Gradient Descent Fails to Learn High-frequency Functions and Modular Arithmetic

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Takhanov, Rustem, Tezekbayev, Maxat, Pak, Artur, Bolatov, Arman, Assylbekov, Zhenisbek
Format: Preprint
Veröffentlicht: 2023
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866929478363512832
author Takhanov, Rustem
Tezekbayev, Maxat
Pak, Artur
Bolatov, Arman
Assylbekov, Zhenisbek
author_facet Takhanov, Rustem
Tezekbayev, Maxat
Pak, Artur
Bolatov, Arman
Assylbekov, Zhenisbek
contents Classes of target functions containing a large number of approximately orthogonal elements are known to be hard to learn by the Statistical Query algorithms. Recently this classical fact re-emerged in a theory of gradient-based optimization of neural networks. In the novel framework, the hardness of a class is usually quantified by the variance of the gradient with respect to a random choice of a target function. A set of functions of the form $x\to ax \bmod p$, where $a$ is taken from ${\mathbb Z}_p$, has attracted some attention from deep learning theorists and cryptographers recently. This class can be understood as a subset of $p$-periodic functions on ${\mathbb Z}$ and is tightly connected with a class of high-frequency periodic functions on the real line. We present a mathematical analysis of limitations and challenges associated with using gradient-based learning techniques to train a high-frequency periodic function or modular multiplication from examples. We highlight that the variance of the gradient is negligibly small in both cases when either a frequency or the prime base $p$ is large. This in turn prevents such a learning algorithm from being successful.
format Preprint
id arxiv_https___arxiv_org_abs_2310_12660
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Gradient Descent Fails to Learn High-frequency Functions and Modular Arithmetic
Takhanov, Rustem
Tezekbayev, Maxat
Pak, Artur
Bolatov, Arman
Assylbekov, Zhenisbek
Machine Learning
Classes of target functions containing a large number of approximately orthogonal elements are known to be hard to learn by the Statistical Query algorithms. Recently this classical fact re-emerged in a theory of gradient-based optimization of neural networks. In the novel framework, the hardness of a class is usually quantified by the variance of the gradient with respect to a random choice of a target function. A set of functions of the form $x\to ax \bmod p$, where $a$ is taken from ${\mathbb Z}_p$, has attracted some attention from deep learning theorists and cryptographers recently. This class can be understood as a subset of $p$-periodic functions on ${\mathbb Z}$ and is tightly connected with a class of high-frequency periodic functions on the real line. We present a mathematical analysis of limitations and challenges associated with using gradient-based learning techniques to train a high-frequency periodic function or modular multiplication from examples. We highlight that the variance of the gradient is negligibly small in both cases when either a frequency or the prime base $p$ is large. This in turn prevents such a learning algorithm from being successful.
title Gradient Descent Fails to Learn High-frequency Functions and Modular Arithmetic
topic Machine Learning
url https://arxiv.org/abs/2310.12660