Dependence Measures for Multivariate Extreme Value Distributions
Neyman Seminar | Wed, September 02, 2009 | 4:00PM - 5:00PM | 1011 Evans Hall
Speaker: Ishay Weissman, Professor of Probability and Statistics, Faculty of Industrial Engineering and Management, Technion-Israel Institute of Technology
The dependence structure of multivariate extremes will be discussed first. Then, two dependence measures will be presented. These measures are suitable for any number of dimensions and are invariant under increasing transformations of the components. They possess an additional desired property, lacked by their competitors, which makes them natural dependence measures for... More >
Real-Time Knowledge Extraction from Massive Time-Series Datastreams
Neyman Seminar | Wed, September 09, 2009 | 4:00PM - 5:00PM | 1011 Evans Hall
Speaker: Joshua Bloom, Associate Professor of Astronomy, Department of Astronomy, UC Berkeley
Classifying entities in any large set of data present computational and algorithmic headaches. But when the data are streaming and, depending on the specific classification, must be acted upon quickly with other (expensive) resources, unique challenges arise. Data collection for "transient" astrophysical phenomena, such as supernova, stellar flares and gamma-ray bursts, has... More >
Traditional Chinese Medicine and Its Challenge to Statisticians
Neyman Seminar | Wed, September 16, 2009 | 4:00PM - 5:00PM | 1011 Evans Hall
Speaker: Wu Xizhi, School of Statistics, Renmin University of China
It contains a brief introduction to the philosophy, methodology, and concept of traditional Chinese medicine and the difference between TCM and its Western counterpart. We describe a statistical method applied to a real data set to predict diseases from symptoms. The obstacles of introducing TCM to the Western world and the difficulties of using classical statistical tools are also discussed. More >
The Costs of Voting: Evidence from a Natural Experiment
Neyman Seminar | Wed, September 23, 2009 | 4:00PM - 5:00PM | 1011 Evans Hall
Speaker: Henry Brady, Professor, Department of Political Science and Department of Public Policy, UC Berkeley
The consolidation of polling places in Los Angeles County during the October 2003 gubernatorial recall election provides a natural experiment on the impact of changing polling places. Overall turnout decreased by a substantial 1.85 percentage points: a drop in polling place turnout of 3.03 percentage points was partially offset by an increase in absentee voting of 1.18... More >
A Bayesian Approach to Transforming Public Gene Expression Repositories into Disease Diagnosis Databases
Neyman Seminar | Wed, September 30, 2009 | 4:00PM - 5:00PM | 1011 Evans Hall
Speaker: Haiyan Huang, Assistant Professor, Department of Statistics, UC Berkeley
The rapid accumulation of gene expression data has offered unprecedented opportunities to study human diseases. The NCBI Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/) is currently the largest database that systematically documents the genome-wide molecular basis of diseases. However, thus far, this resource has been far from fully utilized. In this talk, I will... More >
High-dimensionality effects in the Markowitz problem and other quadratic programs: risk underestimation
Neyman Seminar | Wed, October 07, 2009 | 4:00PM - 5:00PM | 1011 Evans Hall
Speaker: Noureddine El Karoui, Assistant Professor, Department of Statistics, UC Berkeley
It is often the case in statistics and various branches of applied mathematics that one wishes to solve optimization problems involving parameters that are estimated from data. It is therefore natural to try to characterize the relationship between the solution of the optimization problem involving estimated parameters (the sample version) and the solution we would get if we... More >
Time Varying Networks: reverse engineering and analyzing rewiring genetic interactions
Neyman Seminar | Wed, October 14, 2009 | 4:00PM - 5:00PM | 1011 Evans Hall
Speaker: Eric Xing, Associate Professor, Department of Computer Science, Carnegie Mellon University
A plausible representation of the relational information among entities in dynamic systems such as a living cell is a stochastic network that is topologically rewiring and semantically evolving over time. While there is a rich literature in modeling static or temporally invariant networks, until recently, little has been done toward modeling the dynamic processes underlying... More >
Calibration and Prediction Problems in Catchment Scale Hydrology
Neyman Seminar | Wed, October 21, 2009 | 4:00PM - 5:00PM | 1011 Evans Hall
Speaker: Cari Kaufman, Assistant Professor, Department of Statistics, UC Berkeley
Statistical models in environmental applications are making increasing use of the scientific knowledge represented in systems of partial differential equations describing the evolution of environmental processes over time. I will discuss the use of such a model to estimate/predict soil moisture fields under current/future climatic conditions. We use a catchment-level hydrology... More >
Adaptive Sequential Monte Carlo for Computer Models using Computer Experiments and Emulators
Neyman Seminar | Wed, October 28, 2009 | 4:00PM - 5:00PM | 1011 Evans Hall
Speaker: Gardar Johannesson, Applied Statistics & Economics / Engineering Directorate, Lawrence Livermore National Laboratory
As computer model simulations become ever more complex and computationally demanding, so becomes the task of characterizing their accuracy; uncertainty quantification (UQ). Global climate model UQ is a good example. One branch of UQ aims at characterizing input uncertainty and predictive accuracy by comparing model simulations to observations. With a fixed computational budged,... More >
Nonparametric estimation of log-concave densities: some results for one dimension; progress and problems in R^d
Neyman Seminar | Wed, November 04, 2009 | 4:00PM - 5:00PM | 1011 Evans Hall
Speaker: Jon Wellner, Professor, Department of Statistics, University of Washington
I will discuss nonparametric estimation of log-concave densities in $R^1$ and $R^d$. In the case of $R^1$, I will present limit theory for the estimators at fixed points at which the population density has a non-zero second derivative and for the resulting natural mode estimator under a corresponding hypothesis. In the case of $R^d$ with $d \ge 2$ will briefly discuss... More >
Tractable performance bounds for compressed sensing
Neyman Seminar | Wed, November 18, 2009 | 4:00PM - 5:00PM | 1011 Evans Hall
Speaker: Alexandre d'Aspremont, Assistant Professor, Department of Operations Research & Financial Engineering, Bendheim Center for Finance, Princeton University
Recent results in compressed sensing show that, under certain conditions, the sparsest solution to an underdetermined set of linear equations can be recovered by solving a linear program. These results either rely on computing sparse eigenvalues of the design matrix or on properties of its nullspace. So far, no tractable algorithm is known to test these conditions and most... More >
Supervised topic models
Neyman Seminar | Wed, December 02, 2009 | 4:00PM - 5:00PM | 1011 Evans Hall
Speaker: Jon McAuliffe, Adjunct Assistant Professor, Department of Statistics, UC Berkeley
The scale of contemporary electronic text collections has led to growing interest in statistical models based on so-called topics. Formally, a topic is a probability distribution over a vocabulary. Informally, a topic is intended to capture an underlying semantic theme. Most topic models are unsupervised: only the words in the documents are modelled. I will describe... More >