Name: Lucas José Gonçalves
Institution: UERJ, Brazil

The Hamiltonian Monte Carlo method (HMC) is a remarkable advance in Bayesian inference, although only recently has its understanding reached a good level among students and teachers. Today, however, such a method is consolidated as an alternative to MCMC methods, used to obtain samples from a posteriori distribution via iterative simulation with Markov Chains, especially in high-dimensional contexts. The objective of this work is to present the general aspects of this method, whose theoretical basis, formulated in terms of Differential Geometry and Mechanics, still is out of the reach of students of the science area at graduate level. Such an approach sacri fices rigor, but suffices for the presentation of a transient heat transfer application based on the Julia Programming Language