Differential Property Monitoring for Backdoor Detection

Otto Brechelmacher, Dejan Nickovic, Tobias Nießen, Sarah Sallinger (Autor:in und Vortragende:r), Georg Weissenbacher

Publikation: Beitrag in Buch oder TagungsbandVortrag mit Beitrag in TagungsbandBegutachtung

Abstract

A faithful characterization of backdoors is a prerequisite for an effective automated detection. Unfortunately, as we demonstrate, formalization attempts in terms of temporal safety properties prove far from trivial and may involve several revisions. Moreover, given the complexity of the task at hand, a hapless revision of a property may not only eliminate but also introduce inaccuracies in the specification. We introduce a method called differential property monitoring that addresses this challenge by monitoring discrepancies between two versions of a property, and illustrate that this technique can also be used to analyze observations of untrusted components. We demonstrate the utility of the approach using a range of case studies – including the recently discovered xz backdoor.
OriginalspracheEnglisch
TitelFormal Methods and Software Engineering
Untertitel25th International Conference on Formal Engineering Methods, (ICFEM) 2024
Redakteure/-innenKazuhiro Ogata, Dominique Mery, Meng Sun, Shaoying Liu
Herausgeber (Verlag)Springer
Seiten216-236
Band15394
Auflage1
ISBN (elektronisch)978-981-96-0617-7
ISBN (Print)978-981-96-0616-0
DOIs
PublikationsstatusVeröffentlicht - 2024
Veranstaltung25th International Conference on Formal Engineering Methods, ICFEM 2024 - Hiroshima, Hiroshima, Japan
Dauer: 2 Dez. 20246 Dez. 2024

Publikationsreihe

NameLecture Notes in Computer Science
Herausgeber (Verlag)Springer
Band15394
ISSN (Print)0302-9743
ISSN (elektronisch)1611-3349

Konferenz

Konferenz25th International Conference on Formal Engineering Methods, ICFEM 2024
Land/GebietJapan
StadtHiroshima
Zeitraum2/12/246/12/24

Research Field

  • Dependable Systems Engineering

Fingerprint

Untersuchen Sie die Forschungsthemen von „Differential Property Monitoring for Backdoor Detection“. Zusammen bilden sie einen einzigartigen Fingerprint.

Diese Publikation zitieren