Colloquia

Fall 2020

All colloquia will be held virtually

Information about past colloquia is available here.

Date
Speaker
Affiliation
Title
Wednesday, September 9 Ying Wei Columbia University  Integrated Quantile RAnk Test (iQRAT) for gene-level associations
Wednesday, September 16 Tahir Ekin Texas State University Augmented Probability Simulation Methods for Decisions and Games
Wednesday, September 23 Sumanta Basu Cornell University Sparse Identification and Estimation of Large-Scale Vector Autoregressive Moving Averages
Wednesday, September 30 Vladimir Pozdnyakov University of Connecticut Discretely Observed Brownian Motion Governed by a Telegraph Process: Estimation
Wednesday, October 7 Ted Westling University of Massachusetts Joint UConn/UMass Colloquium

Causal Inference with Continuous Exposures

Wednesday, October 14 Nan Laird Harvard T.H. Chan School of Public Health The 25th Pfizer/ASA/UConn Distinguished Statistician Colloquium
Wednesday, October 21 Anru Zhang University of Wisconsin-Madison Statistical Learning for High-dimensional Tensor Data
Wednesday, October 28 Dong-Yun Kim NIH/NHLBI Continuous Monitoring in a Clinical Trial: A Fully Sequential Approach
Wednesday, November 4 Steven Culpepper University of Illinois, Urbana-Champaign Inferring latent structure in polytomous data
Wednesday, November 11 Sofer, Tamar Harvard Medical School Estimating heritability and genetic correlations with (interpretable) method of moment estimators of variance components
Wednesday, November 18 Tiandong Wang Texas A & M University A Directed Preferential Attachment Model with Poisson Measurement
Friday, November 20 Bengt Muthen University of California Recent Advances in Latent Variable Modeling
Wednesday, December 2 Yuping Zhang University of Connecticut  Statistical Learning and Inference for Biological Networks

Ying Wei; Professor, Department of Biostatistics, Mailman School of Public Health, Columbia University

Wednesday, September 9, 2020 4:00 pm

Integrated Quantile RAnk Test (iQRAT) for gene-level associations

Genetic association studies often evaluate the combined group-wise effects of rare and common genetic variants on phenotypes at gene level. Many approaches have been proposed for group-wise association tests, such as the widely used burden tests and sequence kernel association tests with sequencing data. Most of these approaches focus on identifying mean effects. As the genetic associations are complex, we propose an efficient integrated rank test to investigate the genetic effect across the entire distribution/quantile function of a phenotype. The resulting test complements the mean-based analysis and improve efficiency and robustness. The proposed test integrates the rank score test statistics over quantile levels while incorporating Cauchy combination test scheme and Fisher’s method to maximize the power. It generalized the classical quantile-specific rank-score test. Using simulations studies and real Metabochip data on lipid traits, we investigated the performance of the new test in comparison with the burden tests and sequence kernel association tests in multiple scenarios. This is joint work with Tianying Wang and Iuliana Ionita-Laza.

Bio: Ying Wei is a statistician and a Professor of Biostatistics in the Columbia University Mailman School of Public Health, working primarily on quantile regression, semiparametric models of longitudinal data, and their applications.

Wei earned her Ph.D. in Statistics from the University of Illinois at Urbana–Champaign in 2004 and has been a faculty member of Biostatistics at Columbia University, and also an affiliated member of the Data Science Institute ever since.

In 2011, Wei received the Noether Young Scholar Award of the American Statistical Association, “for outstanding early contributions to nonparametric statistics.” In 2015, Wei was elected as a Fellow of the American Statistical Association. Wei is also an elected member of the International Statistical Institute. In 2020 she was named as a Fellow of the Institute of Mathematical Statistics “for contributions to the development, dissemination, and application of mathematical statistics”.

Event address for attendees: 

https://uconn-cmr.webex.com/uconn-cmr/onstage/g.php?MTID=e6c6243ddbe5366041dad514f5a20ac48 

Date and time: 

Wednesday, September 9, 2020 4:00 pm
Eastern Daylight Time (New York, GMT-04:00) 
Duration: 

  1 hour

Tahir Ekin; Associate Professor CIS & Quantitative Methods, Texas State University

Wednesday, September 16, 2020 4:00 pm

Augmented Probability Simulation Methods for Decisions and Games

Expectation function-based decision and game theoretic models require both computation of the utility function and its optimization. This can be computationally challenging especially in cases with continuous and multi-modal sources of uncertainty or complex objective function surfaces. We propose augmented simulation approaches that treat the decision variable(s) as random and construct an augmented distribution in the space of both decisions and random variables. Simulation from this distribution simultaneously solves for the expectation of the objective function and optimization problem. In doing so, we sample more frequently from the marginal decision space in that the objective function has higher values in a maximization problem. This talk introduces augmented probability simulation and its extensions to solve for stochastic programming problems and game theoretic models. There will be a discussion and illustration on a variety of applications such as news-vendor type models, service systems and cybersecurity. 

Bio: Dr. Tahir Ekin is an Associate Professor of Quantitative Methods in McCoy College of Business, Texas State University. His research interests include analytical applications in health care fraud assessment and decision modeling under uncertainty. His book on health care fraud analytics titled “Statistics and Health Care Fraud: How to Save Billions” has been recently published. His work has appeared in a variety of journals including Journal of the Royal Statistical Society Series C, International Statistical Review, Naval Research Logistics and Decision Analysis among others. Dr. Ekin holds a Ph.D. in Decision Sciences from The George Washington University, and a B.S. in Industrial Engineering from Bilkent University, Turkey. He is an elected member of International Statistical Institute and currently serves as Vice President of the International Society of Business and Industrial Statisticians.

Event address for attendees: https://uconn-cmr.webex.com/uconn-cmr/onstage/g.php?MTID=e867ed9f762b20bfb0d97fd2f2786b24d
Date and time:  Wednesday, September 16, 2020 4:00 p.m.
Duration:    1 hour

Sumanta Basu, Assistant Professor, Shayegani Bruno Family Faculty Fellow, Cornell University

Wednesday, September 23, 2020 4:00 pm

Sparse Identification and Estimation of Large-Scale Vector AutoRegressive Moving Averages

The Vector AutoRegressive Moving Average (VARMA) model is fundamental to the theory of multivariate time series; however, identifiability issues have led practitioners to abandon it in favor of the simpler but more restrictive Vector AutoRegressive (VAR) model. We narrow this gap with a new optimization-based approach to VARMA identification built upon the principle of parsimony. Among all equivalent data-generating models, we use convex optimization to seek the parameterization that is “simplest” in a certain sense. A user-specified strongly convex penalty is used to measure model simplicity, and that same penalty is then used to define an estimator that can be efficiently computed.  We establish consistency of our estimators in a double- asymptotic regime.  Our non-asymptotic error bound analysis accommodates both model specification and parameter estimation steps, a feature that is crucial for studying large-scale VARMA algorithms and is largely unexplored in existing literature.  Our analysis also provides new results on infinite-order VAR, elastic net estimation under a singular covariance structure of regressors, and new concentration inequalities for quadratic forms of random variables from Gaussian time series, which may be of independent interest.  We illustrate the advantage of our method over VAR alternatives on simulated and real data examples.

Event address for attendees: https://uconn-cmr.webex.com/uconn-cmr/onstage/g.php?MTID=ea79d2aa6520c0522c9ce17cbf75a5613
Call-in option: US Toll +1-415-655-0002 Access code: 120 644 5968
Date and time:  Wednesday, September 23, 2020 4:00 p.m.
Duration:    1 hour

Vladimir Pozdnyakov, Professor, Department of Statistics, University of Connecticut

Wednesday, September 30, 2020 4:00 pm

Discretely Observed Brownian Motion Governed by a Telegraph Process: Estimation

A Brownian motion whose infinitesimal variance alternates according to a telegraph process is considered. This stochastic process can be employed to model variety of real-word situations. In this work we applied our findings for animal movement analysis. The main goal is to develop an estimation procedure for underlying model parameters when the Brownian Motion governed by telegraph process is observed discretely. Resulting sequence of observations is not Markov. But since the location-state process is Markov, the likelihood estimation can be done with help of Hidden Markov Model tools. Further extensions of the model are discussed. More specifically, we consider (1) introducing an additional hidden state and (2) incorporating measurement errors into the model.  (joint work with Chaoran Hu, Mark Elbroch, Tom Meyer, and Jun Yan)

Event address for attendees: https://uconn-cmr.webex.com/uconn-cmr/onstage/g.php?MTID=e7ef41045207955e8c0476f607757b306
Call-in option: US Toll +1-415-655-0002 Access code: 120 157 9936
Date and time:  Wednesday, September 30, 2020 4:00 p.m.
Duration:    1 hour

Ted Westling, Assistant Professor, Department of Mathematics & Statistics
University of Massachusetts, Amherst

Wednesday, October 7, 2020, 2:00 pm EST. Duration – 1 hour

Causal Inference with Continuous Exposures

Much of the literature on estimating causal effects concerns discrete exposures. Recently, there has been increased interest in continuous exposures; that is, exposures that can take an uncountable number of values. Examples of such exposures include air pollution, pre-vaccination antibody responses, and concentrations of harmful chemicals in the blood. In this talk, I will provide an introduction to the area of causal inference with continuous exposures. I will then provide an overview of some of the recent research concerning nonparametric causal inference with continuous exposures, including my own recent and ongoing research. In particular, I will discuss approaches to nonparametric pointwise and global inference on causal dose-response curves, and, time permitting, inference on alternative causal parameters such as the effects of stochastic and incremental interventions.

Event address for attendees: https://uconn-cmr.webex.com/uconn-cmr/onstage/g.php?MTID=ed71956e994f2ed58b1e820b41491b293
Call-in option: US Toll +1-415-655-0002 Access code: 120 878 4802
Date and time:  Wednesday, October 7, 2020, 2:00 pm EST
Duration:   1 hour

Nan Laird ; Professor Emerita, Harvard T.H. Chan School of Public Health

Wednesday, October 14, 2020 3:00-5:00 pm

The 25th Pfizer/ASA/UConn Distinguished Statistician Colloquium

The department of Statistics at the University of Connecticut is honored to announce the 25th Distinguished Statistician Colloquium. The Pfizer colloquium series ran from 1978 until 2012 and was renewed in 2018. The colloquium series featured C. R. Rao, Bradley Efron, D.R. Cox, Grace Wahba and many more. For a complete list, see https://stat.uconn.edu/pfizer-colloquium/. The purpose of the Colloquium is to provide a forum for a distinguished statistician to share and disseminate their unique perspective and work in the theory and/or application of statistics. Starting from 2018, the series has been co-sponsored by Pfizer, the American Statistical Association, and the Department of Statistics at the University of Connecticut. 

 

This year’s speaker at the 25th Pfizer Distinguished Statistician Series will be Professor Emerita Nan Laird from Harvard T.H. Chan School of Public Health. This year colloquium will be virtual and will take place on Wednesday, October 14th, 2020 from 3:00-5:00pmEST. The interview will be conducted by Christoph Lange (Harvard T.H. Chan School of Public Health) and Joseph Hogan (Brown University). 

The lecture and the interview are open to the public but pre-registration will be needed for a link to the webcast.  The event registration website is:   https://uconnuecs.cvent.com/Pfizer2020   

 

If you have questions, please contact, Juliet Kapsis at juliet.kapsis@uconn.edu or conferences@uconn.edu. 

 

We thank Pfizer and the ASA for their generous financial support. We also thank the members of the selection committee – Dan Meyer and Demissie Alemayehu from Pfizer, Ron Wasserstein and Nancy Flournoy from the ASA, and Dipak Dey (Chair), Joseph Glaz and Ming-Hui Chen from UConn. Professor Chen also represents the New England Statistical Society (NESS). 


Anru Zhang, Assistant Professor, Department of Statistics, University of Wisconsin-Madison

Wednesday, October 21, 2020, 4:00 pm EST. Duration – 1 hour

Statistical Learning for High-dimensional Tensor Data

The analysis of tensor data has become an active research topic in this area of big data. Datasets in the form of tensors, or high-order matrices, arise from a wide range of applications, such as financial econometrics, genomics, and material science. In addition, tensor methods provide unique perspectives and solutions to many high-dimensional problems, such as topic modeling and high-order interaction pursuit, where the observations are not necessarily tensors. High-dimensional tensor problems generally possess distinct characteristics that pose unprecedented challenges to the data science community. There is a clear need to develop new methods, efficient algorithms, and fundamental theory to analyze the high-dimensional tensor data.

In this talk, we discuss some recent advances in high-dimensional tensor data analysis through the consideration of several fundamental and interrelated problems, including tensor SVD and tensor regression. We illustrate how we develop new statistically optimal methods and computationally efficient algorithms that exploit useful information from high-dimensional tensor data based on the modern theories of computation, high-dimensional statistics, and non-convex optimization.

Event address for attendees: https://uconn-cmr.webex.com/uconn-cmr/onstage/g.php?MTID=ea437cb5d4f43c942ad7ae3ddfdf6c7b9
Call-in option: US Toll +1-415-655-0002 Access code: 120 110 6907
Date and time:  Wednesday, October 21, 2020 4:00 p.m. EST
Duration:   1 hour

Dong-Yun Kim, PhD., Office of Biostatistics Research, National Institutes of Health/National Heart, Lung, and Blood Institute

Wednesday, October 28, 2020, 4:00 pm EST. Duration – 1 hour

Continuous Monitoring in a Clinical Trial: A Fully Sequential Approach

In this talk we introduce Sequential Patient Recruit Monitoring (SPRM), a new and efficient accrual monitoring method. Built on a sequential probability ratio test using Woodroofe boundaries, SPRM is an evidence-based, decision-support tool that provides opportunity for corrective action in a timely manner. Suitable for the modern, centralized data management environment, it requires minimal effort to maintain. The method can easily accommodate protocol changes involving sample size and/or recruitment period modification.

We also propose Sequential Event Rate Monitoring (SERM), a new continuous monitoring method for the event rate of time-to-event data in a clinical trial. SERM gives an early warning if the target rate is unlikely to be achieved by the end of study. Since SERM is designed to monitor the overall event rate, blindness of the trial is preserved. If necessary, the method could suggest the number of extra recruitments required for the planned number of primary events. It can also be used to estimate an extension of the follow-up time. We illustrate the methods using data from a well-known Phase III clinical trial.

Bio: Dr. Kim is a mathematical statistician at the Office of Biostatistics Research within National Heart, Lung, and Blood Institute (NHLBI), Bethesda, Maryland. She received PhD in Statistics from the University of Michigan, Ann Arbor in 2003. Before joining NIH in 2013, she held a faculty position at Virginia Tech. Her research interests include: sequential methods in clinical trials, change-point inference, statistical genetics and bioinformatics. At NIH, she has been involved in large NHLBI-sponsored clinical trials and intramural projects, provided scientific reviews of protocols and expert opinions for these projects, and developed novel methodologies for applications in clinical trials. Also, Dr. Kim has years of experience in collaborative research in other areas including mobile health, bioengineering, health services, and environmental science.

Event address for attendees: https://uconn-cmr.webex.com/uconncmr/
onstage/g.php?MTID=e17fdfdfc6a976e3b467ff3fcea6e3348
Call-in option: US Toll +1-415-655-0002 Access code: 120 261 2606
Date and time:  Wednesday, October 28, 2020 4:00 p.m. EST
Duration:   1 hour

Steven Andrew Culpepper, Associate Professor, Department of Statistics, Beckman Institute for Advanced Science and Technology,University of Illinois at Urbana-Champaign

Wednesday, November 4, 2020 4:00 p.m. Duration – 1 hour

Inferring latent structure in polytomous data

Researchers continue to develop and advance latent structure models (LSMs) for research in the social and behavioral sciences. LSMs provide researchers with a framework for providing a fine-grained classification of respondents into substantively meaningful latent classes. Recent research developed confirmatory LSMs for polytomous response data; however, these methods require detailed knowledge about the relationship between the multivariate latent attributes and observed responses. This study advances existing methods by proposing new exploratory methods for inferring the latent structure underlying polytomous response data. We discuss new sufficient conditions for ensuring model parameter identifiability. A novel Bayesian formulation is presented using a variable selection algorithm. An application to the 2012 Programme for International Student Assessment (PISA) problem-solving vignettes is presented to demonstrate the method. We conclude with a brief overview of future research directions for broadening the application of LSMs in social, behavioral, and health research.
Event address for attendees: https://uconn-cmr.webex.com/uconn-cmr/onstage/g.php?MTID=ea5bf20f02e18941dffc2edc0fa6760cb
Call-in option: US Toll +1-415-655-0002 Access code: 120 248 8279
Date and time:  Wednesday, November 4, 2020 4:00 p.m. EST
Duration:   1 hour

Tamar Sofer, PhD, Associate Biostatistician, Brigham and Women’s Hospital, Assistant Professor, Harvard Medical School

Wednesday, November 11, 2020 4:00 p.m. Duration – 1 hour

Estimating heritability and genetic correlations with (interpretable) method of moment estimators of variance components

In genetic association studies, heritability is the proportion of phenotypic variance explained by genetic, and genetic correlation measures the correlation between the additive genetic components underlying two phenotypes. We develop and study moment estimators of heritability and genetic correlation and show that they are easily interpreted as generalizations of standard estimators of residual variance and Pearson correlation in datasets of independent individuals. I will provide examples from data analysis
Event address for attendees: https://uconn-cmr.webex.com/uconn-cmr/onstage/g.php?MTID=ee1fd6e7cd0d0152bfdaf203d9514f99e
Call-in option: US Toll +1-415-655-0002 Access code: 120 496 2492
Date and time:  Wednesday, November 11, 2020 4:00 p.m. EST
Duration:   1 hour

Tiandong Wang, Associate Professor, Department of Statistics, Texas A & M University

Wednesday, November 18, 2020 4:00 p.m. Duration – 1 hour

A Directed Preferential Attachment Model with Poisson Measurement

When modeling a directed social network, one choice is to use the traditional preferential attachment model, which generates power-law tail distributions. In a traditional directed preferential attachment, every new edge is added sequentially into the network. However, for real datasets, it is common to only have coarse timestamps available, which means several new edges are created at the same timestamp. Previous analyses on the evolution of social networks reveal that after reaching a stable phase, the growth of edge counts in a network follows a non-homogeneous Poisson process with a constant rate across the day but varying rates from day to day. Taking such empirical observations into account, we propose a modified preferential attachment model with Poisson measurement, and study its asymptotic behavior. This modified model is then fitted to real datasets, and we see it provides a better fit than the traditional one.
Event address for attendees: https://uconn-cmr.webex.com/uconn-cmr/onstage/g.php?MTID=e8e39873843a2c9e7f281b8a40a57bc12
Call-in option: US Toll +1-415-655-0002 Access code: 120 112 4316
Date and time:  Wednesday, November 18, 2020 4:00 p.m. EST
Duration:   1 hour

Bengt Muthen, Professor Emeritus, University of California, Los Angeles

Friday, November 20, 2020 11:30am -1:00pm

Recent Advances in Latent Variable Modeling


This talk gives an overview of some recent and ongoing latent variable research. Borrowing ideas from multilevel factor analysis, longitudinal SEM in a single-level, wide format is formulated in a new way that finds a well-fitting model 45 years after the writing of the classic Wheaton, Muthen, Alwin, and Summers article. This segues into a generalization of latent transition analysis using the multilevel notion of a random intercept while staying in a single-level, wide format. Turning back to multilevel modeling, the talk considers time series analysis of intensive longitudinal data. This is illustrated by intervention data on electricity consumption and a randomized intervention related to positive and negative affect where cycles play a major role. Finally, the new feature in Mplus Version 8.5 of Bayesian analysis of count, nominal, and binary logit models is presented.

This session is jointly sponsored by the Statistics department and the Research Methods, Measurement, and Evaluation program as part of the Statistical Applications and Quantitative Research Methods colloquium series.
Event address for attendees: https://uconn-cmr.webex.com/uconn-cmr/j.php?MTID=mcffb6bf89878dde1bdb8af0904d19714
Call-in option: Meeting number: 120 655 8452
Password: Mplus
490721dc38f74a50a49d3d752dcfd363Join by video systemDial 1206558452@uconn-cmr.webex.comYou can also dial 173.243.2.68 and enter your meeting number.Join by phone
+1-415-655-0002 US Toll
Access code: 120 655 8452
Date and time:  Friday, Nov 20, 2020 11:30 am | (UTC-05:00) Eastern Time (US & Canada)
Duration:   1 hour 30 minutes

Yuping Zhang, Associate Professor, Department of Statistics, University of Connecticut

Wednesday, December 2, 2020 4:00 p.m. EST, 1-hour duration

Statistical Learning and Inference for Biological Networks

Networks play important roles in elucidating biological systems and complex diseases. It is necessary to develop new methods for network learning and inference based on high-dimensional, heterogeneous and complex biological data generated by modern high-throughput biotechnologies. This talk will feature two new methodologies: 1) a statistical framework for joint learning of multiple heterogeneous exponential Markov Random Fields and its associated variational inference and theory, and 2) statistical inference of network disturbances leveraging topologies and multi-view biological data.
Event address for attendees: https://uconn-cmr.webex.com/uconn-cmr/j.php?MTID=mef718a3bcb13df44454012171bd2d7ce
Date and time:  Wednesday, December 2, 2020 4:00 p.m. EST

Duration:  1-hour duration