Bootstrapping a DQN Replay Memory with Synthetic Experiences

Wenzel Baron Pilar von Pilchau, Anthony Stein, Jörg Hähner

2020

Abstract

An important component of many Deep Reinforcement Learning algorithms is the Experience Replay that serves as a storage mechanism or memory of experienced transitions. These experiences are used for training and help the agent to stably find the perfect trajectory through the problem space. The classic Experience Replay however makes only use of the experiences it actually made, but the stored transitions bear great potential in form of knowledge about the problem that can be extracted. The gathered knowledge contains state-transitions and received rewards that can be utilized to approximate a model of the environment. We present an algorithm that creates synthetic experiences in a nondeterministic discrete environment to assist the learner with augmented training data. The Interpolated Experience Replay is evaluated on the FrozenLake environment and we show that it can achieve a 17% increased mean reward compared to the classic version.

Paper Citation

in Harvard Style

von Pilchau W., Stein A. and Hähner J. (2020). Bootstrapping a DQN Replay Memory with Synthetic Experiences. In Proceedings of the 12th International Joint Conference on Computational Intelligence (IJCCI 2020) - Volume 1: NCTA; ISBN 978-989-758-475-6, SciTePress, pages 404-411. DOI: 10.5220/0010107904040411

in Bibtex Style

@conference{ncta20,
author={Wenzel Baron Pilar von Pilchau and Anthony Stein and Jörg Hähner},
title={Bootstrapping a DQN Replay Memory with Synthetic Experiences},
booktitle={Proceedings of the 12th International Joint Conference on Computational Intelligence (IJCCI 2020) - Volume 1: NCTA},
year={2020},
pages={404-411},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0010107904040411},
isbn={978-989-758-475-6},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 12th International Joint Conference on Computational Intelligence (IJCCI 2020) - Volume 1: NCTA
TI - Bootstrapping a DQN Replay Memory with Synthetic Experiences
SN - 978-989-758-475-6
AU - von Pilchau W.
AU - Stein A.
AU - Hähner J.
PY - 2020
SP - 404
EP - 411
DO - 10.5220/0010107904040411
PB - SciTePress