aiqbench analysis
This research introduces Jeffrey guidance as a principled method to extend diffusion model control beyond standard heuristics by leveraging Bayesian conditioning rules. The framework demonstrates improved fidelity and fairness capabilities when targeting specific embedding distributions or enforcing attribute independence in generated images. This advancement offers researchers more precise tools for manipulating the output distribution of generative models without excessively perturbing their underlying joint structures.
Towards More General Control of Diffusion Models Using Jeffrey Guidance
arXiv:2606.13240v1 Announce Type: cross Abstract: A key strength of diffusion models lies in their flexibility, since their outputs can be controlled at sampling time through guidance. However, beyond simple cases such as conditional sampling, the target distribution is often left implicit, defined only through a sampling rule or a heuristic energy function. To address this, we propose Jeffrey guidance, a principled framework that extends diffusion-model control to applications beyond what standard guidance can express. It leverages Jeffrey's rule of conditioning to update marginal distributions towards a prescribed target, preserving the conditional structure and minimally perturbing the joint distribution. We first demonstrate Jeffrey guidance by targeting a prescribed embedding distribution. With Inception embeddings as the target, this leads to substantial reductions in FID on both CIFAR-10 and FFHQ. We further apply Jeffrey guidance to fairness on CelebA-HQ, updating an unconditional diffusion model to enforce independence between attributes.
Abstract:A key strength of diffusion models lies in their flexibility, since their outputs can be controlled at sampling time through guidance. However, beyond simple cases such as conditional sampling, the target distribution is often left implicit, defined only through a sampling rule or a heuristic energy function. To address this, we propose Jeffrey guidance, a principled framework that extends diffusion-model control to applications beyond what standard guidance can express. It leverages Jeffrey's rule of conditioning to update marginal distributions towards a prescribed target, preserving the conditional structure and minimally perturbing the joint distribution. We first demonstrate Jeffrey guidance by targeting a prescribed embedding distribution. With Inception embeddings as the target, this leads to substantial reductions in FID on both CIFAR-10 and FFHQ. We further apply Jeffrey guidance to fairness on CelebA-HQ, updating an unconditional diffusion model to enforce independence between attributes.
Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Methodology (stat.ME); Machine Learning (stat.ML)
arXiv:2606.13240 [cs.LG]
(or arXiv:2606.13240v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2606.13240 arXiv-issued DOI via DataCite (pending registration)
Full text under the publisher's open licence · 173 words
Related coverage
MMD Guidance: Training-Free Distribution Adaptation for Diffusion Models via Maximum Mean Discrepancy Guidance
arXiv · Jun 10, 2026
Coupled Guidance for Flow Matching
arXiv · Oct 30, 2026
How to Guide Your Language Flow
arXiv · Sep 18, 2026
Asynchronous Multimodal Diffusion Policy Composition via Latency-Aware Guidance Fusion
arXiv · Jul 21, 2026