English

Machine learning for modular multiplication

Machine Learning 2024-03-01 v1 Cryptography and Security

Abstract

Motivated by cryptographic applications, we investigate two machine learning approaches to modular multiplication: namely circular regression and a sequence-to-sequence transformer model. The limited success of both methods demonstrated in our results gives evidence for the hardness of tasks involving modular multiplication upon which cryptosystems are based.

Keywords

Cite

@article{arxiv.2402.19254,
  title  = {Machine learning for modular multiplication},
  author = {Kristin Lauter and Cathy Yuanchen Li and Krystal Maughan and Rachel Newton and Megha Srivastava},
  journal= {arXiv preprint arXiv:2402.19254},
  year   = {2024}
}

Comments

14 pages, 12 figures. Comments welcome!