Projected Coupled Diffusion for Test-Time Constrained Joint Generation

arXiv:2508.10531v3 Announce Type: replace Abstract: Modifications to test-time sampling have emerged as an important extension to diffusion algorithms, with the goal of biasing the generative process to achieve a given objective without having to retrain the entire diffusion model. However, generating jointly correlated samples from multiple pre-trained diffusion models while simultaneously enforcing task-specific constraints without costly retraining has remained challenging. To this end, we propose Projected Coupled Diffusion (PCD), a novel test-time framework for constrained joint generation. PCD introduces a coupled guidance term into the generative dynamics to encourage coordination between diffusion models and incorporates a projection step at each diffusion step to enforce hard constraints. Empirically, we demonstrate the effectiveness of PCD in application scenarios of image-pair generation, object manipulation, and multi-robot motion planning. Our results show improved coupling effects and guaranteed constraint satisfaction without incurring excessive computational costs.

Related

• Neural Network-Based Score Estimation in Diffusion Models: Optimization and Generalization
• BezierFlow: Learning Bezier Stochastic Interpolant Schedulers for Few-Step Generation
• Neural Continuous-Time Markov Chain: Discrete Diffusion via Decoupled Jump Timing and Direction
• On the Interpolation Effect of Score Smoothing in Diffusion Models

Source: arXiv cs.LG | 2026-04-21