DP-KAN: Differentially Private Kolmogorov-Arnold Networks

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Kalinin, Nikita P., Bombari, Simone, Zakerinia, Hossein, Lampert, Christoph H.
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914875059470336
author Kalinin, Nikita P.
Bombari, Simone
Zakerinia, Hossein
Lampert, Christoph H.
author_facet Kalinin, Nikita P.
Bombari, Simone
Zakerinia, Hossein
Lampert, Christoph H.
contents We study the Kolmogorov-Arnold Network (KAN), recently proposed as an alternative to the classical Multilayer Perceptron (MLP), in the application for differentially private model training. Using the DP-SGD algorithm, we demonstrate that KAN can be made private in a straightforward manner and evaluated its performance across several datasets. Our results indicate that the accuracy of KAN is not only comparable with MLP but also experiences similar deterioration due to privacy constraints, making it suitable for differentially private model training.
format Preprint
id arxiv_https___arxiv_org_abs_2407_12569
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle DP-KAN: Differentially Private Kolmogorov-Arnold Networks
Kalinin, Nikita P.
Bombari, Simone
Zakerinia, Hossein
Lampert, Christoph H.
Machine Learning
Cryptography and Security
We study the Kolmogorov-Arnold Network (KAN), recently proposed as an alternative to the classical Multilayer Perceptron (MLP), in the application for differentially private model training. Using the DP-SGD algorithm, we demonstrate that KAN can be made private in a straightforward manner and evaluated its performance across several datasets. Our results indicate that the accuracy of KAN is not only comparable with MLP but also experiences similar deterioration due to privacy constraints, making it suitable for differentially private model training.
title DP-KAN: Differentially Private Kolmogorov-Arnold Networks
topic Machine Learning
Cryptography and Security
url https://arxiv.org/abs/2407.12569