Learning to Play Donkey Kong Using Neural Networks And Reinforcement Learning

Ozkohen, P.M. and Visser, J (2017) Learning to Play Donkey Kong Using Neural Networks And Reinforcement Learning. Bachelor's Thesis, Artificial Intelligence.

Preview	Text AI_BA_2017_PAULOZKOHEN.pdf - Published Version Download (445kB) | Preview
	Text Toestemming.pdf - Other Restricted to Backend only Download (80kB)

Abstract

Neural Networks and Reinforcement Learning have successfully been applied to various games, such as Ms. Pacman and Go. This research combines Multilayer Per-ceptrons and a class of Reinforcement Learning algorithms called Actor-Critic to make an agent learn to play the arcade classic Donkey Kong game. Two neural networks are used in this study, the Actor and the Critic. The Actor neural network learns to select the best action given the game state, the Critic tries to learn the value of being in a certain state. First, a base game-playing performance is obtained by making the agent learn from demonstration data, which is obtained from humans playing the game. After the off-line training of the Actor on the demonstration data, the agent tries to further improve its base performance using feedback from the Critic. The Critic gives feedback by comparing the value of the state before and after taking the action. Results show that an actor pre-trained on demonstration data is able to achieve a good baseline performance.Applying Actor-Critic methods, however, does usually not improve performance, in many cases even decreasing it. Possible reasons include the game not fully being Markovian and other difficulties.

Item Type:	Thesis (Bachelor's Thesis)
Degree programme:	Artificial Intelligence
Thesis type:	Bachelor's Thesis
Language:	English
Date Deposited:	15 Feb 2018 08:30
Last Modified:	15 Feb 2018 08:30
URI:	https://fse.studenttheses.ub.rug.nl/id/eprint/15452

Actions (login required)

View Item