AI RESEARCH
Asymptotic Learning Curves for Diffusion Models with Random Features Score and Manifold Data
arXiv CS.LG
•
ArXi:2603.22962v1 Announce Type: new We study the theoretical behavior of denoising score matching--the learning task associated to diffusion models--when the data distribution is ed on a low-dimensional manifold and the score is parameterized using a random feature neural network. We derive asymptotically exact expressions for the test, train, and score errors in the high-dimensional limit.