Center for Fluid Mechanics
And
The Fluids, Thermal and Chemical Processes Group
Of
The Division of Engineering
Seminar Series
Materials and Construction Research Division National Institute of Standards and Technology Gaithersburg, MD | |
Abstract: The rheological properties of suspensions plays an important role in a wide variety of technological processes and is of fundamental interest. In this presentation I will discuss recent advances in the modeling of suspensions based on a novel cellular automata approach called Dissipative-Particle-Dynamics. The first part of this presentation will focus on the validation of this approach (comparison of simulation results with well known theoretical predictions and other computational methods). In the latter part I will present some recent results concerning the yield stress of hard sphere systems with interparticle interactions.
Brown University Division of Biology and Medicine,
Center for Statistical Sciences Seminar
Candidate for Assistant Professor (Tenure Track) in The Public Health Program/Department of Community Health | |
1st FL, Conference Room 106 (Refreshments at 3:45 p.m.) |
Abstract: The directed acyclic graph causal framework has been used with increasing frequency to address concerns of confounding in a variety of research settings. The directed acyclic graph frmework allows for the representation of causal and conterfactual relations amongst variables: the estimation of causal effects through the g-formula; the detection of independencies through the d-separation criteria: and the implementation of algorithms to determine whether conditioning on a particular set of variables, or none at all, is sufficient to control for confounding as well as algorithms which identify such a set of variables. By incorporating notions of monotonic effects, the directed acyclic graph casual framework can be extended so as to allow for the graphical representation of monotonic effects on causal directed acyclic graphs through signed edges; the use of monotonic effects to determine the sign of the causal effect of an intervention in the presence of intermediate variables; the use of monotonic effects to develop rules governing monotonicity and covariance; and finally the use of monotonicity to determine the sign of the bias when control for confounding is inadequate.
Brown University
Center for Computational Molecular Biology Distinguished Lecture
Series
Dana Farber Cancer Institute at Harvard University Emil T. Hofman Professor of Physics at University of Notre Dame | |
Refreshments will be served at 3:45 pm |
Abstract: Post-genomic biology requires us to move beyond the single gene description, and understand the intricate genetic networks that mediate most cellular processes. In the last few years we learned that cellular networks are not random, but their structure carries the signature of self-organizing processes governed by simple but generic laws. The analysis of the metabolic network and the protein interaction network of several organisms indicates that, despite significant variances in their individual components, these networks display identical topologic and scaling properties. The hubs, highly connected nodes common in such networks, have important implications for the cell's robustness and functionality. I will show that cellular networks have a hierarchical architecture, allowing us to identify the organization of the functional modules embedded in the cellular topology.
For more information see http://www.nd.edu/~networks
BIO: Albert-Laszlo Barabasi, currently at the Dana Farber Cancer Institute at Harvard University, is the Emil T. Hofman Professor of Physics at University of Notre Dame. Educated in Bucharest and Budapest, he received a Ph.D. in physics in 1994 from Boston University. After spending a year at IBM T.J. Watson Research Center he joined Notre Dame in 1995. His research has lead to the discovery and understanding of scale-free networks, capturing the structure of many complex networks in technology and nature, from the World Wide Web to the cell. His current research focuses on applying the concepts developed by his group for characterizing the topology of the www and the Internet to uncovering the structural and topological properties of complex metabolic and genetic networks. He is a Fellow of the American Physical Society and an external member of the Hungarian Academy of Sciences. His recent general audience book entitled Linked: The New Science of Networks (Perseus, 2002) is available on ten languages. For more information see http://www.nd.edu/~alb.
Probability/Statistics Seminar
School of Operations Research and Industrial Engineering, Cornell University | |
Visiting Candidate for the junior Probability/Statisatics Position | |
Abstract: In various applications we are interested in the probability of system failure. For instance ruin of an insurance company, large losses in a financial market or long waiting times in a computer network. Such failures might be expressed as the probability that a certain functional of a stochastic process exceeds a high threshold. In this talk I will present the framework of regular variation for stochastic processes and show how it can be applied to derive asymptotic approximations of failure probabilities. We are specifically interested in explaining what the typical extreme sample paths of heavy-tailed stochastic processes looks like. The idea is to describe the extremal behavior of the process in terms of a limiting measure. Then a mapping theorem can be applied to obtain the asymptotic decay of functionals of the sample paths of the process. We will cover heavy-tailed Levy processes, filtered Levy processes, stochastic integrals, as well as large deviations for multivariate random walks with regularly varying steps.
Population Coding Group Seminar
Learning of Action Through Adaptive Combination of Motor Primitive Nature 406, Oct 2000, pp 742-747 | |
Abstract: Understanding how the brain constructs movements remains a fundamental challenge in neuroscience. The brain may control complex movements through flexible combination of motor primitives1, where each primitive is an element of computation in the sensorimotor map that transforms desired limb trajectories into motor commands. Theoretical studies have shown that a system's ability to learn action depends on the shape of its primitives2. Using a time-series analysis of error patterns, here we show that humans learn the dynamics of reaching movements through a flexible combination of primitives that have gaussianlike tuning functions encoding hand velocity. The wide tuning of the inferred primitives predicts limitations on the brain's ability to represent viscous dynamics. We find close agreement between the predicted limitations and the subjects' adaptation to new force fields. The mathematical properties of the derived primitives resemble the tuning curves of Purkinje cells in the cerebellum. The activity of these cells may encode primitives that underlie the learning of dynamics.
PDE Seminar
Brown University Division of Biology and Medicine,
Center for Statistical Sciences Seminar
Candidate for Assistant Professor (Tenure Track) in The Public Health Program/Department of Community Health | |
1st FL, Conference Room 106 (Refreshments at 3:45 p.m.) |
Abstract: We address the problems of estimating a Receiver Operating Characteristic (ROC) curve when no "gold standard" (GS) test is available. We develop a Bayesian methodology for nonparametric estimation of ROC curves used for evaluation of the accuracy of a continuous-scaled diagnostic procedure, by comparing it with an imperfect reference test. MCMC methods are used to compute the posterior estimates of the sensitivities and specificities used to construct the ROC curve. Our approach is flexible and compares favorably with previous proposals in several aspects: (i) the natural monotonicity property of the resulting ROC curve estimate is assured: (ii) no assumption is needed concerning the shape of the distributions of test values of the diseased and non-diseased in these populations; (iii) the imperfect reference test can be based on any scale.
When information on covariates is available, it can be used to increase the effectiveness of continuous markers in distinguishing between healthy and diseased subjects. In a consecutive work we propose a method for estimation of ROC curves to adjust for covariate effects when the true disease states of test individuals are unknown. The covariates may be correlated with the disease process or with the diagnostic testing procedure, or both. We propose a two-part Bayesian model: first, a logistic regression model for disease prevalence is used to fit the covariates; second, a linear model is used to fit the covariates to the distribution of test scores. We used Markov chain Monte Carlo (MCMC) methods to compute the posterior estimates of the sensitivities and specificities that provide the groundwork for inference concerning the diagnostic procedure's accuracy.
We discuss an application to an analysis of ELISA scores in the diagnostic testing of paratuberculosis (Johne's Disease) for several herds of dairy cows.
<--- 2006 Index