Research Environment

Workshops, Seminars and Colloquia

Students and faculty in the Department of Statistics attend seminars and workshops on topics of current interest throughout the year. Seminars are formally organized by faculty interested in presenting specific topics. Workshops provide an informal setting for discussing work in progress or introducing Department members to new problems. Students may earn academic credit for their participation.

Topics from recent seminars and workshops include: Autonomic Function Data, Computer-Intensive Integration Techniques, Current and Foundational Issues in Experimental Design, Curved Exponential Families and Asymptotic Methods, Data Mining, Design of Clinical Trials, Finitely Additive Measures, Foundations of Statistics, Latent Variable Models, Markov Chain Monte Carlo Methods, PET Imaging, Repeated Measures, Semiparametric Methods, Spatial Statistics, Statistical Graphics, and Wavelets.

Graduate student Ashish Sanil presents his work in the Ph.D. seminar.

Graduate Students Alix Gitelman and Herbie Lee meet with Prof. Larry Wasserman to discuss dissertation work.

There is also a Department Colloquium Series with speakers from outside the Department. The topics span a wide variety of current research areas, and there is often lively discussion following the lecture. All members of the Department attend. In addition, there are occasional "consulting seminars" in which an investigator presents a problem in order to get advice on statistical issues that are raised.

In 1995, the Department hosted the Statistics and Computer Science: The Interface conference and, in 1998, the Department helped host the Spring Meeting of the International Biometric Society Eastern North American Region with the Institute of Mathematical Statistics and Sections of the American Statistical Association. The Department also hosts, every other year, a series of workshops on applications of Bayesian Statistics. Workshop 6, "Case Studies in Bayesian Statistics" was held on Sept. 28-29, 2001, at Carnegie Mellon University. These workshops attract participants form all over the world. In conjunction with these workshops, a special lecture is given by a distinguished statistician to honor the memory of Morris H. DeGroot. DeGroot, who was the first Head of the Department of Statistics when it was founded in 1967, was an inspirational leader at Carnegie Mellon and a major figure in the statistical community.

The Institute for Statistics and its Applications

The Institute for Statistics and its Applications (ISA) was established in the Department of Statistics with support from the National Science Foundation's Group Infrastructure Program. ISA's mission is to train pre-doctoral and young post-doctoral statisticians in cross-disciplinary research and teaching. ISA supports postdoctoral fellows and summer research and teaching fellows at CMU who may be Statisticians interested in cross-disciplinary research and teaching, or who may be psychologists, medical researchers, social or physical scientists, or engineers, who are interested in engaging in Statistical research. Special research programs within ISA include cognitive psychology; functional magnetic resonance imaging; genetics and psychiatric statistics; statistical physics; criminology, governmental statistics and public policy; and environmental statistics.

ISA greatly broadens your opportunities for cross-disciplinary research, and increases your interactions with professionals in Statistics and allied disciplines, by participating in ISA research programs and through contact with ISA research and teaching fellows.

Collaborative Research and Consulting

Faculty members in Carnegie Mellon's Department of Statistics are active in collaborative research and consulting, and students are strongly encouraged to participate. Besides formal exposure to statistical consulting in the Statistical Practice course, and their extensive project work in Advanced Data Analysis, students typically encounter a broad variety of statistical scientific problems as they progress through the program. These provide valuable experience, financial support, and sometimes inspiration for the doctoral thesis.

Some examples of recent Advanced Data Analysis projects are: The Meridional Temperature Gradient and Climate Change, The Relationship Between AFDC Benefits and State Birth and Abortion Rates: Evidence From State-Level Data, Analysis of Functional MRI for a Motor Task, Model Selection for Consumer Loan Application Data, Why do certain women have earlier menopause than the others?, An Analysis of Currency Options and Exchange Rate Distributions, Analysis of Arrest Rates, The Relationship Between Form and Function in the Speech of Specifically Language Imparied Children, Analysis of Data from a Fiberglass Manufacturing Process, and Postmortem Analysis of Neuron Distributions in the Locus Coeruleus of Alcoholics and Suicidal Victims.

Computing Environment

The Department of Statistics operates its own computer facilities which provide students with experience using over eighty advanced graphics workstations. The workstations are interconnected by a departmental Fast Ethernet which, in turn, is connected to University and worldwide networks. Twenty workstations are located in two department-owned computer rooms; these computers are designated primarily for graduate student use. In addition, there are several DOS/Windows PC's, and laser printers. The Department also has a graphics laboratory with equipment for producing computer-animated video tapes and computer-generated color graphics.

Students enjoy unlimited access to the computing facilities. Ken Shirakawa and Johnny Lam (left)

and Stella Salvatierra and Anita Araneda (right) work at graphics workstations in the department.

Shingo Oue, Prof. Bill Eddy, and Audris Mockus explore data visualization ideas using air traffic data in the Graphic Lab.

In 2000, the Department of Statistics at Carnegie Mellon, with partial support from a National Science Foundation SCREMS grant, purchased a 128 CPU mini-supercomputer. This computer, combined with the department's existing 16 processor computer, provides unprecedented computing capacity. The new machine has a total memory of 34 gigabytes, a total disk capacity of more than 1 terabyte, and a computational throughput of approximately 1 teraflop.

All students have full unlimited access to all of the computing equipment. Students and faculty use the facilities in a variety of creative ways. Some work on very large datasets such as data collected from complex manufacturing processes or large-scale surveys such as the U.S. National Crime Survey and the U.S. Census. Others run computationally intensive simulations of stochastic systems, study numerical methods for implementing Bayesian inference, or experiment with novel ways of visualizing data. The Department and University support a wide range of standard and experimental statistical software.

The Department believes that computation is an essential tool for applied statistics and that the theory of computational methods is an important area of research. For these reasons, the Department provides an introduction to the computing environment to all new students. Furthermore, the hardware and software are upgraded on a regular basis to ensure that the facilities remain current.

Next Topic, Recent Ph.D. Dissertations



CMU HomeH&SS HomeStatistics