Reducing the cost of machine learning differential attacks using bit selection and a partial ML-distinguisher

Simple item page
dc.contributor.authorEbrahimi, Amirhossein
dc.contributor.authorRegazzoni, Francesco
dc.contributor.authorPalmieri, Paolo
dc.contributor.funderScience Foundation Irelanden
dc.date.accessioned2023-05-17T15:04:44Z
dc.date.available2023-05-09T15:09:11Zen
dc.date.available2023-05-17T15:04:44Z
dc.date.issued2023-04-01
dc.date.updated2023-05-09T14:09:14Zen
dc.description.abstractIn a differential cryptanalysis attack, the attacker tries to observe a block cipher’s behavior under an input difference: if the system’s resulting output differences show any non-random behavior, a differential distinguisher is obtained. While differential cryptanlysis has been known for several decades, Gohr was the first to propose in 2019 the use of machine learning (ML) to build a distinguisher. In this paper, we present the first Partial Differential (PD) ML distinguisher, and demonstrate its effectiveness on cipher SPECK32/64. As a PD-ML-distinguisher is based on a selection of bits rather than all bits in a block, we also study if different selections of bits have different impact in the accuracy of the distinguisher, and we find that to be the case. More importantly, we also establish that certain bits have reliably higher effectiveness than others, through a series of independent experiments on different datasets, and we propose an algorithm for assigning an effectiveness score to each bit in the block. By selecting the highest scoring bits, we are able to train a partial ML-distinguisher over 8-bits that is almost as accurate as an equivalent ML-distinguisher over the entire 32 bits (68.8% against 72%), for six rounds of SPECK32/64. Furthermore, we demonstrate that our obtained machine can reduce the time complexity of the key-averaging algorithm for training a 7-round distinguisher by a factor of 25 at a cost of only 3% in the resulting machine’s accuracy. These results may therefore open the way to the application of (partial) ML-based distinguishers to ciphers whose block size has so far been considered too large.
dc.description.statusPeer reviewed
dc.description.versionAccepted Version
dc.format.mimetypeapplication/pdfen
dc.identifier.citationEbrahimi, A., Regazzoni, F. and Palmieri, P. (2023) ‘Reducing the cost of machine learning differential attacks using bit selection and a partial ml-distinguisher’, in G.-V. Jourdan, L. Mounier, C. Adams, F. Sèdes, and J. Garcia-Alfaro (eds) Foundations and Practice of Security, FPS 2022, Lecture Notes in Computer Science, vol 13877, Cham: Springer Nature Switzerland, pp. 123-141. https://doi.org/10.1007/978-3-031-30122-3_8.en
dc.identifier.doi10.1007/978-3-031-30122-3_8
dc.identifier.endpage141
dc.identifier.isbn978-3-031-30121-6
dc.identifier.isbn978-3-031-30122-3
dc.identifier.journaltitleLecture Notes in Computer Science
dc.identifier.startpage123
dc.identifier.urihttps://hdl.handle.net/10468/14480
dc.identifier.volume13877
dc.language.isoenen
dc.publisherSpringer
dc.relation.ispartofFoundations and Practice of Security. FPS 2022
dc.relation.projectinfo:eu-repo/grantAgreement/SFI/SFI Centres for Research Training Programme::Data and ICT Skills for the Future/18/CRT/6222/IE/SFI Centre for Research Training in Advanced Networks for Sustainable Societies/
dc.rights© 2023. This version of the article has been accepted for publication, after peer review and is subject to Springer Nature’s AM terms of use, but is not the Version of Record and does not reflect post-acceptance improvements, or any corrections. The Version of Record is available online at: https://doi.org/10.1007/978-3-031-30122-3_8.
dc.subjectDifferential cryptanalysisen
dc.subjectMachine Learning based cryptanalysisen
dc.subjectPartial ML-distinguisheren
dc.titleReducing the cost of machine learning differential attacks using bit selection and a partial ML-distinguisheren
dc.typeArticle (peer-reviewed)en
dc.typeConference_Itemen
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
partial_machine_learning_distinguisher.pdf
Size:
635.24 KB
Format:
Adobe Portable Document Format
Description:
Accepted version
Download
Collections
Computer Science - Conference Items