MCMC using Hamiltonian dynamics

This is a past event

15 people went

Human Spaces

216 Haven Street · Saint Louis, MO

How to find us

Don't worry if you get to the location and find yourself in a residential neighborhood—that's exactly where you're supposed to be!

Location image of event venue

Details

Michael Schmidt will be presenting on "MCMC using Hamiltonian dynamics" by Radford M. Neal from "Handbook of Markov Chain Monte Carlo." Michael will help guide us through some terrifying equations to show how Hamiltonian Monte Carlo can make solving numerically difficult problems possible.

From the abstract: "Hamiltonian dynamics can be used to produce distant proposals for the Metropolis algorithm, thereby avoiding the slow exploration of the state space that results from the diffusive behaviour of simple random-walk proposals. Though originating in physics, Hamiltonian dynamics can be applied to most problems with continuous state spaces by simply introducing fictitious “momentum” variables."

You can see the paper here: https://arxiv.org/pdf/1206.1901.pdf

For more information on other MCMC methods see the paper here: https://www.webpages.uidaho.edu/~stevel/565/literature/macKayIntro%20MCMC.pdf

Michael Schmidt is a mega-nerd who enjoys programming and data-science; he enjoys physics and math as well.

Our sponsor this month is Beacon Hill Technologies. Be sure to thank them for providing food for us!

As always, Lisa Rokusek will be hosting us at Human Spaces. Thanks to her for putting up with us month after month!