BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CERN//INDICO//EN
BEGIN:VEVENT
SUMMARY:Local convergence for mean-field particle approximation of Wassers
 tein gradient flow
DTSTART:20260414T075000Z
DTEND:20260414T085000Z
DTSTAMP:20260411T111300Z
UID:indico-event-15617@indico.math.cnrs.fr
DESCRIPTION:Speakers: Pierre Monmarché (Université Gustave Eiffel)\n\nWa
 sserstein gradient flows of free energies are non-linear PDE which can be 
 interpreted in terms of McKean-Vlasov self-interacting diffusion processes
 . In practice\, using propagation of chaos\, they are approximated by syst
 ems of mean-field interacting particles. Up to now\, most works have studi
 ed the long-time convergence of these flows and the associated particles i
 n the relatively simple situation where the free energy has a unique criti
 cal point which is  the global minimizer\, and the invariant Gibbs measur
 e of the N-particle system satisfies a log-Sobolev inequality independent 
 from N. In this talk we focus on the other situation\, which occurs for th
 e case for the Curie-Weiss model in a multi-well potential at low temperat
 ure. In that case\, the particle system is metastable and propagation of c
 haos cannot hold uniformly in time. We will see how it is still possible t
 o get local convergence estimates of practical interest\, which describe a
  fast convergence (i.e. at a rate independent from N) toward a local minim
 izer and uniform propagation of chaos for a time scale  exponentially lar
 ge with N.\n\nhttps://indico.math.cnrs.fr/event/15617/
LOCATION:Salle K. Johnson (1R3-1er étage)
URL:https://indico.math.cnrs.fr/event/15617/
END:VEVENT
END:VCALENDAR