Guardado en:
Detalles Bibliográficos
Autores principales: Le, Xuan-Bach, Wagner, Dominik, Witzman, Leon, Rabinovich, Alexander, Ong, Luke
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:https://arxiv.org/abs/2410.12175
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866929545099083776
author Le, Xuan-Bach
Wagner, Dominik
Witzman, Leon
Rabinovich, Alexander
Ong, Luke
author_facet Le, Xuan-Bach
Wagner, Dominik
Witzman, Leon
Rabinovich, Alexander
Ong, Luke
contents Linear temporal logic (LTL) and, more generally, $ω$-regular objectives are alternatives to the traditional discount sum and average reward objectives in reinforcement learning (RL), offering the advantage of greater comprehensibility and hence explainability. In this work, we study the relationship between these objectives. Our main result is that each RL problem for $ω$-regular objectives can be reduced to a limit-average reward problem in an optimality-preserving fashion, via (finite-memory) reward machines. Furthermore, we demonstrate the efficacy of this approach by showing that optimal policies for limit-average problems can be found asymptotically by solving a sequence of discount-sum problems approximately. Consequently, we resolve an open problem: optimal policies for LTL and $ω$-regular objectives can be learned asymptotically.
format Preprint
id arxiv_https___arxiv_org_abs_2410_12175
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Reinforcement Learning with LTL and $ω$-Regular Objectives via Optimality-Preserving Translation to Average Rewards
Le, Xuan-Bach
Wagner, Dominik
Witzman, Leon
Rabinovich, Alexander
Ong, Luke
Machine Learning
Artificial Intelligence
Linear temporal logic (LTL) and, more generally, $ω$-regular objectives are alternatives to the traditional discount sum and average reward objectives in reinforcement learning (RL), offering the advantage of greater comprehensibility and hence explainability. In this work, we study the relationship between these objectives. Our main result is that each RL problem for $ω$-regular objectives can be reduced to a limit-average reward problem in an optimality-preserving fashion, via (finite-memory) reward machines. Furthermore, we demonstrate the efficacy of this approach by showing that optimal policies for limit-average problems can be found asymptotically by solving a sequence of discount-sum problems approximately. Consequently, we resolve an open problem: optimal policies for LTL and $ω$-regular objectives can be learned asymptotically.
title Reinforcement Learning with LTL and $ω$-Regular Objectives via Optimality-Preserving Translation to Average Rewards
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2410.12175