City Research Online

Abstract

The combination of Deep Learning and Reinforcement Learning, termed Deep Reinforcement Learning Networks (DRLN), offers the possibility of using a Deep Learning Neural Network to produce an approximate Reinforcement Learning value table that allows extraction of features from neurons in the hidden layers of the network. This paper presents a two stage technique for training a DRLN on features extracted from a DRLN trained on a identical problem, via the implementation of the Q-Learning algorithm, using TensorFlow. The results show that the extraction of features from the hidden layers of the Deep Q-Network improves the learning process of the agent (4.58 times faster and better) and proves the existence of encoded information about the environment which can be used to select the best action. The research contributes preliminary work in an ongoing research project in modeling features extracted from DRLNs.

Publication Type:	Article
Additional Information:	The final publication is available through Springer, https://doi.org/10.1007/978-3-319-92537-0_25.
Publisher Keywords:	Reinforcement Learning, Neural Networks, Deep Learning, Feature Extraction, TensorFlow
Departments:	School of Science & Technology > Computer Science

Export

Downloads