Larry Wasserman

Larry A. Wasserman

I began my undergraduate life as a computer science student. One year of typing punch cards convinced me to change fields. (Yes. I am old enough to remember punch cards.) Ironically, as a statistician, I often now work closely with computer scientists. Indeed, the lines between the two fields are blurring and, in fact, I have a strong connection with CALD (the Center for Automated Learning and Discovery) in the School of Computer Science. Fortunately, they've eliminated the punch cards. I've worked with colleagues in computer science on language modeling, analysis of simulation algorithms, and fast methods for statistical computation.

Statistics has points of contact with many fields, not just computer science. For example, I work with astrophysicists, statistical geneticists, and criminologists. In each case, I find that statistics is very helpful for solving scientific problems. But it works in the other direction, too. These scientific applications often lead to interesting methodological and theoretical statistics questions. The result is that I spend time doing both applications and theory. On the theoretical side, my interests largely have been in large sample theory, mixture models, nonparametric inference, Bayesian methods, multiple testing, and causal inference.

Some Related Publications

Shen, X. and Wasserman, L. (2000). Rates of convergence of posterior distributions. To appear: Annals of Statistics.

Genovese, C. and Wasserman, L. (2000). Rates of convergence for the Gaussian mixture sieve. Annals of Statistics, 28, 1105-1127.

Wasserman, L. (2000). Asymptotic inference for mixture models using data dependent priors. J. Roy. Statist. Soc. B, 62, 159-180.

Devlin, B., Roeder, K., and Wasserman, L. (2000). Genomic control for association studies: A semiparametric test to detect excess-haplotype sharing. Biostatistics, 1, 369-388.

Connolly, A.J., Nichol, A.J., Moore, A.W., Schneider, J., Genovese, C., and Wasserman, L. (2000). Fast algorithms and efficient statistics: Density estimation in large astronomical datasets. To appear: Astrophysics Journal.

CMU HomeH&SS HomeStatistics