Error as Signal: Stiffness-Aware Diffusion Sampling via Embedded Runge-Kutta Guidance

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Kong, Inho, Lee, Sojin, Hong, Youngjoon, Kim, Hyunwoo J.
Natura: Preprint
Pubblicazione: 2026
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866908977482170368
author Kong, Inho
Lee, Sojin
Hong, Youngjoon
Kim, Hyunwoo J.
author_facet Kong, Inho
Lee, Sojin
Hong, Youngjoon
Kim, Hyunwoo J.
contents Classifier-Free Guidance (CFG) has established the foundation for guidance mechanisms in diffusion models, showing that well-designed guidance proxies significantly improve conditional generation and sample quality. Autoguidance (AG) has extended this idea, but it relies on an auxiliary network and leaves solver-induced errors unaddressed. In stiff regions, the ODE trajectory changes sharply, where local truncation error (LTE) becomes a critical factor that deteriorates sample quality. Our key observation is that these errors align with the dominant eigenvector, motivating us to leverage the solver-induced error as a guidance signal. We propose Embedded Runge-Kutta Guidance (ERK-Guid), which exploits detected stiffness to reduce LTE and stabilize sampling. We theoretically and empirically analyze stiffness and eigenvector estimators with solver errors to motivate the design of ERK-Guid. Our experiments on both synthetic datasets and the popular benchmark dataset, ImageNet, demonstrate that ERK-Guid consistently outperforms state-of-the-art methods. Code is available at https://github.com/mlvlab/ERK-Guid.
format Preprint
id arxiv_https___arxiv_org_abs_2603_03692
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Error as Signal: Stiffness-Aware Diffusion Sampling via Embedded Runge-Kutta Guidance
Kong, Inho
Lee, Sojin
Hong, Youngjoon
Kim, Hyunwoo J.
Computer Vision and Pattern Recognition
Artificial Intelligence
Classifier-Free Guidance (CFG) has established the foundation for guidance mechanisms in diffusion models, showing that well-designed guidance proxies significantly improve conditional generation and sample quality. Autoguidance (AG) has extended this idea, but it relies on an auxiliary network and leaves solver-induced errors unaddressed. In stiff regions, the ODE trajectory changes sharply, where local truncation error (LTE) becomes a critical factor that deteriorates sample quality. Our key observation is that these errors align with the dominant eigenvector, motivating us to leverage the solver-induced error as a guidance signal. We propose Embedded Runge-Kutta Guidance (ERK-Guid), which exploits detected stiffness to reduce LTE and stabilize sampling. We theoretically and empirically analyze stiffness and eigenvector estimators with solver errors to motivate the design of ERK-Guid. Our experiments on both synthetic datasets and the popular benchmark dataset, ImageNet, demonstrate that ERK-Guid consistently outperforms state-of-the-art methods. Code is available at https://github.com/mlvlab/ERK-Guid.
title Error as Signal: Stiffness-Aware Diffusion Sampling via Embedded Runge-Kutta Guidance
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2603.03692