Zobrazeno 1 - 5
of 5
pro vyhledávání: '"Mbacke, Sokhna Diarra"'
Diffusion models are one of the most important families of deep generative models. In this note, we derive a quantitative upper bound on the Wasserstein distance between the data-generating distribution and the distribution learned by a diffusion mod
Externí odkaz:
http://arxiv.org/abs/2312.05989
Since their inception, Variational Autoencoders (VAEs) have become central in machine learning. Despite their widespread use, numerous questions regarding their theoretical properties remain open. Using PAC-Bayesian theory, this work develops statist
Externí odkaz:
http://arxiv.org/abs/2310.04935
We extend PAC-Bayesian theory to generative models and develop generalization bounds for models based on the Wasserstein distance and the total variation distance. Our first result on the Wasserstein distance assumes the instance space is bounded, wh
Externí odkaz:
http://arxiv.org/abs/2302.08942
Autor:
Mbacke, Sokhna Diarra
À cause de la complexité croissante des systèmes informatiques, ces derniers sont aujourd’hui étudiés au moyen de multiples modèles et formalismes. Ainsi, il est nécessaire de développer des théories qui unifient différentes approches aaf
Externí odkaz:
http://hdl.handle.net/20.500.11794/30666
Autor:
Mbacke, Sokhna Diarra
Due to their increasing complexity, today’s computer systems are studied using multiple models and formalisms. Thus, it is necessary to develop theories that unify different approaches in order to limit the risks of errors when moving from one form
Externí odkaz:
http://hdl.handle.net/20.500.11794/33008