AI RESEARCH
Memorisation, convergence and generalisation in generative models
arXiv CS.LG
•
ArXi:2605.21402v1 Announce Type: cross Generative neural networks set? To settle this question, Kadkhodaie, Guth, Simoncelli and Mallat (ICLR '24) trained diffusion models independently on disjoint subsets of a dataset and showed that they converge to nearly the same density when the number of