BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Kernel Approximation of Wasserstein and Fisher-Rao Gradient flows
DTSTART:20241127T160000
DTEND:20241127T170000
DTSTAMP:20260513T144050Z
UID:0415111513eb727cf885e351461850a5a13892a30ae1263fd0e4eebf
CATEGORIES:Conferences - Seminars
DESCRIPTION:Prof. JJ Zhu (WIAS Berlin)\nAbstract:\nGradient flows have eme
 rged as a powerful framework for analyzing machine learning and statistica
 l inference algorithms. Motivated by several applications in statistical i
 nference\, generative models\, generalization and robustness of learning a
 lgorithms\, I will provide a few new results regarding the kernel approxim
 ation of gradient flows\, such as a hidden link between the gradient flows
  of kernel maximum-mean discrepancy and relative entropies. These findings
  not only advance our theoretical understanding but also provide practical
  tools for enhancing machine learning algorithms. I will showcase inferenc
 e and sampling algorithms using a new kernel approximation of the Wasserst
 ein-Fisher-Rao (a.k.a. Hellinger-Kantorovich) gradient flows\, which have 
 better convergence characterization and improved performance in computatio
 n.\n\nThe talk is based on the joint works with Alexander Mielke.\n\nSpeak
 er Bio:\nJia-Jie Zhu (https://jj-zhu.github.io/) is a machine learner\, ap
 plied mathematician\, and research group leader at the Weierstrass Institu
 te\, Berlin. Previously\, he worked as a postdoctoral researcher in machin
 e learning at the Max-Planck-Institute for Intelligent Systems\, Tübingen
 \, and received his Ph.D. training in optimization\, at the University of 
 Florida\, USA. He is interested in the intersection of machine learning\, 
 analysis\, and optimization\, on topics such as gradient flows of probabil
 ity measures\, optimal transport\, and robustness of learning and optimiza
 tion algorithms.\n\n\n 
LOCATION:CM 1 517 https://plan.epfl.ch/?room==CM%201%20517
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR