What, Indeed, is an Achievable Provable Guarantee for Learning-Enabled Safety Critical Systems

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Bensalem, Saddek, Cheng, Chih-Hong, Huang, Wei, Huang, Xiaowei, Wu, Changshun, Zhao, Xingyu
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908056298717184
author Bensalem, Saddek
Cheng, Chih-Hong
Huang, Wei
Huang, Xiaowei
Wu, Changshun
Zhao, Xingyu
author_facet Bensalem, Saddek
Cheng, Chih-Hong
Huang, Wei
Huang, Xiaowei
Wu, Changshun
Zhao, Xingyu
contents Machine learning has made remarkable advancements, but confidently utilising learning-enabled components in safety-critical domains still poses challenges. Among the challenges, it is known that a rigorous, yet practical, way of achieving safety guarantees is one of the most prominent. In this paper, we first discuss the engineering and research challenges associated with the design and verification of such systems. Then, based on the observation that existing works cannot actually achieve provable guarantees, we promote a two-step verification method for the ultimate achievement of provable statistical guarantees.
format Preprint
id arxiv_https___arxiv_org_abs_2307_11784
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle What, Indeed, is an Achievable Provable Guarantee for Learning-Enabled Safety Critical Systems
Bensalem, Saddek
Cheng, Chih-Hong
Huang, Wei
Huang, Xiaowei
Wu, Changshun
Zhao, Xingyu
Machine Learning
Artificial Intelligence
Software Engineering
Systems and Control
Machine learning has made remarkable advancements, but confidently utilising learning-enabled components in safety-critical domains still poses challenges. Among the challenges, it is known that a rigorous, yet practical, way of achieving safety guarantees is one of the most prominent. In this paper, we first discuss the engineering and research challenges associated with the design and verification of such systems. Then, based on the observation that existing works cannot actually achieve provable guarantees, we promote a two-step verification method for the ultimate achievement of provable statistical guarantees.
title What, Indeed, is an Achievable Provable Guarantee for Learning-Enabled Safety Critical Systems
topic Machine Learning
Artificial Intelligence
Software Engineering
Systems and Control
url https://arxiv.org/abs/2307.11784