The Serial Scaling Hypothesis

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Liu, Yuxi, Preechakul, Konpat, Kuwaranancharoen, Kananart, Bai, Yutong
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866908999224393728
author Liu, Yuxi
Preechakul, Konpat
Kuwaranancharoen, Kananart
Bai, Yutong
author_facet Liu, Yuxi
Preechakul, Konpat
Kuwaranancharoen, Kananart
Bai, Yutong
contents While machine learning has advanced through massive parallelization, we identify a critical blind spot: some problems are fundamentally sequential. These "inherently serial" problems-from mathematical reasoning to physical simulations to sequential decision-making-require sequentially dependent computational steps that cannot be efficiently parallelized. We formalize this distinction in complexity theory, and demonstrate that current parallel-centric architectures face fundamental limitations on such tasks. Then, we show for first time that diffusion models despite their sequential nature are incapable of solving inherently serial problems. We argue that recognizing the serial nature of computation holds profound implications on machine learning, model design, and hardware development.
format Preprint
id arxiv_https___arxiv_org_abs_2507_12549
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle The Serial Scaling Hypothesis
Liu, Yuxi
Preechakul, Konpat
Kuwaranancharoen, Kananart
Bai, Yutong
Machine Learning
Computational Complexity
68Q15, 68Q10, 68T07
F.1.1; F.1.3; I.2.6
While machine learning has advanced through massive parallelization, we identify a critical blind spot: some problems are fundamentally sequential. These "inherently serial" problems-from mathematical reasoning to physical simulations to sequential decision-making-require sequentially dependent computational steps that cannot be efficiently parallelized. We formalize this distinction in complexity theory, and demonstrate that current parallel-centric architectures face fundamental limitations on such tasks. Then, we show for first time that diffusion models despite their sequential nature are incapable of solving inherently serial problems. We argue that recognizing the serial nature of computation holds profound implications on machine learning, model design, and hardware development.
title The Serial Scaling Hypothesis
topic Machine Learning
Computational Complexity
68Q15, 68Q10, 68T07
F.1.1; F.1.3; I.2.6
url https://arxiv.org/abs/2507.12549