Evolutionary algorithms for optimising reinforcement learning policy approximation

dc.contributor.advisorBassett, Bruce
dc.contributor.authorCuningham, Blake
dc.date.accessioned2020-02-19T12:18:14Z
dc.date.available2020-02-19T12:18:14Z
dc.date.issued2019
dc.date.updated2020-02-19T12:17:41Z
dc.description.abstractReinforcement learning methods have become more efficient in recent years. In particular, the A3C (asynchronous advantage actor critic) approach demonstrated in Mnih et al. (2016) was able to halve the training time of the existing state-of-the-art approaches. However, these methods still require relatively large amounts of training resources due to the fundamental exploratory nature of reinforcement learning. Other machine learning approaches are able to improve the ability to train reinforcement learning agents by better processing input information to help map states to actions - convolutional and recurrent neural networks are helpful when input data is in image form that does not satisfy the Markov property. The specific required architecture of these convolutional and recurrent neural network models is not obvious given infinite possible permutations. There is very limited research giving clear guidance on neural network structure in a RL (reinforcement learning) context, and grid search-like approaches require too many resources and do not always find good optima. In order to address these, and other, challenges associated with traditional parameter optimization methods, an evolutionary approach similar to that taken by Dufourq and Bassett (2017) for image classification tasks was used to find the optimal model architecture when training an agent that learns to play Atari Pong. The approach found models that were able to train reinforcement learning agents faster, and with fewer parameters than that found by OpenAI’s model in Blackwell et al. (2018) - a superhuman level of performance.
dc.identifier.apacitationCuningham, B. (2019). <i>Evolutionary algorithms for optimising reinforcement learning policy approximation</i>. (). ,Faculty of Science ,Department of Statistical Sciences. Retrieved from http://hdl.handle.net/11427/31170en_ZA
dc.identifier.chicagocitationCuningham, Blake. <i>"Evolutionary algorithms for optimising reinforcement learning policy approximation."</i> ., ,Faculty of Science ,Department of Statistical Sciences, 2019. http://hdl.handle.net/11427/31170en_ZA
dc.identifier.citationCuningham, B. 2019. Evolutionary algorithms for optimising reinforcement learning policy approximation.en_ZA
dc.identifier.ris TY - Thesis / Dissertation AU - Cuningham, Blake AB - Reinforcement learning methods have become more efficient in recent years. In particular, the A3C (asynchronous advantage actor critic) approach demonstrated in Mnih et al. (2016) was able to halve the training time of the existing state-of-the-art approaches. However, these methods still require relatively large amounts of training resources due to the fundamental exploratory nature of reinforcement learning. Other machine learning approaches are able to improve the ability to train reinforcement learning agents by better processing input information to help map states to actions - convolutional and recurrent neural networks are helpful when input data is in image form that does not satisfy the Markov property. The specific required architecture of these convolutional and recurrent neural network models is not obvious given infinite possible permutations. There is very limited research giving clear guidance on neural network structure in a RL (reinforcement learning) context, and grid search-like approaches require too many resources and do not always find good optima. In order to address these, and other, challenges associated with traditional parameter optimization methods, an evolutionary approach similar to that taken by Dufourq and Bassett (2017) for image classification tasks was used to find the optimal model architecture when training an agent that learns to play Atari Pong. The approach found models that were able to train reinforcement learning agents faster, and with fewer parameters than that found by OpenAI’s model in Blackwell et al. (2018) - a superhuman level of performance. DA - 2019 DB - OpenUCT DP - University of Cape Town KW - statistical sciences LK - https://open.uct.ac.za PY - 2019 T1 - Evolutionary algorithms for optimising reinforcement learning policy approximation TI - Evolutionary algorithms for optimising reinforcement learning policy approximation UR - http://hdl.handle.net/11427/31170 ER - en_ZA
dc.identifier.urihttp://hdl.handle.net/11427/31170
dc.identifier.vancouvercitationCuningham B. Evolutionary algorithms for optimising reinforcement learning policy approximation. []. ,Faculty of Science ,Department of Statistical Sciences, 2019 [cited yyyy month dd]. Available from: http://hdl.handle.net/11427/31170en_ZA
dc.language.rfc3066eng
dc.publisher.departmentDepartment of Statistical Sciences
dc.publisher.facultyFaculty of Science
dc.subjectstatistical sciences
dc.titleEvolutionary algorithms for optimising reinforcement learning policy approximation
dc.typeMaster Thesis
dc.type.qualificationlevelMasters
dc.type.qualificationnameMSc
Files
Original bundle
Now showing 1 - 1 of 1
Loading...
Thumbnail Image
Name:
thesis_sci_2019_cuningham_blake.pdf
Size:
7.19 MB
Format:
Adobe Portable Document Format
Description:
License bundle
Now showing 1 - 1 of 1
Loading...
Thumbnail Image
Name:
license.txt
Size:
0 B
Format:
Item-specific license agreed upon to submission
Description:
Collections