------------------------------------------------------------ 
Title: APS Study on Boost-Phase Intercept Systems
Speaker: David Hafemeister
Physics Department
California Polytechnic State University

Date: September 4, 2003
Location: 101 Corcoran Hall 
Time: 4:15 p.m.
------------------------------------------------------------ 
Ever since 1983, there has been a hope (the sine qua non) to be able to attack the
first few minutes of missile launch, before MIRVs and chemical bomblets could be
released. We begin with a 10-minute history of defense in space events. Then, we
briefly analyze the following boost-phase attack systems in terms of location,
time-lines, energy on target, and duty factor (summarizing the 2003 APS report):
The 1980s x-ray laser weapon, the airborne laser (ABL) and space-base
interceptors (SBI). The role of countermeasures to defeat boost-phased defenses 
will be summarized. 
------------------------------------------------------------
Title: QTL analysis and genetic architecture of complex traits: 
where have we been and where are we going?

Speaker: Professor Zhao-Bang Zeng
Department of Statistics, North Carolina State University

Date: September 5, 2003
Location: Funger Hall 310
Time: 11:00 a.m.
------------------------------------------------------------ 
QTL analysis has changed the study of genetic basis of complex traits. With the availability of 
dense genome-wide molecular markers, it is now possible to map many QTL with medical 
implication or biological and agricultural importance into genome positions for further study.
Early research in QTL mapping analysis was mainly concerned with accurate and robust 
localization of individual QTL. More recent research, at least in experimental animal and plant 
populations, has been shifted to the inference of overall genetic architecture of complex traits, 
which may include the number, positions, effects, interaction and pleiotropy of QTL. After a 
brief introduction of the subject and early progress, I will concentrate the talk on some recent
development of statistical methodology to infer the genetic architecture of complex traits with
examples of analysis on some Drosophila data. In the end, I will briefly discuss some of our 
current efforts to extend the analysis to natural populations such as human population, with SNP 
data and to combine QTL analysis with microarray gene expression data.
------------------------------------------------------------ 

Title: Bioterrorism: Calibrating Risks and Responses
Speaker: Dr. Brad Roberts
Institute for Defense Analyses

Date: October 2, 2003
Location: Cochran Hall room 101, 725 21st St. NW
Time: 4:30 p.m.
------------------------------------------------------------ 

Are the risks of bioterrorism increasing or declining two years into the global war on terrorism and nearly half a year into the war to expel Saddam from Iraq? Such risks are difficult to calibrate, as they have varied sources, including not just al Qaeda and "rogue states" but also copycatters, the American militia movement, and loner terrorists.

Uncertainty about the risks fuels debate about the necessary responses. The strategic vocabulary of counterterrorism has had to expand to encompass a set of challenges not previously contemplated. The Bush administration has pursued some radical "solutions," with results as yet uncertain. This symposium will review the evolving risks of bioterrorism in the light of changing circumstances and sketch out key issues in the Bush response.
------------------------------------------------------------ 

Title: On the Robustness of the Predictive Distribution for Sampling from Finite Populations
Speaker: Professor Sudip Bose
Department of Statistics, George Washington University

Date: October 3, 2003
Location: Funger Hall 310
Time: 11:00 a.m.

-------------------------------------------------------------

Suppose that one is interested in sampling from a finite population where items are classified into K categories. Suppose that there is a (prior) probability distribution on the number of items (in the population) that belong to each category. Assume that for different population sizes, these (prior) probability distributions, satisfy a natural and reasonable property, which we call the "generating" property. Then the distribution of a sample selected without replacement from the population does not depend on the population size, and by extension, the conditional distribution of a second sample, given the first sample, does not depend on the population size. The "generating" property is satisfied by probability distributions in the multivariate Polya-Eggenberger family. This has immediate implications for process control or quality control situations, for inspection of errors in financial documents, and so on.

----------------------------------------------------------

Title: Funding Opportunities in the Department of Defense
Speaker: Dr. Wendy Martinez
Office of Naval Research

Date: October 10, 2003
Location: Funger Hall 310
Time: 11:00 a.m.
------------------------------------------------------------ 
This talk will provide information on funding opportunities in the
Department of Defense, including the Navy, Air Force, Army and DARPA.

I will discuss various programs of interest to both new and experienced researchers. 
These include the Young Investigator Program, Multidisciplinary Research Program 
of the URI, sabbatical and summer programs, and many others. I will also highlight 
some of the ONR programs that accept unsolicited research proposals in areas such as
probability, statistics, machine learning, computer science and others.
------------------------------------------------------------ 

Title: Multivariate extensions of Univariate Life Distributions
Speaker: Professor S.P.Mukherjee
Calcutta Universities

Date: October 17, 2003
Location: Funger Hall 310
Time: 11:00 a.m.
------------------------------------------------------------ 
There have been many approaches to deriving multivariate generalizations of univariate life distributions. The one we wanted to accept was the mutivariate extension of some characterizations of univariate distributions that we had derived in respect of Weibull and other distributions. The multivariate extension that we formulated for the joint distribution of the vector of component lives with specified marginals, introducing a common association parameter. The extension includes quite a few multivariate distributions as particular cases. We then verified some closure properties and some characterizations for the univariate case for their validity. Some extension of the approach to a few less often used distributions like the lomax and a finite range distribution.

--------------------------------------------------------------

Title: Conditional U-Statistics with Applications in Discriminant Analysis, ARMA: Processes and Hidden Markov Models

Speaker: Professor Madan L. Puri
              Indiana University, Bloomington, Indiana

Date: October 31, 2003
Location: Funger Hall 310
Time: 11:00 a.m.
------------------------------------------------------------ 
Stute (Ann. Probab. (1991), Ann. Statist. (1994)) introduced a class of conditional U-statistics which generalize the Nadaraya-Watson estimate of a regression function. Under the usual iid set-up, Stute proved the asymptotic normality, weak and strong consistency and the universal consistency of the estimate in the rth mean. Here we extend Stute's results from the independent case to the dependent case. Applications to discriminant analysis, ARMA processes and hidden Markov models are provided. The work is in collaboration with Professor Michel Harel (C.N.R.S. Toulouse, France).
------------------------------------------------------------ 

Title: Nonparametric Bayesian Estimation of Spectral Densities

Speaker: Professor Anindya Roy
              Department of Mathematics and Statistics, University of Maryland, Baltimore County

Date: November 14, 2003
Location: Funger Hall 310
Time: 11:00 a.m.
------------------------------------------------------------ 
We propose a Bayesian approach to estimating the spectral density of a time series. A nonparametric prior on the spectral density is described through Bernstein polynomials. Because the true likelihood is very complicated the posterior is obtained by updating the prior with the Whittle likelihood. We describe a Markov chain Monte Carlo algorithm for sampling from the posterior distribution. We apply our methodology to river flow data from the Missouri river. We also establish consistency of the posterior distribution. In proving consistency we derive contiguity of the exact Gaussian likelihood and the Whittle likelihood.
------------------------------------------------------------ 

Title: Sample Size and Power of Randomized Design

Speaker: Professor Feifang Hu
             Department of Statistics, University of Virginia

Date: November 17, 2003
Location: Funger Hall 310
Time: 11:00 a.m.
------------------------------------------------------------ 

For randomized designs, the power and sample size are usually obtained by ignoring the randomness of the allocation in the literature. But the power is a random variable for a fixed sample size n, when a randomized design is used. In this talk, we focus on the power function (random) and the sample size of two-arm (drug versus control) randomized clinical trials. Based on asymptotic properties, we derive a power function for each fixed sample size. Then a formula of sample size is derived for randomized designs. This formula is applied to several important randomization procedures. Some simulation studies are reported

------------------------------------------------------------ 

Title: "NCI's Biostatistics Grant Portfolio and NIH Funding Mechanism"

Speaker: Dr. Ram Tiwari
             NCI/NIH

Date: December 5, 2003
Location: Funger Hall 310
Time: 11:00 a.m.
------------------------------------------------------------ 

The talk consists of two parts. In Part I, I will talk about our newly released website: www.statfund.cancer.gov , which contains information about a large proportion of NIH's funded grants in Biostatistics. These grants are housed in the Division of Cancer Control and Population Sciences at the National Cancer Institute (NCI). I will also discuss various funding opportunities in (Bio)statistics at NCI. In Part II, I will go over NIH's funding mechanisms and discuss the grant review process at NIH in great detail.

Last Updated: 11/15/03

--------------------------------------------------------------------------------
The contact person is Reza Modarres at Reza@gwu.edu or 202-994-6359.