Plenary Presentations Our three plenary speakers will be:  Wednesday, 1:15pm-2:15pm Confederation III David Aldous Professor, Department of Statistics University of California at Berkeley Berkeley, CA 94720-3860 Telephone:510-642-3295 E-mail: aldous@stat.berkeley.edu Webpage: http://stat-www.berkeley.edu/users/aldous/ Research Interests: discrete probability models, statistical physics and analysis of algorithms, Markov chains, probabilistic combinatorics, phylogenetic tree models, rare events Talk title: Flows through Random Networks Abstract: While the phrase “flows through random networks” calls to mind a variety of related topics, it is less easy to exhibit any core mathematical theory. The talk will give a theoretician’s overview of existing related theories, of some current research topics, and some programs for future research.  Thursday, 1:15pm-2:15pm - Confederation III Onno Boxma Professor, Department of Mathematics and Computer Science Eindhoven University of Technology HG 9.14 P.O. Box 513 5600 MB Eindhoven The Netherlands Tel: +31 (40) 247 2858 Email: boxma@win.tue.nl Webpage: http://www.win.tue.nl/~boxma/ Research Interests: applied probability, queueing theory, performance analysis of computer, communication, and production systems Talk title: Processor sharing Abstract: A major service mechanism in queueing models is processor sharing (PS). In a PS queue, the service rate is equally shared among all customers present. Thus, when there are n > 0 customers present, each customer receives service at rate 1/n. Originally, the PS paradigm emerged as an idealization of round-robin scheduling mechanisms in time-shared computer systems. In recent years, the PS discipline has attracted renewed interest as a convenient modeling abstraction for bandwidth sharing protocols in high-speed networks. In this talk, various aspects of PS queues are discussed, with an emphasis on sojourn times. We pay special attention to tail asymptotics for sojourn times in PS queues, both for light-tailed and heavy-tailed service requirements. The talk is based on collaborations with Sem Borst, Sindo Nunez-Queija, Regina Egorova, Bert Zwart and others.  Friday, 1:15pm-2:15pm Confederation III Darrell Duffie James I. Miller Professor of Finance Graduate School of Business 518 Memorial Way Stanford University Stanford, CA 94305-5015 Telephone: (650) 723-1976 Fax: (650) 725-7979 Email: duffie@stanford.edu Webpage: http://www.stanford.edu/~duffie/ Research Interests: Incomplete security markets; derivatives markets; financial risk management; capital asset pricing theory; the dynamic spanning role of security markets; preference theory under uncertainty; security design; term structures of interest rates; credit risk modeling, including valuation of corporate and sovereign debt, swaps, and credit derivatives Talk title: Estimating Term Structures of Actual and Risk-neutral Default Probabilities Abstract: The talk will focus on doubly stochastic models of correlated corporate default timing. A model of the term structure of conditional default probabilities is estimated with firm specific and macro-economic data, for all public non-financial U.S. firms. The doubly-stochastic property permits a decomposition of the maximum likelihood estimation problem into separate problems: (1) maximum likelihood estimation of the parameters determining the dependence of stochastic default intensity on covariates, and (2) maximum likelihood estimation of the parameters determining the stochastic evolution of the covariates. The doubly stochastic property is also tested, by virtue of a random time change under which the counting process for total defaults to date is, conditional on the paths of covariates, a constant intensity Poisson process. This talk is based on collaborations with Sanjiv Das, Andeas Eckner, Guillamue Horel, Nikunj Kapasia, Leandro Saita, and Ke Wang.