Safety and optimality in learning-based control at low computational cost

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Baumann, Dominik, Kowalczyk, Krzysztof, Rojas, Cristian R., Tiels, Koen, Wachel, Pawel
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866910940909273088
author Baumann, Dominik
Kowalczyk, Krzysztof
Rojas, Cristian R.
Tiels, Koen
Wachel, Pawel
author_facet Baumann, Dominik
Kowalczyk, Krzysztof
Rojas, Cristian R.
Tiels, Koen
Wachel, Pawel
contents Applying machine learning methods to physical systems that are supposed to act in the real world requires providing safety guarantees. However, methods that include such guarantees often come at a high computational cost, making them inapplicable to large datasets and embedded devices with low computational power. In this paper, we propose CoLSafe, a computationally lightweight safe learning algorithm whose computational complexity grows sublinearly with the number of data points. We derive both safety and optimality guarantees and showcase the effectiveness of our algorithm on a seven-degrees-of-freedom robot arm.
format Preprint
id arxiv_https___arxiv_org_abs_2505_08026
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Safety and optimality in learning-based control at low computational cost
Baumann, Dominik
Kowalczyk, Krzysztof
Rojas, Cristian R.
Tiels, Koen
Wachel, Pawel
Systems and Control
Machine Learning
Applying machine learning methods to physical systems that are supposed to act in the real world requires providing safety guarantees. However, methods that include such guarantees often come at a high computational cost, making them inapplicable to large datasets and embedded devices with low computational power. In this paper, we propose CoLSafe, a computationally lightweight safe learning algorithm whose computational complexity grows sublinearly with the number of data points. We derive both safety and optimality guarantees and showcase the effectiveness of our algorithm on a seven-degrees-of-freedom robot arm.
title Safety and optimality in learning-based control at low computational cost
topic Systems and Control
Machine Learning
url https://arxiv.org/abs/2505.08026