AI RESEARCH

On Forgetting and Stability of Score-based Generative models

arXiv CS.LG

ArXi:2601.21868v2 Announce Type: replace-cross Understanding the stability and long-time behavior of generative models is a fundamental problem in modern machine learning. This paper provides quantitative bounds on the sampling error of score-based generative models by leveraging stability and forgetting properties of the Marko chain associated with the reverse-time dynamics.