Projected Latent Markov Chain Monte Carlo: Conditional Sampling of Normalizing Flows

Chris Cannella · Mohammadreza Soltani · VAHID TAROKH

Keywords: [ Missing Data Inference ] [ Markov chain Monte Carlo ] [ Normalizing flows ] [ Conditional Sampling ]

[ Abstract ]
[ Paper ]
Tue 4 May 5 p.m. PDT — 7 p.m. PDT


We introduce Projected Latent Markov Chain Monte Carlo (PL-MCMC), a technique for sampling from the exact conditional distributions learned by normalizing flows. As a conditional sampling method, PL-MCMC enables Monte Carlo Expectation Maximization (MC-EM) training of normalizing flows from incomplete data. Through experimental tests applying normalizing flows to missing data tasks for a variety of data sets, we demonstrate the efficacy of PL-MCMC for conditional sampling from normalizing flows.

Chat is not available.