3 papers
cs.LG2026
Folded Transport MCMC: Eliminating Label Switching by Sampling on a Fundamental Domain
Jun Hu
In Bayesian mixture models and other exchangeable-component models, the posterior is invariant under permutation of component labels, creating m! equivalent modes-the label-switchi…
cs.LG2026
Self-Certifying Transport MCMC via Dual Spectral-Gap Certificates
Jun Hu
We propose CerT-MCMC, a framework that equips learned-transport Markov chain Monte Carlo with automatic, rigorous convergence certificates. A normalising flow maps a Gaussian refer…
cs.LG2026
Non-Vacuous Certification of Transport MCMC via Oscillation-Controlled Normalizing Flows
Jun Hu
Transport MCMC trains a normalizing flow to precondition Metropolis--Hastings proposals, achieving high empirical efficiency on challenging posteriors; yet no prior work produces a…