A universal compression theory for lottery ticket hypothesis and neural scaling laws

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Wang, Hong-Yi, Luo, Di, Poggio, Tomaso, Chuang, Isaac L., Ziyin, Liu
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917301205336064
author Wang, Hong-Yi
Luo, Di
Poggio, Tomaso
Chuang, Isaac L.
Ziyin, Liu
author_facet Wang, Hong-Yi
Luo, Di
Poggio, Tomaso
Chuang, Isaac L.
Ziyin, Liu
contents When training large-scale models, the performance typically scales with the number of parameters and the dataset size according to a slow power law. A fundamental theoretical and practical question is whether comparable performance can be achieved with significantly smaller models and substantially less data. In this work, we provide a positive and constructive answer. We prove that a generic permutation-invariant function of $d$ objects can be asymptotically compressed into a function of $\operatorname{polylog} d$ objects with vanishing error, which is proved to be the optimal compression rate. This theorem yields two key implications: (Ia) a large neural network can be compressed to polylogarithmic width while preserving its learning dynamics; (Ib) a large dataset can be compressed to polylogarithmic size while leaving the loss landscape of the corresponding model unchanged. Implication (Ia) directly establishes a proof of the dynamical lottery ticket hypothesis, which states that any ordinary network can be strongly compressed such that the learning dynamics and result remain unchanged. (Ib) shows that a neural scaling law of the form $L\sim d^{-α}$ can be boosted to an arbitrarily fast power law decay, and ultimately to $\exp(-α' \sqrt[m]{d})$.
format Preprint
id arxiv_https___arxiv_org_abs_2510_00504
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle A universal compression theory for lottery ticket hypothesis and neural scaling laws
Wang, Hong-Yi
Luo, Di
Poggio, Tomaso
Chuang, Isaac L.
Ziyin, Liu
Machine Learning
Disordered Systems and Neural Networks
Information Theory
When training large-scale models, the performance typically scales with the number of parameters and the dataset size according to a slow power law. A fundamental theoretical and practical question is whether comparable performance can be achieved with significantly smaller models and substantially less data. In this work, we provide a positive and constructive answer. We prove that a generic permutation-invariant function of $d$ objects can be asymptotically compressed into a function of $\operatorname{polylog} d$ objects with vanishing error, which is proved to be the optimal compression rate. This theorem yields two key implications: (Ia) a large neural network can be compressed to polylogarithmic width while preserving its learning dynamics; (Ib) a large dataset can be compressed to polylogarithmic size while leaving the loss landscape of the corresponding model unchanged. Implication (Ia) directly establishes a proof of the dynamical lottery ticket hypothesis, which states that any ordinary network can be strongly compressed such that the learning dynamics and result remain unchanged. (Ib) shows that a neural scaling law of the form $L\sim d^{-α}$ can be boosted to an arbitrarily fast power law decay, and ultimately to $\exp(-α' \sqrt[m]{d})$.
title A universal compression theory for lottery ticket hypothesis and neural scaling laws
topic Machine Learning
Disordered Systems and Neural Networks
Information Theory
url https://arxiv.org/abs/2510.00504