Stochastic Gradient MCMC with Repulsive Forces

11/30/2018
by   Víctor Gallego, et al.
0

We propose a unifying view of two different families of Bayesian inference algorithms, SG-MCMC and SVGD. We show that SVGD plus a noise term can be framed as a multiple chain SG-MCMC method. Instead of treating each parallel chain independently from others, the proposed algorithm implements a repulsive force between particles, avoiding collapse. Experiments in both synthetic distributions and real datasets show the benefits of the proposed scheme.

READ FULL TEXT

Please sign up or login with your details

Forgot password? Click here to reset