Efficient Models Deep Reinforcement Learning for NetHack Strategies

Yasuhiro Onuki, Yasuyuki Tahara, Akihiko Ohsuga, Yuichi Sei

2025

Abstract

Deep reinforcement learning (DRL) has been widely used in agent research across various video games, demonstrating its effectiveness. Recently, there has been increasing interest in DRL research in complex environments such as Roguelike games. These games, while complex, offer fast execution speeds, making them useful as a testbeds for DRL agents. Among them, the game NetHack has gained of research attention. In this study, we aim to train a DRL agent for efficient learning with reduced training costs using the NetHack Learning Environment (NLE). We propose a method that incorporates a variational autoencoder (VAE). Additionally, since the rewards provided by the NLE are sparse, which complicates training, we also trained a DRL agent with additional rewards. As a result, although we expected that using the VAE would allow for more advantageous progress in the game, contrary to our expectations, it proves ineffective. Conversely, we find that the additional rewards are effective.

Paper Citation

in Harvard Style

Onuki Y., Tahara Y., Ohsuga A. and Sei Y. (2025). Efficient Models Deep Reinforcement Learning for NetHack Strategies. In Proceedings of the 17th International Conference on Agents and Artificial Intelligence - Volume 1: ICAART; ISBN 978-989-758-737-5, SciTePress, pages 556-563. DOI: 10.5220/0013253100003890

in Bibtex Style

@conference{icaart25,
author={Yasuhiro Onuki and Yasuyuki Tahara and Akihiko Ohsuga and Yuichi Sei},
title={Efficient Models Deep Reinforcement Learning for NetHack Strategies},
booktitle={Proceedings of the 17th International Conference on Agents and Artificial Intelligence - Volume 1: ICAART},
year={2025},
pages={556-563},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013253100003890},
isbn={978-989-758-737-5},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 17th International Conference on Agents and Artificial Intelligence - Volume 1: ICAART
TI - Efficient Models Deep Reinforcement Learning for NetHack Strategies
SN - 978-989-758-737-5
AU - Onuki Y.
AU - Tahara Y.
AU - Ohsuga A.
AU - Sei Y.
PY - 2025
SP - 556
EP - 563
DO - 10.5220/0013253100003890
PB - SciTePress