ART for Diffusion Sampling: A Reinforcement Learning Approach to Timestep Schedule

Excerpt

arXiv:2601.18681v3 Announce Type: replace Abstract: We consider time discretization for score-based diffusion models to generate samples from a learned reverse-time dynamic on a finite grid. Uniform and hand-crafted grids can be suboptimal given a budget on the number of time steps. We introduce Adaptive Reparameterized Time (ART), which controls the clock speed of a reparameterized time variable to redistribute computation along the sampling trajectory while preserving the terminal time, with t