Generalizing Hamiltonian Monte Carlo with Neural Networks
arXiv:1711.09268
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 106x 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.
arXiv commentICLR 2018
References in corpus (16)
- Adam: A Method for Stochastic Optimization
- Deep Unsupervised Learning using Nonequilibrium Thermodynamics
- Density estimation using Real NVP
- Stochastic Gradient Hamiltonian Monte Carlo
- A note on the evaluation of generative models
- Markov Chain Monte Carlo and Variational Inference: Bridging the Gap
- Deep Generative Stochastic Networks Trainable by Backprop
- Accelerate Monte Carlo Simulations with Restricted Boltzmann Machines
- Two phase transitions in the fully frustrated model
- A-NICE-MC: Adversarial Training for MCMC
- Can Boltzmann Machines Discover Cluster Updates ?
- Hamiltonian Monte Carlo Without Detailed Balance
- Riemannian Manifold Hamiltonian Monte Carlo
- Hamiltonian Annealed Importance Sampling for partition function estimation
- Magnetic Hamiltonian Monte Carlo
- A Markov Jump Process for More Efficient Hamiltonian Monte Carlo
Cited by in corpus (16)
- i-flow: High-dimensional Integration and Sampling with Normalizing Flows
- Denoising Diffusion Implicit Models
- Deep Energy Estimator Networks
- Equivariant Hamiltonian Flows
- Revisiting Bayesian Autoencoders with MCMC
- Adaptive Meta-Learning Stochastic Gradient Hamiltonian Monte Carlo Simulation for Bayesian Updating of Structural Dynamic Models
- Stochastic Normalizing Flows
- Metropolis-Hastings view on variational inference and adversarial training
- Deep Markov Chain Monte Carlo
- Adaptive Monte Carlo augmented with normalizing flows
- MetFlow: A New Efficient Method for Bridging the Gap between Markov Chain Monte Carlo and Variational Inference
- Self-Tuning Hamiltonian Monte Carlo for Accelerated Sampling
- Delayed rejection Hamiltonian Monte Carlo for sampling multiscale distributions
- -class Classification: an Anomaly Detection Method for Highly Imbalanced or Incomplete Data Sets
- Approximate Inference in Discrete Distributions with Monte Carlo Tree Search and Value Functions
- Neural Mode Jump Monte Carlo