Keynote Speakers

Arnaud Doucet arnaud doucet photo

Department of Statistics, Oxford University, U.K.

Bio: Arnaud is Professor of Statistics at the University of Oxford. He received his Ph.D. in engineering from University of Paris XI in 1997. After completing his Ph.D., he joined Cambridge University as postdoc. He then took up permenant positions at the University of Melbourne, Cambridge University, University of British Columbia and the Institute of Statistical Mathematics. He has been professor at Oxford since 2011.

His research interests include Bayesian statistics; Stochastic Simulation; Sequential Monte Carlo; Markov Chain Monte Carlo; Time Series. He is primarily interested in the development and study of novel Monte Carlo methods for inference in complex stochastic models.

He is a former Canada research chair in stochastic computation at UBC and fellow of Hertford College at the University of Oxford. He has authored over 70 per-reviewed publications with over 26000 citations. He has advised over 20 doctoral students and post-docs with more than 10 holding faculty positions at various world-class universities across the world.

Title: On a New Class of Pseudo-Marginal Algorithms

Time: 18 Dec, 9:00-10:00

Venue: UTown Auditorium 2

Abstract: The use of unbiased estimators within the Metropolisā€”Hastings has found numerous applications in Bayesian statistics. The resulting so-called pseudo-marginal algorithm allows us to deal with intractable likelihood functions which are unbiasedly estimated using, for example, importance sampling or particle filters. However, recent theoretical results have established that the computational cost of pseudo-marginal methods is for many common applications of order TĀ² for T data points at each iteration. This cost is prohibitive for large datasets. I will present new procedures which can provably significantly reduce it. On various applications, the efficiency of computations is increased by several orders of magnitude.

 

Yongdai Kim

Department of Statistics, Seoul National University, South KoreaKim

Bio: Yongdai Kim is Professor of Statistics at Seoul National University. He received his Ph.D. in Statistics from Ohio State University. After completing his Ph.D., he worked at National Institues of Health as a biostatistician. In 1999, he came back to Korea and took up permanent positions are Hankuk University of Foriegn Studies and Ewha Wonmans Univesirty. IN 2004, he joined Seoul National University as an assistant professor and was promoted to full professor in 2011.

His research interests include Bayesian nonparametrics, Markov Chain Monte Carlo, Regularized methods and Variable Selection for High-dimensional Regression model and optimization algorithms.

He is currently the chair of Department of Statistics, Seoul National University. He authored over 50 peer-reviewed publications. He first proved the Bernstein-von Mises theorem for a certain semiparametric Bayesian model in 2004 and the oracle property for ultra-high dimensional models in 2008. He has supervised more than 10 doctoral students and post-docs.

Title: On model selection for ultra-high dimensional models

Time: 17 Dec, 9:20-10:20

Venue: UTown Auditorium 2

Abstract: Model selection is one of the most important topics for ultra-high dimensional models where the number of covariates is much larger than the sample size. Various methods including penalized regression approached and information criteria have been proposed. In this talk, first I review what have been done so far for model selection on ultra-high dimensions, and explain what have not been done yet. Then, I will introduce recent results including fast computation and data-adaptive information criterion.

 

Samuel Kou

Department of Statistics, Harvard University, U.S.A. Sam Kou 05

Bio: Samuel Kou is Professor of Statistics at Harvard University. He received his Ph.D. in statistics from Stanford University in 2001 under the supervision of Professor Bradley Efron. After completing his Ph.D., he joined Harvard University as an Assistant Professor. He was promoted to full professor in 2008. 

His research interests include stochastic inference in single molecule biophysics, chemistry, and biology; Bayesian inference for stochastic models; nonparametric statistical methods; model selection and empirical Bayes methods; Monte Carlo methods; and economic and financial modeling.

He is the recipient of the COPSS (Committee of Presidents of Statistical Societies) Presidents' Award; a U.S. National Science Foundation CAREER Award; the Raymond J. Carroll Young Investigator Award; the Institute of Mathematical Statistics Richard Tweedie Award; and the American Statistical Association Outstanding Statistical Application Award. He is an elected Fellow of the American Statistical Association, an elected member of the International Statistical Institute, and a Medallion Lecturer and an elected Fellow of the Institute of Mathematical Statistics
.

Title: Big data, Google and disease detection: the statistical story

Time: 19 Dec, 11:20-12:20

Venue: UTown Auditorium 2

Abstract: Big data collected from the internet have generated significant interest in not only the academic community but also industry and government agencies. They bring great potential in tracking and predicting massive social trends or activities. We focus on tracking disease epidemics in this talk. We will discuss the applications, in particular Google Flu Trends, some of the fallacy and the statistical implications. We will propose a new model that utilizes publicly available online data to estimate disease epidemics. Our model outperforms all previous real-time tracking models for influenza epidemics at the national level of the US. We will also draw some lessons for big data applications.