Asymptotically Exact Negative Guidance of Diffusion Models via Positive-Unlabeled Learning
Abstract
Negative guidance of diffusion models generates samples that avoid an undesired condition, specified either by reference samples or by a conditional score alongside an unconditional model. Existing methods inject repulsive terms derived from local information at each sample, so the resulting sampler does not target an explicit density, and its bias cannot be systematically reduced by increasing computation. We instead present a negative-guidance method that samples from an explicit target density, which uniformly handles both reference-sample and score-based specifications and suppresses probability mass near the undesired condition with an avoidance strength that parametrizes the target density itself. Realizing this target requires estimating, at each diffusion timestep, how likely each sample is to belong to the undesired condition. We cast this estimation as a Positive-Unlabeled learning problem that admits lightweight online training along the diffusion trajectory, using condition samples as positives and diffusion particles as unlabeled data. Because the guided particles drift from the unconditional marginal, we further correct the residual mismatch via Sequential Monte Carlo, yielding a sampler that is asymptotically exact in the particle limit and that turns any existing negative-guidance method into its proposal kernel. Empirically, our method approaches the ground-truth target on Gaussian mixtures and Pareto-dominates existing methods on trade-offs between avoidance and generation quality, distributional bias, and diversity, including in an image-generation setting with a pretrained score network.
Related
- First-Passage Time Problems for Jump-Diffusion Processes with Positive and Negative Jumps
- Partial AUC Maximization from Positive-unlabeled Data
- AUC Maximization from Biased Positive-unlabeled Data with Confidence
- Fraud Detection via Bayesian Positive-Unlabeled Learning with Gaussian Processes and Multilayer Networks
- When are time-to-event models a waste of time? Bridging mixture cure models and positive-unlabeled learning for binary classification under right-censoring
BibTeX
@misc{takagi2026asymptotically,
title = {{Asymptotically Exact Negative Guidance of Diffusion Models via Positive-Unlabeled Learning}},
author = {Hirohane Takagi and Masashi Sugiyama},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.19802},
eprint = {2610.19802},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.19802},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Asymptotically Exact Negative Guidance of Diffusion Models via Positive-Unlabeled Learning
AU - Hirohane Takagi
AU - Masashi Sugiyama
PY - 2026
DA - 2026/10/21
PB - arXiv
DO - 10.48550/arxiv.2610.19802
UR - https://arxiv.org/abs/2610.19802
AB - Negative guidance of diffusion models generates samples that avoid an undesired condition, specified either by reference samples or by a conditional score alongside an unconditional model. Existing methods inject repulsive terms derived from local information at each sample, so the resulting sampler does not target an explicit density, and its bias cannot be systematically reduced by increasing computation. We instead present a negative-guidance method that samples from an explicit target density, which uniformly handles both reference-sample and score-based specifications and suppresses probability mass near the undesired condition with an avoidance strength that parametrizes the target density itself. Realizing this target requires estimating, at each diffusion timestep, how likely each sample is to belong to the undesired condition. We cast this estimation as a Positive-Unlabeled learning problem that admits lightweight online training along the diffusion trajectory, using condition samples as positives and diffusion particles as unlabeled data. Because the guided particles drift from the unconditional marginal, we further correct the residual mismatch via Sequential Monte Carlo, yielding a sampler that is asymptotically exact in the particle limit and that turns any existing negative-guidance method into its proposal kernel. Empirically, our method approaches the ground-truth target on Gaussian mixtures and Pareto-dominates existing methods on trade-offs between avoidance and generation quality, distributional bias, and diversity, including in an image-generation setting with a pretrained score network.
ER -