Timezone: »

 
Poster
Entropy-based adaptive Hamiltonian Monte Carlo
Marcel Hirt · Michalis Titsias · Petros Dellaportas

Thu Dec 09 12:30 AM -- 02:00 AM (PST) @
in Poster Session 5 »

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.

Keywords: Generative Model

Author Information

Marcel Hirt (University College London)
Michalis Titsias (DeepMind)
Petros Dellaportas (University College London and Athens University of Economics)

More from the Same Authors