A Comparison of Multi-task learning and Single-task learning Approaches

Thomas Marquet; Elisabeth Oswald

Paper 2023/611

A Comparison of Multi-task learning and Single-task learning Approaches

Thomas Marquet, University of Klagenfurt

Elisabeth Oswald

, University of Klagenfurt, University of Birmingham

Abstract

In this paper, we provide experimental evidence for the benefits of multi-task learning in the context of masked AES implementations (via the ASCADv1-r and ASCADv2 databases). We develop an approach for comparing single-task and multi-task approaches rather than comparing specific resulting models: we do this by training many models with random hyperparameters (instead of comparing a few highly tuned models). We find that multi-task learning has significant practical advantages that make it an attractive option in the context of device evaluations: the multi-task approach leads to performant networks quickly in particular in situations where knowledge of internal randomness is not available during training.

Note: Additional experiments on a reduced dataset

Metadata

Available format(s): PDF
Category: Attacks and cryptanalysis
Publication info: Published elsewhere. Minor revision. AIHWS 2023
DOI: 10.1007/978-3-031-41181-6
Keywords: Side-channel attacks Multi-task learning Masking Deep Learning
Contact author(s): thomas marquet @ aau at
elisabeth oswald @ aau at
History: 2023-10-05: last of 3 revisions; 2023-04-28: received; See all versions
Short URL: https://ia.cr/2023/611
License: CC BY

BibTeX

@misc{cryptoeprint:2023/611,
      author = {Thomas Marquet and Elisabeth Oswald},
      title = {A Comparison of Multi-task learning and Single-task learning Approaches},
      howpublished = {Cryptology ePrint Archive, Paper 2023/611},
      year = {2023},
      doi = {10.1007/978-3-031-41181-6},
      note = {\url{https://eprint.iacr.org/2023/611}},
      url = {https://eprint.iacr.org/2023/611}
}