Trust-Region Noise Search (TRS)
Publication 路 Black-box alignment for diffusion and flow models.
Niklas Schweiger, K. Ram, Daniel Cremers
Accepted @ ReALM-GEN Workshop, ICLR 2026.
馃搫 Download Paper
We propose a simple trust-region based search algorithm (TRS) which treats the pre-trained generative and reward models as black boxes, only optimizing the source noise to achieve reward-agnostic alignment.
Abstract
Aligning generative models (like diffusion or flow models) with specific user preferences often requires differentiable reward functions or expensive fine-tuning. We propose Trust-Region Noise Search (TRS), a simple yet effective black-box alignment algorithm.
TRS treats the generative model and the reward function as completely opaque. Instead of updating model weights, it optimizes the source noise (the latent space) using a trust-region search. This makes it applicable to non-differentiable rewards and avoids the catastrophic forgetting associated with full fine-tuning.
Key Methodology
The core idea of TRS is to iteratively explore the noise space $\mathbb{R}^M$ to find regions that map to samples with higher rewards.
Impact
Our results demonstrate that TRS can steer models towards high-aesthetic scores in text-to-image tasks and optimal docking scores in molecule design鈥攁ll without ever calculating a single gradient through the generative model itself.
- Status: Download Paper PDF 路 Accepted @ ReALM-GEN Workshop @ ICLR 2026 (Rio de Janeiro)
- Collaborators: K. Ram, Prof. Daniel Cremers
- Focus: High-efficiency alignment, non-differentiable rewards, black-box optimization.