Guided Trajectory Optimization with Sparse Scaling for Test-Time Diffusion

TL;DR

This paper introduces RTS, a reward-guided trajectory scaling method that enhances diffusion model performance by actively optimizing noise exploration and prioritizing key denoising steps, leading to state-of-the-art results.

Contribution

RTS is a novel approach combining reward-guided noise optimization and sparse scaling with PCA analysis to improve diffusion model outputs at test time.

Findings

Outperforms baselines by 15.6% in GenEval Score

Achieves 60.4% improvement in ImageReward score

Sets new state-of-the-art in diffusion model evaluation

Abstract

The efficient Test-Time Scaling (TTS) paradigm offers a promising perspective for enhancing the generation performance of diffusion models. However, current solutions are limited to a static, pre-defined noise pool and suffer from inflexible noise exploration across the denoising trajectory. To bridge this gap, we propose RTS, a novel Reward-guided Trajectory Scaling method to fully unlock the generative potential of diffusion models. Unlike existing methods, RTS facilitates the synthesis of refined, high-fidelity images via two core innovations: 1) a reward-guided noise optimization strategy to actively direct the search towards promising regions; and 2) a sparse test-time scaling framework together with a PCA-driven curvature analysis scheme to prioritize key intermediate steps in the entire denoising space, effectively compressing the search space. Experiments show our approach outperforms baselines by 15.6% across GenEval Score, and a 60.4% enhancement in ImageReward score, setting a new SOTA while providing a practical guideline for more effective test-time scaling across diffusion-specific architectures.