Analyzing Exact Output Regions of Reinforcement Learning Policy Neural Networks for High-Dimensional Input-Output Spaces

Torben Logemann, Eric Veith

2024

Abstract

Agent systems based on deep reinforcement learning have achieved remarkable success in recent years. They have also been applied to a variety of research topics in the field of power grids, as such agents promise real resilience. However, deep reinforcement learning agents cannot guarantee behavior, as the mapping of the entire input space to the output of even a simple feed-forward neural network cannot be accurately explained. For critical infrastructures, such black box models are not acceptable. To ensure an optimized trade-off between learning performance and explainability, this paper relies on efficient regularizable feed-forward neural networks and presents an extension of the algorithm NN2EQCDT to transform the networks into pruned decision trees with significantly fewer nodes to be accurately explained. In this paper, we present a methodological approach to further analyze the decision trees for high-dimensional input-output spaces and analyze an agent for a power grid experiment.

Paper Citation

in Harvard Style

Logemann T. and Veith E. (2024). Analyzing Exact Output Regions of Reinforcement Learning Policy Neural Networks for High-Dimensional Input-Output Spaces. In Proceedings of the 1st International Conference on Explainable AI for Neural and Symbolic Methods - Volume 1: EXPLAINS; ISBN 978-989-758-720-7, SciTePress, pages 96-107. DOI: 10.5220/0012928000003886

in Bibtex Style

@conference{explains24,
author={Torben Logemann and Eric Veith},
title={Analyzing Exact Output Regions of Reinforcement Learning Policy Neural Networks for High-Dimensional Input-Output Spaces},
booktitle={Proceedings of the 1st International Conference on Explainable AI for Neural and Symbolic Methods - Volume 1: EXPLAINS},
year={2024},
pages={96-107},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0012928000003886},
isbn={978-989-758-720-7},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 1st International Conference on Explainable AI for Neural and Symbolic Methods - Volume 1: EXPLAINS
TI - Analyzing Exact Output Regions of Reinforcement Learning Policy Neural Networks for High-Dimensional Input-Output Spaces
SN - 978-989-758-720-7
AU - Logemann T.
AU - Veith E.
PY - 2024
SP - 96
EP - 107
DO - 10.5220/0012928000003886
PB - SciTePress