Dr. Vinayak Rao Wins the International Society of Bayesian Analysis Savage Award

02-25-2014

Congratulations to Dr. Vinayak Rao of Purdue Statistics, who recently won the ISBA Savage Award in Theory and Methods for his Ph.D. thesis titled Markov Chain Monte Carlo for continuous-time, discrete state systems.This award is bestowed each year to two outstanding doctoral dissertations in Bayesian econometrics and statistics one each in : Theory and Methods, and Applied Methodology. The award was named in honor of Leonard J. “Jimmie” Savage, born in 1917. His most noted work was the 1954 book Foundations of Statistics, in which he put forward a theory of subjective and personal probability and statistics which forms one of the strands underlying Bayesian statistics and has applications to game theory.

Dr. Rao's thesis concerned a new class of Markov chain Monte Carlo algorithms for continuous-time systems. Such systems are important to describe temporal activity that is asynchronous with multiple time-scales (such as neuronal spiking, activity in gene networks or check-ins by FourSquare users). While continuous-time is appropriate to model such systems, computation involves approximations like discretizing time and is biased and computationally expensive. Dr. Rao's thesis developed and studied a class of algorithms that have no error by introducing random discretizations. Besides being exact and computationally inexpensive, this class of methods allows people to easily use ideas from the familiar discrete-time literature.