BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:An analysis of the noise schedule for score-based generative model
 s - Antonio Ocello (Ecole Polytechnique Paris)
DTSTART:20240716T140000Z
DTEND:20240716T143000Z
UID:TALK219043@talks.cam.ac.uk
DESCRIPTION:Score-based generative models (SGMs) aim at estimating a targe
 t data distribution by learning score functions using only noise-perturbed
  samples from the target. Recent literature has focused extensively on ass
 essing the error between the target and estimated distributions\, gauging 
 the generative quality through the Kullback-Leibler (KL) divergence and Wa
 sserstein distances. Under mild assumptions on the data distribution\, we 
 establish an upper bound for the KL divergence between the target and the 
 estimated distributions\, explicitly depending on any time-dependent noise
  schedule. Under additional regularity assumptions\, taking advantage of f
 avorable underlying contraction mechanisms\, we provide a tighter error bo
 und in Wasserstein distance compared to state-of-the-art results. In addit
 ion to being tractable\, this upper bound jointly incorporates properties 
 of the target distribution and SGM hyperparameters that need to be tuned d
 uring training.
LOCATION:External
END:VEVENT
END:VCALENDAR