denoising-diffusion-probabilistic-models
[ICLR 2026] SPREAD: Sampling-based Pareto front Refinement via Efficient Adaptive Diffusion