ResearchTrend.AI
  • Papers
  • Communities
  • Events
  • Blog
  • Pricing
Papers
Communities
Social Events
Terms and Conditions
Pricing
Parameter LabParameter LabTwitterGitHubLinkedInBlueskyYoutube

© 2025 ResearchTrend.AI, All rights reserved.

  1. Home
  2. Papers
  3. 2204.12965
8
21

Particle algorithms for maximum likelihood training of latent variable models

27 April 2022
Juan Kuntz
Jen Ning Lim
A. M. Johansen
    FedML
ArXivPDFHTML
Abstract

(Neal and Hinton, 1998) recast maximum likelihood estimation of any given latent variable model as the minimization of a free energy functional FFF, and the EM algorithm as coordinate descent applied to FFF. Here, we explore alternative ways to optimize the functional. In particular, we identify various gradient flows associated with FFF and show that their limits coincide with FFF's stationary points. By discretizing the flows, we obtain practical particle-based algorithms for maximum likelihood estimation in broad classes of latent variable models. The novel algorithms scale to high-dimensional settings and perform well in numerical experiments.

View on arXiv
Comments on this paper