Entropy-based adaptive Hamiltonian Monte Carlo
arXiv:2110.14625
Abstract
Hamiltonian Monte Carlo (HMC) is a popular Markov Chain Monte Carlo (MCMC) algorithm to sample from an unnormalized probability distribution. A leapfrog integrator is commonly used to implement HMC in practice, but its performance can be sensitive to the choice of mass matrix used therein. We develop a gradient-based algorithm that allows for the adaptation of the mass matrix by encouraging the leapfrog integrator to have high acceptance rates while also exploring all dimensions jointly. In contrast to previous work that adapt the hyperparameters of HMC using some form of expected squared jumping distance, the adaptation strategy suggested here aims to increase sampling efficiency by maximizing an approximation of the proposal entropy. We illustrate that using multiple gradients in the HMC proposal can be beneficial compared to a single gradient-step in Metropolis-adjusted Langevin proposals. Empirical evidence suggests that the adaptation method can outperform different versions of HMC schemes by adjusting the mass matrix to the geometry of the target distribution and by providing some control on the integration time.
arXiv commentTo appear in NeurIPS (2021)
References in corpus (24)
- Adam: A Method for Stochastic Optimization
- Density estimation using Real NVP
- A Conceptual Introduction to Hamiltonian Monte Carlo
- On the ergodicity properties of some adaptive MCMC algorithms
- Residual Flows for Invertible Generative Modeling
- On Russian Roulette Estimates for Bayesian Inference with Doubly-Intractable Likelihoods
- Geometric integrators and the Hamiltonian Monte Carlo method
- Rapid Mixing of Hamiltonian Monte Carlo on Strongly Log-Concave Distributions
- NeuTra-lizing Bad Geometry in Hamiltonian Monte Carlo Using Neural Transport
- Leave Pima Indians alone: binary regression as a benchmark for Bayesian computation
- Variable transformation to obtain geometric ergodicity in the random-walk Metropolis algorithm
- On the convergence of Hamiltonian Monte Carlo
- Relaxing Bijectivity Constraints with Continuously Indexed Normalising Flows
- Fast mixing of Metropolized Hamiltonian Monte Carlo: Benefits of multi-step gradients
- Optimal Convergence Rate of Hamiltonian Monte Carlo for Strongly Logconcave Distributions
- tfp.mcmc: Modern Markov Chain Monte Carlo Tools Built for Modern Hardware
- Banded Matrix Operators for Gaussian Markov Models in the Automatic Differentiation Era
- Faster Hamiltonian Monte Carlo by Learning Leapfrog Scale: an offline randomized solution
- A general perspective on the Metropolis-Hastings kernel
- Gradient-based Adaptive Markov Chain Monte Carlo
- Stochastic Chebyshev Gradient Descent for Spectral Optimization
- A Condition Number for Hamiltonian Monte Carlo
- A Neural Network MCMC sampler that maximizes Proposal Entropy
- Sampling by Divergence Minimization