Larry A. Wasserman is the UPMC Professor of Statistics and Data Science in the Department of Statistics and Data Science and the Machine Learning Department at Carnegie Mellon University. He received my Ph.D. from the University of Toronto in 1988. He is known for his work on high dimensional inference, nonparametric inference, machine learning, topological data analysis and astrostatistics. Much of his work has been devoted to providing statistical foundations for algorithms in machine learning. He was awarded the COPSS Presidents’ Award in 1999 for the outstanding statistician under age 40, the Statistical Society of Canada Prize in Statistics in 2002, the DeGroot Prize in 2006 and the Reitz Lecturer in 2013. He is an elected fellow of the American Statistical Association, Institute of Mathematical Statistics, the American Association for the Advancement of Science. He is a member of the National Academy of Sciences.

His research papers frequently focus on high dimensional inference, machine learning, clustering, nonparametric inference, topological data analysis and astrostatistics. I have also written two advanced statistics textbooks: “All of Statistics,” and “All of Nonparametric Statistics.”

Public Lecture

March 14, 2018, 4:00 p.m. - 5:30 p.m.

102 Thomas Building

"The Greatest Debate in the History of Science"

The field of Statistics is concerned with quantifying uncertainty. But since its beginning, the field has been haunted by a debate about how to define uncertainty. The two dominant schools of thought are the Bayesian approach --- which regards probability as degree of belief --- and the frequentist approach, which regards probability as long-run frequencies. This debate has infected other sciences such as astronomy, biology and machine learning. Unlike many other philosophical debates in science, this debate has serious practical consequences. The medications you take were approved using statistical techniques and the choice of these techniques hinges on these philosophical questions.

Scientific Lecture

March 15, 2018, 4:00 p.m. - 5:00 p.m.

104 Thomas Building

"High Dimensional Multinomials and Unsmooth Densities "

 I will discuss a fundamental problem in Statistics: testing if a sample was drawn from a given distribution. Despite the long history of this problem, much is still not known especially when we allow for high dimensions and low smoothness. In this setting, we find the local minimax testing rates and we give some tests that are optimal. We show that these tests have much higher power than standard tests. This is joint work with Sivaraman Balakrishnan.