En poursuivant votre navigation sur ce site, vous acceptez l'utilisation d'un simple cookie d'identification. Aucune autre exploitation n'est faite de ce cookie. OK

Documents 49J40 3 results

Filter
Select: All / None
Q
Déposez votre fichier ici pour le déplacer vers cet enregistrement.
y

Barycenters for transport costs - Delon, Julie (Author of the conference) | CIRM H

Multi angle

optimal transport; Wasserstein barycenters

60A10 ; 49J40 ; 49K21 ; 49N15

Bookmarks Report an error
Déposez votre fichier ici pour le déplacer vers cet enregistrement.
y
Bayesian posterior distributions can be numerically intractable, even by the means of Markov Chain Monte Carlo methods. Bayesian variational methods can then be used to compute directly (and fast) a deterministic approximation of these posterior distributions. In this course, I describe the principles of the variational methods and their application in Bayesian inference, review main theoretical results and discuss their use on examples.

62F15 ; 62H12 ; 49J40

Bookmarks Report an error
Déposez votre fichier ici pour le déplacer vers cet enregistrement.
y

A critical Poincaré-Sobolev inequality - van den Bosch, Hanne (Author of the conference) | CIRM H

Multi angle

We study a specific Poincaré-Sobolev inequality in bounded domains, that has recently been used to prove a semi-classical bound on the kinetic energy of fermionic many-body states. The corresponding inequality in the entire space is precisely scale invariant and this gives rise to an interesting phenomenon. Optimizers exist for some (most ?) domains and do not exist for some other domains, at least for the isosceles triangle in two dimensions. In this talk, I will discuss bounds on the constant in the inequality and the proofs of existence and non-existence.
This is joint work with Rafael Benguria and Cristobal Vallejos (PUC, Chile)[-]
We study a specific Poincaré-Sobolev inequality in bounded domains, that has recently been used to prove a semi-classical bound on the kinetic energy of fermionic many-body states. The corresponding inequality in the entire space is precisely scale invariant and this gives rise to an interesting phenomenon. Optimizers exist for some (most ?) domains and do not exist for some other domains, at least for the isosceles triangle in two dimensions. ...[+]

35Q40 ; 49J40 ; 47J20

Bookmarks Report an error