EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Straggler- and Adversary-Tolerant Secure Distributed Matrix Multiplication Using Polynomial Codes

Eimear Byrne iD, Oliver W Gnilke iD, Jörg Kliewer iD

Entropy · 31.01.2023 · 25 (2) · 266

DOI: 10.3390/e25020266 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Entropy, online erschienen am 31.01.2023. Verfasst von Eimear Byrne und 2 weiteren Autor:innen. Beteiligte Einrichtung: School of Mathematics and Statistics, University College Dublin, D04 V1W8 Dublin, Ireland. Die Arbeit wurde bisher 13-mal zitiert (Crossref) und nennt 32 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: UCD Seed Funding - Horizon Scanning Scheme, U.S. National Science Foundation grants, UCD Seed Funding-Horizon Scanning scheme.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Large matrix multiplications commonly take place in large-scale machine-learning applications. Often, the sheer size of these matrices prevent carrying out the multiplication at a single server. Therefore, these operations are typically offloaded to a distributed computing platform with a master server and a large amount of workers in the cloud, operating in parallel. For such distributed platforms, it has been recently shown that coding over the input data matrices can reduce the computational delay by introducing a tolerance against straggling workers, i.e., workers for which execution time significantly lags with respect to the average. In addition to exact recovery, we impose a security constraint on both matrices to be multiplied. Specifically, we assume that workers can collude and eavesdrop on the content of these matrices. For this problem, we introduce a new class of polynomial codes with fewer non-zero coefficients than the degree +1. We provide closed-form expressions for the recovery threshold and show that our construction improves the recovery threshold of existing schemes in the literature, in particular for larger matrix dimensions and a moderate to large number of colluding workers. In the absence of any security constraints, we show that our construction is optimal in terms of recovery threshold.

Schlagwörter

Stichwörter der Autor:innen

distributed computationdistributed learninginformation theoretic securitymatrix multiplicationpolynomial codes

Autor:innen und Institutionen

  • School of Mathematics and Statistics, University College Dublin, D04 V1W8 Dublin, Ireland.
  • Department of Mathematical Sciences, Aalborg University, 9220 Aalborg, Denmark.
  • Department of Electrical and Computer Engineering, New Jersey Institute of Technology, Newark, NJ 07410, USA.

ORCID-Profile: Eimear Byrne, Oliver W Gnilke, Jörg Kliewer

Förderung

  • UCD Seed Funding - Horizon Scanning Scheme
  • U.S. National Science Foundation grants
  • UCD Seed Funding-Horizon Scanning scheme
  • U.S. National Science Foundation

Zitieren

APA

Byrne, E., Gnilke, O. W., & Kliewer, J. (2023). Straggler- and Adversary-Tolerant Secure Distributed Matrix Multiplication Using Polynomial Codes. Entropy, 25(2), 266. https://doi.org/10.3390/e25020266

Vancouver

Byrne E, Gnilke OW, Kliewer J. Straggler- and Adversary-Tolerant Secure Distributed Matrix Multiplication Using Polynomial Codes. Entropy. 2023;25(2):266. doi:10.3390/e25020266

BibTeX
@article{Byrne2023,
  author = {Byrne, Eimear and Gnilke, Oliver W and Kliewer, Jörg},
  title = {Straggler- and Adversary-Tolerant Secure Distributed Matrix Multiplication Using Polynomial Codes},
  journal = {Entropy},
  year = {2023},
  volume = {25},
  number = {2},
  pages = {266},
  doi = {10.3390/e25020266}
}
Kostenlos starten →