English

Bayesian Learning via Neural Schr\"odinger-F\"ollmer Flows

Machine Learning 2022-10-27 v9 Machine Learning

Abstract

In this work we explore a new framework for approximate Bayesian inference in large datasets based on stochastic control (i.e. Schr\"odinger bridges). We advocate stochastic control as a finite time and low variance alternative to popular steady-state methods such as stochastic gradient Langevin dynamics (SGLD). Furthermore, we discuss and adapt the existing theoretical guarantees of this framework and establish connections to already existing VI routines in SDE-based models.

Keywords

Cite

@article{arxiv.2111.10510,
  title  = {Bayesian Learning via Neural Schr\"odinger-F\"ollmer Flows},
  author = {Francisco Vargas and Andrius Ovsianas and David Fernandes and Mark Girolami and Neil D. Lawrence and Nikolas Nüsken},
  journal= {arXiv preprint arXiv:2111.10510},
  year   = {2022}
}
R2 v1 2026-06-24T07:45:37.136Z