Machine learning for modular multiplication

Kristin Lauter; Cathy Yuanchen Li; Krystal Maughan; Rachel Newton; Megha Srivastava

Machine learning for modular multiplication

Kristin Lauter, Cathy Yuanchen Li, Krystal Maughan, Rachel Newton, Megha Srivastava

Research output: Chapter in Book/Report/Conference proceeding › Chapter › peer-review

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.

Original language	English
Title of host publication	Research Directions in Number Theory - Women in Numbers 6
Publisher	Springer
Publication status	Accepted/In press - 17 Sept 2024

Publication series

Name	Association for Women in Mathematics Series
Publisher	Springer

Cite this

@inbook{ad6fb0e37bbb4d8b8c5e956f1b2ea579,

title = "Machine learning for modular multiplication",

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.",

author = "Kristin Lauter and Li, {Cathy Yuanchen} and Krystal Maughan and Rachel Newton and Megha Srivastava",

year = "2024",

month = sep,

day = "17",

language = "English",

series = "Association for Women in Mathematics Series",

publisher = "Springer",

booktitle = "Research Directions in Number Theory - Women in Numbers 6",

}

TY - CHAP

T1 - Machine learning for modular multiplication

AU - Lauter, Kristin

AU - Li, Cathy Yuanchen

AU - Maughan, Krystal

AU - Newton, Rachel

AU - Srivastava, Megha

PY - 2024/9/17

Y1 - 2024/9/17

N2 - 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.

AB - 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.

M3 - Chapter

T3 - Association for Women in Mathematics Series

BT - Research Directions in Number Theory - Women in Numbers 6

PB - Springer

ER -

Machine learning for modular multiplication

Abstract

Publication series

Fingerprint

Cite this