Persi Diaconis

I work in statistics and probability trying to make sense of the world in the face of randomness. In probability, I work on problems such as "how many times should a deck of cards be shuffled to mix it?" The answer is "about seven." This is abstracted to bounding the time to stationarity of Markov Chains and Random walks on groups. It leads to the design of new Monte Carlo algorithms and developing fresh understanding of group characters and structure. In statistics, I have tried to develop new ways of visualizing high dimensional data. A second focus has been understanding the Bayesian approach to statistics. This involves working on foundational issues like the nature of randomness as well as understanding the costs and benefits of mixing a-priori information with data in high-dimensional problems.