GENERALIZING HAMILTONIAN MONTE CARLO WITH NEURAL NETWORKS By: Daniel Levy et, al.

[QiXuanWang]

Link: https://arxiv.org/pdf/1711.09268.pdf

Yu:
Don't know whether it can be used to generate samples more efficiently?

Abstract:
We present a general-purpose method to train Markov chain Monte Carlo kernels, parameterized by deep neural networks, that converge and mix quickly to
their target distribution. Our method generalizes Hamiltonian Monte Carlo and is
trained to maximize expected squared jumped distance, a proxy for mixing speed.
We demonstrate large empirical gains on a collection of simple but challenging
distributions, for instance achieving a 106× improvement in effective sample size
in one case, and mixing when standard HMC makes no measurable progress in a
second. Finally, we show quantitative and qualitative gains on a real-world task:
latent-variable generative modeling. We release an open source TensorFlow implementation of the algorithm.