Research Focus Areas

Biological Sciences Applications

Management and analysis of information about biological systems. The goal is often to identify and classify biological organisms.

Computational Statistics

Rapid prototyping of new computational methods to apply to complex problems. These are typically MCMC-based methods, and the problems often involve massive data sets with troublesome characteristics.

Information Integration Technology (IIT)

Developing a framework of methodologies to support decision-making under uncertainty. The framework allows for diverse information sources (models, test data, expert judgment, etc.) and merges techniques from several disciplines (statistics, graph theory, knowledge representation, etc.) to develop performance estimates for complex systems.

Monte Carlo Methods

Extending traditional Monte Carlo methods, e.g. through the use of biasing (importance sampling) techniques to improve convergence in simulations of time-dependent physical processes.

Reliability

Investigations into system performance and availability and how they change, e.g. with time or with improved or degraded materials or processes. Determining optimal experimental design is often the first step, and then we analyze information from those experiments and other sources, such as operational data and computer models.

Statistical Population Bounding

Determining bounds that contain a desired fraction of a given population, as opposed to confidence limits, which bound the mean of a population. Characteristics we consider include aging, multiple populations, assessment of measurement process, and bounds on probabilities.

Statistical Computer Model Evaluation

Developing methods for uncertainty quantification, in order to assess the quality of mathematical predictions from computer models. We seek to determine the degree to which prediction and real outcomes can be expected to differ, and to find causes for the difference.