Registered Data

[00383] A kernel-based method for Schrödinger bridges

  • Session Time & Room : 3E (Aug.23, 17:40-19:20) @D514
  • Type : Contributed Talk
  • Abstract : We propose a kernel-based machine learning method for numerical solutions of Schrödinger's bridge problems. In this method, the terminal time distribution constraint is described by a kernel-based metric on the space of probability measures.
  • Classification : 93E20
  • Format : Talk at Waseda University
  • Author(s) :
    • Yumiharu Nakano (Tokyo Institute of Technology)