I am generally interested in statistical inference for stochastic processes, in particular, univariate and multivariate stochastic differential equations and stochastic partial differential equations. More specifically, my contributions are in the area of "exact" inference, that is, inference that does not depend on user-selected grids, or approximations. A significant component in this area is the use of a Bernoulli factory, i.e. an algorithm which simulates exact Bernoulli random variates with an unknown success probability. I work on developing exact Markov chain Monte Carlo techniques which only use approximations of the intractable target probability density function. Such methods are computationally intensive and thus I am also interested in high-performance GPU computing.

Radu Herbei joined the statistics faculty in 2006. He currently serves as the Publications Officer for the Statistical Computing section of the American Statistical Association. His research has been funded by NSF and ONR.