report (Refereed)
Series
Trita-NA, ISSN 0348-2952 ; 0434
National Category
Computer Sciences
Identifiers
urn:nbn:se:kth:diva-8798 (URN)
Available from: 2005-11-23 Created: 2005-11-23 Last updated: 2018-01-13Bibliographically approved
3. Modular neural networks and reinforcement learning
Open this publication in new window or tab >>Modular neural networks and reinforcement learning
2004 (English)Report (Other academic)
Abstract [en]

We investigate the effect of modular architecture in an artificial neural network for a reinforcement learning problem. Using the supervised backpropagation algorithm to solve a two-task problem, the network performance can be increased by using networks with modular structures. However, using a modular architecture to solve a two-task reinforcement learning problem will not increase the performance compared to a non-modular structure. We show that by combining a modular structure with a modular reward signal the network learns significantly faster.

Place, publisher, year, edition, pages
Stockholm: KTH Royal Institute of Technology, 2004. 6 p.
Series
Trita-NA, ISSN 0348-2952 ; 0434
National Category
Computer Sciences
Identifiers
urn:nbn:se:kth:diva-8799 (URN)
Note
QC 20111216Available from: 2005-11-23 Created: 2005-11-23 Last updated: 2018-01-13Bibliographically approved
4. Paralled reinforcement learning using multiple reward signals
Open this publication in new window or tab >>Paralled reinforcement learning using multiple reward signals
2006 (English)In: Neurocomputing, ISSN 0925-2312, E-ISSN 1872-8286, Vol. 69, no 16/18, 2171-2179 p.Article in journal (Refereed) Published
Abstract [en]

Reinforcement learning is a quite general learning paradigm that can be used to solve a large set of problems. For complex problems it has been shown that by using task decomposition it may be possible for the system to learn faster. One common approach is to construct systems with multiple modules, where each module learns a sub-task. We present a parallel learning method for agents with an actor–critic architecture based on artificial neural networks. The agents have multiple modules, where the modules can learn in parallel to further increase learning speed. Each module solves a sub-problem and receives its own separate reward signal with all modules trained concurrently. We use the method on a grid world navigation task and show that parallel learning can significantly reduce learning time.

Keyword
Neural networks, Modularity, Reinforcement learning
National Category
Computer Sciences
Identifiers
urn:nbn:se:kth:diva-8800 (URN)10.1016/j.neucom.2005.07.008 (DOI)000241044700033 ()2-s2.0-33748416863 (Scopus ID)
Note

Updated from submitted to published, 20111216. QC 20111216

Available from: 2005-11-23 Created: 2005-11-23 Last updated: 2018-01-13Bibliographically approved
5. Evolving multiple reward signals
Open this publication in new window or tab >>Evolving multiple reward signals
2004 (English)In: Trita-NA, no 35Article in journal (Refereed) Published
Place, publisher, year, edition, pages
KTH, 2004
National Category
Computer and Information Sciences
Identifiers
urn:nbn:se:kth:diva-8801 (URN)
Note
QC 20111007Available from: 2005-11-23 Created: 2005-11-23 Last updated: 2018-01-13Bibliographically approved

Open Access in DiVA

No full text

Search in DiVA

By author/editor
Raicevic, Peter
By organisation
Numerical Analysis and Computer Science, NADA
Computer Sciences

Search outside of DiVA

GoogleGoogle Scholar

isbn
urn-nbn

Altmetric score

isbn
urn-nbn
Total: 310 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • harvard1
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf