Diffusion and flow models generate outputs from random noise. What if we could search for better noise? TRS does exactly that: it treats the generative model and any reward function as black boxes, and uses a trust-region search to find noise inputs that produce higher-reward outputs — without retraining or backpropagating through anything.
Key highlights
Optimizing the noise samples of diffusion and flow models is an increasingly popular approach to align these models to target rewards at inference time. However, we observe that these approaches are usually restricted to differentiable or cheap reward functions, the formulation of the underlying pre-trained generative model, or are memory/compute inefficient. We instead propose a simple trust-region based search algorithm (TRS) which treats the pre-trained generative and reward models as a black-box and only optimizes the source noise. Our approach achieves a good balance between global exploration and local exploitation, and is versatile and easily adaptable to various generative settings with minimal hyperparameter tuning. We evaluate TRS across text-to-image, molecule and protein design tasks, and obtain significantly improved output samples over the base generative models and other inference-time alignment approaches which optimize the source noise sample, or even the entire reverse-time sampling noise trajectories in the case of diffusion models.
A generative model maps a noise vector to an output. A reward function scores it. TRS searches for the noise inputs that maximize the reward — treating everything as a black box.
Same Stable Diffusion 1.5 model, same prompt — TRS only changes the source noise. Drag to compare.
Pick a prompt and compare TRS against every baseline under the same SDXL model and budget — TRS (highlighted) reaches the most faithful prompt alignment.
Rotate the actual optimized structures in 3D (rendered with Mol*, the rcsb.org/3d-view engine). Switch task and baseline — TRS lands closer to target molecule properties (lower loss) and folds into more designable protein backbones under the same budget.
Explore the scaling study interactively — switch task, reward model, and toggle baselines. TRS consistently reaches the highest reward under the same compute budget across all domains.
ECCV 2026
@misc{schweiger2026trustregionnoisesearchblackbox,
title={Trust-Region Noise Search for Black-Box Alignment of Diffusion and Flow Models},
author={Niklas Schweiger and Daniel Cremers and Karnik Ram},
year={2026},
eprint={2603.14504},
archivePrefix={arXiv},
primaryClass={cs.LG},
url={https://arxiv.org/abs/2603.14504},
}