Show simple item record

dc.contributor.authorAslansefat, K.
dc.contributor.authorKabir, Sohag
dc.contributor.authorAbdullatif, Amr R.A.
dc.contributor.authorVasudevan, Vinod
dc.contributor.authorPapadopoulos, Y.
dc.date.accessioned2021-08-10T11:11:33Z
dc.date.accessioned2021-09-23T09:43:02Z
dc.date.available2021-08-10T11:11:33Z
dc.date.available2021-09-23T09:43:02Z
dc.date.issued2021-08
dc.identifier.citationAslansefat K, Kabir S, Abdullatif ARA et al (2021) Toward Improving Confidence in Autonomous Vehicle Software: A Study on Traffic Sign Recognition Systems. Computer. 54(8): 66-76.en_US
dc.identifier.urihttp://hdl.handle.net/10454/18591
dc.descriptionYesen_US
dc.description.abstractThis article proposes an approach named SafeML II, which applies empirical cumulative distribution function-based statistical distance measures in a designed human-in-the loop procedure to ensure the safety of machine learning-based classifiers in autonomous vehicle software. The application of artificial intelligence (AI) and data-driven decision-making systems in autonomous vehicles is growing rapidly. As autonomous vehicles operate in dynamic environments, the risk that they can face an unknown observation is relatively high due to insufficient training data, distributional shift, or cyber-security attack. Thus, AI-based algorithms should make dependable decisions to improve their interpretation of the environment, lower the risk of autonomous driving, and avoid catastrophic accidents. This paper proposes an approach named SafeML II, which applies empirical cumulative distribution function (ECDF)-based statistical distance measures in a designed human-in-the-loop procedure to ensure the safety of machine learning-based classifiers in autonomous vehicle software. The approach is model-agnostic and it can cover various machine learning and deep learning classifiers. The German Traffic Sign Recognition Benchmark (GTSRB) is used to illustrate the capabilities of the proposed approach.en_US
dc.description.sponsorshipThis work was supported by the Secure and Safe MultiRobot Systems (SESAME) H2020 Project under Grant Agreement 101017258.en_US
dc.language.isoenen_US
dc.publisherIEEE
dc.relation.isreferencedbyhttps://doi.org/10.1109/MC.2021.3075054en_US
dc.rights© 2021 IEEE. Reproduced in accordance with the publisher's self-archiving policy.en_US
dc.subjectAutonomous systemsen_US
dc.subjectSafety assuranceen_US
dc.subjectAI safetyen_US
dc.subjectStatistical distance measureen_US
dc.subjectSafeMLen_US
dc.subjectSafe machine learningen_US
dc.titleToward Improving Confidence in Autonomous Vehicle Software: A Study on Traffic Sign Recognition Systemsen_US
dc.status.refereedYesen_US
dc.date.Accepted2021-04-19
dc.date.application2021-08-03
dc.typeArticleen_US
dc.type.versionAccepted manuscripten_US
dc.date.updated2021-08-10T10:11:39Z
refterms.dateFOA2021-09-23T09:43:30Z
dc.openaccess.statusGreenen_US


Item file(s)

Thumbnail
Name:
Abdullatif_ARA_2021.pdf
Size:
1.105Mb
Format:
PDF
Description:
Accepted manuscript
Thumbnail
Name:
Toward_Improving_Confidence_in ...
Size:
1.684Mb
Format:
PDF
Description:
Published version - keep suppressed

This item appears in the following Collection(s)

Show simple item record