Bounded Exploration with World Model Uncertainty in Soft Actor-Critic Reinforcement Learning Algorithm

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Qiao, Ting, Williams, Henry, Valencia, David, MacDonald, Bruce
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866929620473872384
author Qiao, Ting
Williams, Henry
Valencia, David
MacDonald, Bruce
author_facet Qiao, Ting
Williams, Henry
Valencia, David
MacDonald, Bruce
contents One of the bottlenecks preventing Deep Reinforcement Learning algorithms (DRL) from real-world applications is how to explore the environment and collect informative transitions efficiently. The present paper describes bounded exploration, a novel exploration method that integrates both 'soft' and intrinsic motivation exploration. Bounded exploration notably improved the Soft Actor-Critic algorithm's performance and its model-based extension's converging speed. It achieved the highest score in 6 out of 8 experiments. Bounded exploration presents an alternative method to introduce intrinsic motivations to exploration when the original reward function has strict meanings.
format Preprint
id arxiv_https___arxiv_org_abs_2412_06139
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Bounded Exploration with World Model Uncertainty in Soft Actor-Critic Reinforcement Learning Algorithm
Qiao, Ting
Williams, Henry
Valencia, David
MacDonald, Bruce
Machine Learning
Systems and Control
One of the bottlenecks preventing Deep Reinforcement Learning algorithms (DRL) from real-world applications is how to explore the environment and collect informative transitions efficiently. The present paper describes bounded exploration, a novel exploration method that integrates both 'soft' and intrinsic motivation exploration. Bounded exploration notably improved the Soft Actor-Critic algorithm's performance and its model-based extension's converging speed. It achieved the highest score in 6 out of 8 experiments. Bounded exploration presents an alternative method to introduce intrinsic motivations to exploration when the original reward function has strict meanings.
title Bounded Exploration with World Model Uncertainty in Soft Actor-Critic Reinforcement Learning Algorithm
topic Machine Learning
Systems and Control
url https://arxiv.org/abs/2412.06139