{"id":"arxiv/2605.10727","url":"https://arxiv.org/abs/2605.10727","tags":["cs.LG","math.DG"],"score":0,"title":"Kernel-Gradient Drifting Models","domain":"cs.LG","source":"arxiv","authors":["Maria Esteban-Casadevall","Jorge Carrasco-Pollo","Max Welling","Jan-Willem van de Meent","Erik J. Bekkers","Floor Eijkelboom"],"reviews":0,"subject":"cs.LG","summary":"We propose kernel-gradient drifting, a one-step generative modeling framework that replaces the fixed Euclidean displacement direction in drifting models with directions induced by the kernel itself. Standard drifting is attractive because it enables fast, high-quality generation without distilling a large pretrained diffusion model, but its theory is currently understood mainly for Gaussian kernels, where the drift coincides with smoothed score matching and is identifiable. Our gradient-based reformulation exposes this score-based structure for general kernels: the resulting drift is the…","category":"research","comments":0,"disputed":false,"division":"cs","mentions":0,"discussed":0,"published":"2026-05-11","reactions":{}}