Bio

Etienne Bernard holds a PhD in physics from ENS Paris and is now a postdoc at MIT. In his thesis “Algorithms and applications of the Monte Carlo method: Two-dimensional melting and perfect sampling,” he designed Markov chain Monte Carlo algorithms in order to solve condensed matter problems. He still works on problems related to Monte Carlo algorithms, as well as on non-equilibrium statistical physics. He is also interested in topics related to statistics such as Bayesian inferences, probabilistic graphical models, and machine learning.

Project: Tree Algorithms to Store the Probability Distribution of Large Datasets

My project deals with computing and storing the probability distribution function of a large dataset (~terabytes). The storing process has to be fast, and the final distribution should use little memory space. Moreover, the algorithm has to be able to handle very different distributions.

Favorite Four-Color, Four State Turing Machine

Rule 118092828950265984164380