wustl.jpg           Contact: Todd Kuffner, email: kuffner@wustl.edu                 


Workshop Description

The Second Workshop on Higher-Order Asymptotics and Post-Selection Inference (WHOA-PSI)^{2} seeks to build upon the success of the first workshop, by presenting the latest developments in post-selection inference, and discussing how tools from higher-order asymptotics can both elucidate important properties of post-selection inference procedures, as well as suggest new directions which may ultimately yield more accurate small-sample performance. The workshop format is intended to encourage collaboration and lively discussion, and to give a voice to all participants with online discussion forums (a result of a successful experiment from the first workshop). More specific details will soon be posted below.

This conference supports the Non-Discrimination Statement of the Association for Women in Mathematics (AWM).

We are grateful to have the financial sponsorship of the Dean of the Faculty of Arts & Sciences and the Department of Mathematics at Washington University in St. Louis.

Latest News (23rd February, 2017): Registration is now open. You can go to the registration page by clicking here. More information about registration is below. At this time, our application for funding to support junior participants is still pending.

Organizing Committee

Todd Kuffner lead organizer
Washington University in St. Louis, Assistant Professor, Dept. of Mathematics
John Kolassa
Rutgers University, Professor, Dept. of Statistics and Biostatistics
Ryan Tibshirani
Carnegie Mellon University, Associate Professor, Dept. of Statistics

Confirmed Speakers

Tentative Speakers

Registration Information

Clicking here will take you to the registration page. The registration fee is $300 and is nonrefundable. Registration includes meeting fees, breakfasts, lunches and coffee breaks on all 3 days, as well as a banquet dinner on Saturday 12th August. Registration will close either when the room capacity is reached, or on Wednesday, 19th July, at 11:59PM, Central Standard Time. You can see how many places are remaining on the registration page.

Dates and Times

The workshop is a full 3 days. The talks will begin around 8am on Saturday August 12th, and will end by 5pm on Monday August 14th.

Lodging Information

All workshop participants will have to make their own lodging arrangements. The Knight Center is holding blocks of rooms at a special rate ($119/night) for the nights of Fri 8/11, Sat 8/12, Sun 8/13 and Mon 8/14. Guests will need to call the Knight Center direct at 314-933-9400 or toll free 866-933-9400 and ask for rooms under the room block:  WHOA-PSI Workshop
Other nearby hotels include: (1) The Moonrise Hotel (http://moonrisehotel.com), which is within walking distance. (2) Clayton Plaza Hotel (http://www.cpclayton.com), which offers a free shuttle service.

Potential Topics include (but are certainly not limited to):   Participants: feel free to send me updates!

Principles and general views of post-selection inference, for example
Benjamini (2010). `Simultaneous and selective inference: current successes and future challenges', Biometrical Journal 52, 708-721.
Taylor & Tibshirani (2015), `Statistical learning and selective inference', Proceedings of the National Academy Sciences 112, 7629-7634.
Leeb & Potscher (2005), `Model selection and inference: facts and fiction', Econometric Theory 21, 21-59.

Comparisons of post-selection inference procedures, for example
Leeb, Potscher & Ewald (2015), `On various confidence intervals post-model-selection', Statistical Science 30, 216-227.

Incorporating resampling and asymptotic refinements into inference procedures relevant for this workshop, for example

Stephen M.S. Lee and Yilei Wu (2017). Resampling-based post-model-selection inference for linear regression models.
McCarthy, Zhang, Brown, Berk, Buja, George & Zhao (2017). Calibrated Percentile Double Bootstrap for Robust Linear Regression Inference, Statistica Sinica, accepted.

Post-selection inference and selective inference, for example
Hyun, G'Sell & Tibshirani (2016), `Exact post-selection inference for changepoint detection and other generalized lasso problems', arXiv: 1606.03552
Taylor & Tibshirani (2016), `Post-selection inference for L1-penalized likelihood models', arXiv: 1602.07358
Fithian, Taylor, Tibshirani & Tibshirani (2015+), `Selective sequential model selection', arXiv: 1512.02565
Tibshirani, Taylor, Lockhart, Tibshirani (2015+), `Exact post-selection inference for sequential regression procedures', J. Amer. Statist. Assoc., to appear.
Lockhart, Taylor, Tibshirani & Tibshirani (2014), `A significance test for the lasso', Annals of Statistics 42, 413-468.
Tibshirani, Rinaldo, Tibshirani & Wasserman (2015), `Uniform asymptotic inference and the bootstrap after model selection', arXiv: 1506.06266
Tian & Taylor (2015), `Asymptotics of selective inference', arXiv: 1501.03588
Lee, Sun, Sun & Taylor (2015), `Exact post-selection inference with the lasso', to appear in the Annals of Statistics.

Simultaneous inference, false discovery rates (FDR), false coverage statement rates (FCR), family-wise error rates (FWER), for example
Berk, Brown, Buja, Zhang & Zhao (2013), `Valid post-selection inference', Annals of Statistics 41, 802-837.
Benjamini (2010), `Discovering the false discovery rate', J. Roy. Statist. Soc. Ser. B 72, 405-416.
Benjamini & Yekutieli (2005), `False discovery rate-adjusted multiple confidence intervals for selected parameters', J. Amer. Statist. Assoc. 100, 71-93.
G'Sell, Wager, Chouldechova & Tibshirani (2015+), `Sequential selection procedures and false discovery rate control', J. Roy. Statist. Soc. Ser. B, to appear.
Barber & Candes (2015), `Controlling the false discovery rate via knockoffs', Annals of Statistics 43, 2055-2085.
Su, Bogdan & Candes (2016+), `False discoveries occur early on the Lasso path', arXiv: 1511.01957.

Bayesian post-selection inference, for example
Panigrahi, Taylor & Weinstein (2016). `Bayesian post-selection inference in the linear model', arXiv: 1605.08824
Yekutieli (2012). `Adjusted Bayesian inference for selected parameters', J. Roy. Statist. Soc. Ser. B, 74(3), 515-541.

Bagging and Boosting, for example
Bradic (2016). `Randomized maximum-contrast selection: subagging for large-scale regression', Elec. J. Statist. 10(1), 121-170.
Li & Bradic (2015). `Boosting in the presence of outliers: adaptive classification with non-convex loss functions', arXiv: 1510.01064.
Efron (2014), `Estimation and accuracy after model selection', J. Amer. Statist. Assoc. 109, 991-1007.
Buhlmann & Yu (2002), `Analyzing bagging', Annals of Statistics 30, 927-961.

High-dimensional inference, for example
Fan, Shao & Zhou (2015), `Are discoveries spurious? Distributions of Maximum Spurious Correlations and their applications', arXiv: 1502.04237
Cai & Guo (2015), `Confidence intervals for high-dimensional linear regression: minimax rates and adaptivity', arXiv: 1506.05539
Ning & Liu (2015), `A general theory of hypothesis tests and confidence regions for sparse high dimensional models', arXiv: 1412.8765
Ning, Zhao & Liu (2015), `A likelihood ratio framework for high dimensional semiparametric regression', arXiv: 1412.2295
Shah & Samworth (2013), `Variable selection with error control: another look at stability selection', J. Roy. Statist. Soc. B 75, 55-80.
Meinshausen & Buhlmann (2010), `Stability selection', J. Roy. Statist. Soc. Ser. B 72, 417-473.
van de Geer, Buhlmann, Ritov & Dezeure (2014), `On asymptotically optimal confidence regions and tests for high-dimensional models', Annals of Statistics 42, 1166-1202.
Javanmard & Montanari (2015+), `Hypothesis testing in high-dimensional regression under the Gaussian random design model: asymptotic theory', IEEE Trans. Inform. Theory, to appear.
Liu & Yu (2013), `Asymptotic properties of Lasso+mLS and Lasso+Ridge in sparse high-dimensional linear regression', Electronic J. Statist. 7, 3124-3169.
Zhang & Zhang (2014), `Confidence intervals for low-dimensional parameters in high-dimensional linear models', J. Roy. Statist. Soc. Ser. B 76, 217-242.
Belloni, Chernozhukov & Hansen, `Inference methods for high-dimensional sparse econometric models', Advances in Economics & Econometrics, Econometric Society World Congress 2010.

Selection and inference for weak signals, for example
Shi & Qu (2016). `Weak signal identification and inference in penalized model selection', Annals of Statistics, to appear.
Jeng (2016). `Detecting weak signals in high dimensions', J. Multivariate Statist. 147, 234-246.

The aspects of the above topics and other post-selection inference procedures which will be emphasized in the workshop are those related to higher-order asymptotics, including both analytic- and resampling-based tools and refinements, some of which are described in:
Some recent references for post-selection inference include
Chapter 3 of Fithian (2015), Topics in Adaptive Inference, Ph.D. thesis, Stanford University.
Chapter 6 of Hastie, Tibshirani & Wainwright (2015), Statistical Learning with Sparsity: The Lasso and Generalizations, Chapman & Hall.
Chapters 10-11 of Buhlmann & van de Geer (2011), Statistics for High-Dimensional Data, Springer.

Local Information

For those arriving early or thinking about staying longer, St. Louis is a lovely place to visit. Besides the iconic Gateway Arch and the nearby Old Courthouse which houses exhibits on the Dred Scott case, St. Louis has a stunning botanical garden, a high density of good restaurants (BBQ is a specialty), and is close to many rivers (Missouri, Mississippi and Meremac) which are great for float trips. There are many nearby parks and nature reserves which are excellent for hiking, as well as a wolf sanctuary. Mark Twain's boyhood home lies an hour north of the city. Anheuser-Busch is headquartered in St. Louis and offers tours of the brewery (requires advance booking due to popularity). The workshop takes place one week before the second scheduled US Presidential Election debate, which will be held on campus. For those unfamiliar with the institution, Washington University in St. Louis is a leading national research university, ranked 23rd in the world in the 2016 Academic Ranking of World Universities. Our statistics presence is concentrated in the Dept. of Mathematics. You are encouraged to look around this beautiful campus on the western edge of St. Louis, which faces Forest Park, the site of the 1904 World's Fair and home to the Saint Louis Zoo and Saint Louis Art Museum (both free admission, walking distance from campus). For baseball fans, the Pittsburgh Pirates are in town the weekend of the workshop, playing the St. Louis Cardinals at Busch Stadium.