7PMVNF

t
*TTVF
 IMS Bulletin

August 2016 Your new Council members

CONTENTS We are pleased to con!rm the results of the "#$% IMS Council elections. &e President- 1 Election results Elect is Alison Etheridge. &e !ve new Council members are Jean Bertoin, Songxi Chen, Elizaveta Levina, Simon Tavaré and Cun-Hui Zhang. 2–3 Members’ News: Jogesh Babu, Joseph Hilbe, David &e new Council members and President-Elect will serve IMS for three years, starting van Dyk, Clifford Spiegelman, o'cially at the IMS Business Meeting, held this year at the World Congress in Toronto. Rebecca Doerge, Dipak Dey, &ey will join the following Council members: Peter Bühlmann, Florentina Bunea, Alyson Wilson, Bin Yu Geo(rey Grimmett, Aad van der Vaart and Naisyin Wang (who are on Council for a 4 IMS Fellows further year), and Andreas Buja, Gerda Claeskens, Nancy Heckman, Kavita Ramanan and Ming Yuan (whose terms last until August "#$)). &e new Council members will replace 5 Call for Nominations; New Rick Durrett, Ste(en Lauritzen, Susan Murphy, Jonathan Taylor and Jane-Ling Wang. Researchers Group Mixer IMS Council also recently appointed a new Treasurer, Zhengjun Zhang. Zhengjun 6 Student Puzzle 15 is a professor in the Department of Statistics at the University of Wisconsin–Madison. 7 Profile: Larry Wasserman &erefore, starting at the Business Meeting, the new IMS Executive Committee will be Jon Wellner (President), Richard Davis (Past President), Alison Etheridge (President- 8 Special Lectures: Sara van de Geer, Bin Yu, Nanny Elect), Zhengjun Zhang (Treasurer), Judith Rousseau (Program Secretary), and Aurore Wermuth, Vanessa Didelez, Delaigle (Executive Secretary). Former President Erwin Bolthausen and Treasurer Jean Peter Diggle Opsomer will leave the Executive Committee this year. We express our gratitude to all those who give their time and energy to assist the IMS. 12 New IMS Monograph &anks too, to those who also stood for election this year. And !nally, thanks to all the 13 Project Euclid news IMS members who took the time to vote! 14 Recent papers: Bayesian Analysis; Brazilian Journal of Probability and Statistics

15 Vlada’s Point 16 Treasurer’s Report 2015 18 Meetings 23 Employment Opportunities Alison Etheridge Jean Bertoin Songxi Chen 24 International Calendar President-Elect Council 2016–19 Council 2016–19 27 Information for Advertisers

Read it online at bulletin.imstat.org Elizaveta Levina Simon Tavaré Cun-Hui Zhang Council 2016–19 Council 2016–19 Council 2016–19 IMS Bulletin 2 . IMS Bulletin Volume 45 . Issue 5 Volume 45 r Issue 5 August 2016 IMS Members’ News ISSN 1544-1881 International Astrostatistics Association Spiegelman receives Don Owen Award Contact information IMS Fellow G. Jogesh Babu of Pennsylvania &e "#$% Don Owen Award, given by the IMS Bulletin Editor: Anirban DasGupta State University, and member Joseph M. ASA’s San Antonio Chapter, was presented Assistant Editor: Tati Howell Hilbe of Arizona State University, have been to Clifford Spiegelman at the ,%th annual Contributing Editors: Robert Adler, Peter awarded the International Astrostatistics Conference of Texas Statisticians. Cli( is Bickel, Stéphane Boucheron, David Hand, Vlada Limic, Xiao-Li Meng, Dimitris Association (IAA) Outstanding a professor at Texas A&M University. He Politis, Terry Speed and Hadley Wickham Contributions to Astrostatistics medal, the earned his doctoral degree in statistics from top award given to members of the global Northwestern University in $+-%; taught Contact the IMS Bulletin by email: astrostatistics and astroinformatics commu- at Florida State, Northwestern, and Johns e [email protected] w http://bulletin.imstat.org nity by the IAA. Both were also elected IAA Hopkins; and served as a scientist at the https://www.facebook.com/IMSTATI fellows, as was another IMS fellow, David National Bureau of Standards for nine years van Dyk of Imperial College, London. before joining Texas A&M in $+)-. Contact the IMS regarding your dues, membership, subscriptions, orders or &e International Astrostatistics change of address: Association was founded as an independent Doerge receives ACE Fellowship IMS Dues and Subscriptions Office scienti!c association for astrostatistics Rebecca W. Doerge, the Trent and Judith 9650 Rockville Pike, Suite L3503A and astroinformatics in "#$", developing Anderson Distinguished Professor of Bethesda, MD 20814-3998 USA from the International Statistical Institute Statistics and President’s Fellow for Big Data t 877-557-4674 [toll-free in USA] astrostatistics committee and network. &e and Simulation at Purdue, has been awarded t +1 216 295 5661[international] goal of the association from its outset has Fellow of American Council of Education f +1 301 634 7099 been to foster collaboration between statis- (ACE) for "#$%–$-. Each university e [email protected] ticians and astronomers. It also has a goal of nominates only one candidate who shows encouraging the production of educational promise of being an academic leader for ACE Contact the IMS regarding any other matter, including advertising, copyright books, articles, white papers, and tutorials in fellowship. Read the list of fellows at http:// permission, offprint orders, copyright statistics for the bene!t of the astronomical www.acenet.edu/news-room/Pages/ACE- transfer, societal matters, meetings, fellows community. See http://iaa.mi.oa-brera.inaf.it/ Fellows-Class-of-2016-17.aspx. nominations and content of publications: Executive Director, Elyse Gustafson Dipak Dey appointed Editor-in-Chief of Sankhya IMS Business Office Professor Dipak K. Dey, Board of Trustees Distinguished Professor and Associate Dean of the PO Box 22718, Beachwood OH 44122, USA College of Liberal Arts and Sciences at the University of Connecticut, has been appointed t 877-557-4674 [toll-free in USA] Editor-in-Chief of the prestigious scienti!c journal Sankhya, the Indian journal of statistics. t +1 216 295 5661[international] Dipak is a former editor of the IMS Bulletin, and is a Fellow of IMS, ISBA and ASA, and f +1 216 295 5661 an elected member of the ISI. He is a past-president of the International Indian Statistical e [email protected] Association and in "#$* received the Outstanding Statistician Award from the Connecticut chapter of the ASA. Executive Committee Sankhya is published by &e Indian Statistical Institute. &e journal was founded by President: Richard Davis Prasanta Chandra Mahalanobis in $+,, and is now published in two series, A and B, in [email protected] collaboration with Springer. All submissions to are now online, using: President-Elect: Jon Wellner Sankhya [email protected] For Series A: http://www.springer.com/statistics/journal/13171 Past President: Erwin Bolthausen For Series B: http://www.springer.com/statistics/journal/13571 [email protected] Sankhya publishes research articles in the broad areas of theoretical statistics, probability Treasurer: Jean Opsomer and applied statistics. Series A primarily covers theoretical statistics and probability (including [email protected] stochastic processes). Series B primarily covers all areas of applied statistics (including applied Program Secretary: Judith Rousseau [email protected] probability, applied stochastic processes, econometrics and statistical computing). Reviews Executive Secretary: Aurore Delaigle and discussion articles in areas of current research activity are also published. Each volume has [email protected] four parts: Series A issues are in February and August, Series B in May and November. = access published papers online

August . 2016 IMS Bulletin . 3 IMS Journals and Publications Annals of Statistics: Ed George and Tailen Hsing http://imstat.org/aos http://projecteuclid.org/aos Annals of Applied Statistics: Tilmann Gneiting IMS Members’ News http://imstat.org/aoas http://projecteuclid.org/aoas Army Wilks Award to Alyson Wilson Annals of Probability: Maria Eulalia Vares http://imstat.org/aop &is year’s Army Wilks Award winner is Alyson Wilson, professor of statistics at North http://projecteuclid.org/aop Carolina State University. &e award—established to commemorate the career of Samuel Annals of Applied Probability: Bálint Tóth http://imstat.org/aap S. Wilks and his service to the Army—is given to a deserving individual who has made a http://projecteuclid.org/aoap substantial contribution to statistical methodology and application a(ecting the practice or Statistical Science: Peter Green application of statistics to problems in defense and security. &e Army Wilks Award is given http://imstat.org/sts http://projecteuclid.org/ss periodically at the Conference on Applied Statistics in Defense (CASD), which was held IMS Collections October $+–"", "#$., at George Mason University in Fairfax, Virginia. http://imstat.org/publications/imscollections.htm http://projecteuclid.org/imsc IMS Monographs and IMS Textbooks: David Cox Bin Yu to give keynote at Women in Statistics and Data Science conference http://imstat.org/cup/ &e Conference for Women in Statistics and Data Science (WSDS), organised by the ASA, is IMS Co-sponsored Journals and to be held October "#–"" in Charlotte, North Carolina. WSDS "#$% will bring hundreds of Publications statistical practitioners and data scientists together in celebration of women in statistics and Electronic Journal of Statistics: Domenico Marinucci data science. &e focus is to empower women statisticians, biostatisticians, and data scientists http://imstat.org/ejs http://projecteuclid.org/ejs by exchanging ideas and presenting technical talks on important, modern, and cutting-edge Electronic Journal of Probability: Brian Rider research; discussing how to establish fruitful multidisciplinary collaborations; and showcasing http://ejp.ejpecp.org the accomplishments of successful women professionals. Electronic Communications in Probability: Registration is open now, and the housing deadline is September "#. Conference registra- Sandrine Péché http://ecp.ejpecp.org tion ends October *. Travel awards are available: apply by August ). Find out more about the Current Index to Statistics: George Styan conference at https://ww2.amstat.org/meetings/wsds/2016/index.cfm http://www.statindex.org log into members' area at imstat.org One of the featured speakers at WSDS is Bin Yu (along with Cynthia Clark, Stacy Journal of Computational and Graphical Statistics: Lindborg and Wendy Martinez). Bin was interviewed about her career in interdisciplin- Diane Cook ary statistics for Amstat News: you can read the article at http://magazine.amstat.org/ http://www.amstat.org/publications/jcgs log into members' area at imstat.org blog/2016/06/01/wsds16/ Statistics Surveys: Donald Richards In the interview, Bin talks about four of her current projects. One of these, a long-term http://imstat.org/ss collaboration with biologists Erwin Frise and Sue Celniker of Lawrence Berkeley National http://projecteuclid.org/ssu Probability Surveys: Ben Hambly Lab (LBNL) that uses novel spatial gene expression data to understand how organs are formed http://imstat.org/ps in the fruit-/y, was pro!led in the Berkeley Lab news center (Mapping a Cell’s Destiny: New http://www.i-journals.org/ps/ http://newscenter.lbl. Berkeley Lab Tool Speeds Discovery of Spatial Patterns in Gene Networks, IMS-Supported Journals gov/2016/05/04/mapping-cells-destiny/). Bin has been working on this with her students Siqi ALEA: Latin American Journal of Probability and Wu and Karl Kumbier and former postdocs Statistics: Victor Perez Abreu http://alea.impa.br/english Antony Joseph and Siva Balakrishnan; they Annales de l'Institut Henri Poincaré (B): Gregory also work with Wei Xu’s computer science Miermont, Christophe Sabot http://imstat.org/aihp team at Tshinghua University to scale up the http://projecteuclid.org/aihp Bayesian Analysis: Bruno Sansó computations by building upon open-source http://ba.stat.cmu.edu platforms Spark and Fiji. Bernoulli: Holger Dette She says, “&is is my favorite data science http://www.bernoulli-society.org/ http://projecteuclid.org/bj project since it represents an iterative knowl- Brazilian Journal of Probability and Statistics: edge discovery process that is complete with Francisco Louzada Neto http://imstat.org/bjps wet-lab knockout experiments, statistical and http://projecteuclid.org/bjps Stochastic Systems: Assaf Zeevi machine learning methodology development, BerkeleyLab Berkeleyand UC Wu, Siqi Credit http://www.i-journals.org/ssy/ and so0ware development for other groups to go a0er heterogeneous building blocks hidden in their data, spatial or not. &is project also motivated exciting theoretical work on dictionary IMS-Affiliated Journals Probability and Mathematical Statistics: K. Bogdan, learning. &e theoretical study has made us go back to practice for the next step of devising M. Musiela, J. Rosiński, W. Szczotka, & W.A. Woyczyński uncertainty measures. It would not have been possible without my amazing student, Siqi Wu.” http://www.math.uni.wroc.pl/~pms 4 . IMS Bulletin Volume 45 . Issue 5

IMS Fellows 2016

Congratulations to the  new IMS Fellows elected this year! They will be presented at the IMS Presidential Address and Awards session at the World Congress, on Monday, July  at :pm.

Marek Biskup, for excellent research in the Nina Gantert, for her in/uential interface of probability theory and statistical contributions in the study of motion in physics. random media.

Emery Neal Brown, for exceptional Elizaveta Levina, for fundamental contributions to mathematical and statistical contributions to high-dimensional statistics, modeling and analysis of neuroscience data; particularly to network modeling. and for outstanding leadership and service to the profession.

Herold Dehling, for fundamental Eyal Lubetzky, for fundamental contributions at the interface of probability contributions to the cut-o( phenomenon and statistics, particularly to asymptotics and the dynamics of the Ising model. for dependent data structures and empirical processes.

Mathias Drton, for outstanding contributions Nicolai Meinshausen, for exceptional to mathematical statistics, particularly to contributions to high-dimensional and graphical modeling and algebraic statistics computational statistics, causal inference with algorithmic applications. and widely recognized research in climate sciences.

Lutz Duembgen, for outstanding Eric Moulines, for his excellent contributions contributions to mathematical statistics, in asymptotic theory for time-dependent particularly to nonparametrics, shape- models, particularly to hidden Markov constrained inference and empirical models and stochastic algorithms. processes; and for his exceptional leadership and service to the profession.

Alison Etheridge, for outstanding research Carl Eric Mueller, for fundamental and on measure-valued stochastic processes and in/uential contributions in stochastic partial applications to population biology; and di(erential equations and measure-valued for international leadership and impressive di(usions; and for his important services to service to the profession. the community. August . 2016 IMS Bulletin . 5

Peiyong Annie Qu, for her in/uential Harry van Zanten, for excellent and contributions to estimating equations, sustained contributions to the frequentist semiparametric inference on correlated asymptotic properties of Bayesian procedures. data; and for her outstanding service to the profession.

Marc A. Suchard, for outstanding Hao Helen Zhang, for her in/uential contributions at the interface between contributions to nonparametric statistics, applied mathematics, statistics and feature selection, high-dimensional statistics, epidemiology, particularly to computational and machine learning and data mining; and methodology for stochastic processes, for her excellent service to the profession. Bayesian statistics, and population genetics.

Call For Nominations: C.R. and Bhargavi Rao Prize for Outstanding Research in Statistics &e C.R. and Bhargavi Rao Prize was established to honor and recognize outstanding and in/uential innovations in the theory and practice of mathematical statistics, international leadership in directing statistical research, and pioneering contributions by a recognized leader in the !eld of statistics. &e Rao Prize is awarded by the Department of Statistics at Penn State University to a nominee selected by the members of the Rao Prize Committee. C.R. Rao, Emeritus Professor of Statistics at Penn State, held the Eberly Chair in Statistics at Penn State from $+)) to "##$. Previous Rao Prize recipients are: Bradley Efron, Jayaram Sethuraman, Lawrence D. Brown, Peter J. Bickel, James O. Berger, Herman Cherno( and Sir David Cox. For additional information, see If you’re attending JSM in Chicago, come http://stat.psu.edu/rao-prize. Nominations for the "#$- Rao Prize should be submitted by and meet other early-career researchers at December ,  by email to [email protected] or by regular New Researchers Group Mixer mail to: Chair, Rao Prize Selection Committee, !"# $omas Building, Penn State University, University Park, PA %#&'"-"%%%. on Monday, August !, "#!$ &e Rao Prize shall be awarded in odd numbered years. &e award recipient will receive a medal, a cash prize, and an invitation to visit (om $:%#–&:## pm Penn State and give a talk as part of a day-long workshop. Joliet Nominations should include a letter describing the nominee’s in the Room in the Hilton outstanding contributions to leadership and research in statistics, a current curriculum vita, and two supporting letters. 6 . IMS Bulletin Volume 45 . Issue 5

Student Puzzle Corner 15

It is the turn of a problem on probability this time. We will consider a problem that looks like a Student members of problem on analysis. Many of you know that analysis and probability share a strong synergistic the IMS are invited to relationship; there are a number of classic texts on how analysis and probability feed into each submit solutions (to [email protected] other. The problem will be left slightly open ended to whet your imagination. Here is the exact with subject “Student problem of this issue: Puzzle Corner”). The deadline is August , (a) Let f be a given function on the unit interval [, ]. Define now a sequence of functions fn by _k _k+ _k _k+ . The names and the rule fn(x) = The average value of f over the interval [n , n ], if x [n , n ]. affiliations of (up to) Deadline AugustWhat is the weakest sufficient condition you can provide under which fn(x) f(x) for almost the first  student all x? Give a proof of your claim. members to submit correct solutions, and (b) For extra credit only: Fix an > . Is it true that for some set B with ∫Bdx < , the answer to the prob- supxB |fn(x)−f(x)| ? That is, is it true that in fact outside of a set of arbitrarily lem, will be published small measure, fn will converge uniformly to f? in the next issue of the  Bulletin. The Editor’s decision is final.

Solution to Puzzle 14

Editor Anirban DasGupta writes: Γ(α + β + n) 2 n n+α−2 β−2 Γ(α+3n) Γ(α+β+n) = Γ(α + n) Γ(β) ∫6 p p (2−p) dp = Γ(α+n) Γ(α+β+3n)

By Stirling’s approximation to Γ(x) as x → ∞, and using the nota- tion ~ to mean that the ratio of the two sequences converges to 2, we now get −α−3n α 3n 3n+α −α−β−n α+β n n+α+β θ ~ e (2+3n ) (3n) e (2+ n ) n ~ 3−β. n −α−β−3n α+β 3n 3n+α+β −α−n α n n+α e (2+ 2n ) (3n) e (2+n ) n Tom Berrett Promit Ghosal We had at Oxford University, at Notice that the limit of θn does not depend on α; this is because Columbia University, and Haozhe Zhang at the Iowa State only the local behavior of g near p = 2 matters for the limiting behav-

University [above, le)–right] send us correct answers to both parts of ior of θn. In particular, if p has a uniform prior, then the limit of θn is the previous puzzle; congratulations to all of them. Let us recall the ½ : you ought to be ;6:;6 about the question will our Sun last for the

problem. Suppose given p, X2,X3, · · · are iid Bernoulli with parameter next ../ billion years? Very interesting! p. Let p have a prior distribution with a Lebesgue density g. Suppose Consider now the case of a general prior density g with in!nitely

θn = Pg(Xn+2 = · · · = X3n = 2 | X2 = · · · = Xn = 2) denote the Bayesian smooth local behavior at p = 2. Let k be de!ned as the unique non- predictive probability that the next n trials will all be successes if negative integer such that g(2) = · · · = g(k−2)(2) = 6, g(k)(2) ≠ 6. &en, 2 3n the previous have all been so. &en, !nd the limit of the sequence as in the case of the special Beta prior, θ = ∫6 p g(p)dp. n n 2 n ∫6 p g(p)dp θn when g is a general Beta density with parameters α, β, and for a general g which is in!nitely smooth at p = 2. Using Watson’s Lemma, and once again, Stirling’s approximation, 2 n (−2)k g(k)(2) Γ(n+2) Γ(k+2) (−2)k g(k)(2) First consider the case of a Beta prior. In that case, the posterior ∫6 p g(p)dp ~ ~ . k! Γ(n+k+3) nk+2 distribution of p given that X2 = · · · = Xn = 2 is Beta with parameters n + α and β. &erefore, 2 (3n)k+2 θ = E [P(X = · · · = X = 2 |X = · · · = X = 2, p)] Hence, θ ~ = 3−k−2. n p |X2=···=Xn=2 n+2 3n 2 n n 2 = E (pn) nk+2 p |X2=···=Xn=2 August . 2016 IMS Bulletin . 7

Profile: Larry Wasserman

As we reported in the last issue, Larry Wasserman was among five IMS Fellows recently elected to the US National Academy of Sciences. We’ll be writing about them them in the coming issues, and to start us off, Stephen E. Fienberg from Carnegie Mellon University, writes a profile about his colleague and friend:

In May, Larry A. Wasserman was elected to membership in the US National Academy of Sciences. Born in Windsor, Ontario, Larry attended the University of Toronto for both his BSc and PhD. He joined the Department of Statistics at Carnegie Mellon University (CMU) as a postdoctoral fellow in $+)) following completion of his PhD thesis on belief functions, written under the direction of Rob Tibshirani. His thesis was later recognized by the Pierre Robillard Award of the Statistical Society of Canada. Larry quickly rose through the faculty ranks and has remained at CMU, becoming a mainstay in both the Department of Statistics and the Machine Learning Department. Currently, he is also a member of the McWilliams Center for Cosmology and one of the conveners of the Topological Statistics and Statistical Machine Learning groups. In $+++, Larry received the Presidents’ Award of the Committee of Presidents of Statistical Societies, for the Outstanding Statistician Under the Age of *#, and three years later, in "##", the Centre de Larry Wasserman (left, pictured with Ed George) presented the Recherches Mathematique de Montreal-Statistical Society of Canada 2013 IMS Rietz Lecture at JSM in Montreal Prize in Statistics. A Fellow of the American Statistical Association, the IMS, and the American Association for the Advancement of agent provocateur, sometimes tongue in cheek but most o0en with Science, Larry delivered the "#$, IMS Rietz Lecture, on topological purpose and insight. I o(er two examples. inference. His essay in the recently publish COPSS volume, $e Rise of Larry’s early research at CMU led to his recognition as a leading the Machines, begins: “Statistics is the science of learning from data. Bayesian innovator who has integrated insights into the foundations Machine Learning (ML) is the science of learning from data. &ese of inference with incisive and applicable methodological contribu- !elds are identical in intent although they di(er in their history, tions. But increasingly over time he challenged conventional Bayesian conventions, emphasis and culture. &ere is no denying the success thinking and argued for frequency-styled calibration of statistical and importance of the !eld of Statistics for science and, more gener- tools. His contributions range from de!nitive treatments of Bayesian ally, for society. I’m proud to be a part of the !eld. &e focus of this robustness and modern nonparametric estimation, mixture models, essay is on one challenge (and opportunity) to our Field: the rise of multiple testing, privacy, and causal inference, and highly successful Machine Learning.” collaborations with astrophysicists and statistical geneticists. His In a "#$" essay on refereeing, Larry begins: “Our current peer legions of collaborators come from across his departments at CMU review is an authoritarian system resembling a priesthood or a guild. and across multiple !elds and countries. It made sense in the $%##’s when it was invented. Over ,## years later A proli!c contributor to the theoretical and methodological we are still using the same system. It is time to modernize and democ- statistics literature, Larry is also an expositor par excellence and the ratize our approach to scienti!c publishing.” He goes on to propose author of two highly praised textbooks: All of Statistics (the winner of a system of open publication with posted reviews and commentary, the "##. DeGroot Prize from the International Society for Bayesian including rebuttals. Analysis) and All of Nonparametric Statistics. He is a gi0ed lecturer Larry is a great colleague, one who is pushing the boundaries of and his Statistical Learning class in the machine Learning Department our !eld and who is constantly challenging accepted wisdom and is legendary both for its insightfulness and for its demands on the conventional thinking. He does so with zest, humor, and great insight. students. Yet they /ock to attend. His election to the National Academy of Sciences recognizes these Beyond his technical and expository contributions, Larry is an contributions, as well as his more traditional ones. 8 . IMS Bulletin Volume 45 . Issue 5

Preview of Wald lectures: Sara van de Geer

Sara van de Geer is Professor of Statistics in the Department of Mathematics, ETH Zürich. Her work focuses on mathematical statistics, for example, theory for M-estimators in high/infinite dimensions, adaptation to unknown sparsity, semi-paramet- ric theory, confidence sets in high-dimensional models, and concentration of measure for high-dimensional and nonparametric problems. She has (co-)authored four monographs, most recently lecture notes for the Saint-Flour Probability Summer School. She was a council member of the Swiss National Science Foundation –, and is President of the Bernoulli Society -. She is a Knight in the Order of Orange-Nassau, a member of the German Academy of Sciences Leopoldina, and a correspondent of the Dutch Royal Academy of Sciences. Sara’s three Wald Lectures will be given at the World Congress in Toronto, on July ,  and .

High-dimensional statistics: a triptych

High-dimensional statistics concerns the examples: the 2-norm, norms generated from estimator under sparsity conditions and show situation where the number of parameters p cones, the sorted 2-norm, the nuclear norm that it reaches the lower bound. is (much) larger than the number of observa- for matrices and an extension to tensors. In the third lecture, we examine the tions n. &is is quite common nowadays, and We then show sharp oracle inequalities for a null space property for sparsity inducing it has led to the development of new statis- broad class of loss functions. norms. &e null space property ensures exact tical methodology. &ese lectures present a &e second lecture addresses the recovery of certain sparsity patterns and is selected overview of mathematical theory for construction of asymptotic con!dence moreover a key ingredient for oracle results. sparsity inducing methods. intervals for parameters of interest. Here, We derive this property for the Gram matrix In the !rst lecture we will highlight the we restrict ourselves to the linear and the based on n copies of a p-dimensional random main ingredients for proving sharp oracle graphical model. We prove asymptotic nor- variable X , where we require moment condi- inequalities for regularized empirical risk mality of de-biased estimators. We consider tions for !nite dimensional projections of X minimizers. &e regularization penalty will asymptotic lower bounds for the variance or the more general small ball property. be taken to be a norm Ω on p-dimensional of an approximately unbiased estimator of &e lectures are based on joint work with Euclidean space. Important is that the a one-dimensional parameter as well as Le Andreas Elsener, Jana Janková, Alan Muro norm Ω is has a particular feature which we Cam-type lower bounds. We ascertain the and Benjamin Stucky. term the triangle property. We present as approximate unbiasedness of the de-biased Rietz lecture preview: Bin Yu

Bin Yu is Chancellor’s Professor in the Departments of Statistics and of Electrical Engineering & Computer Science at the University of California at Berkeley. Her current research interests focus on solving high-dimensional data problems through developments of statistics and machine learning methodologies, algorithms, and theory. Her group is engaged in interdisciplinary research with scientists from genomics, neuroscience, and medicine. She is a Member of the National Academy of Sciences and Fellow of the American Academy of Arts and Sciences. She was a Guggenheim Fellow in , and President of IMS in –. She was an ICIAM Invited Lecturer in  and Tukey Memorial Lecture of Bernoulli Society in . She is a Fellow of IMS, ASA, IEEE and AAAS. She has served or is serving on numerous journal editorial boards, including JMLR, AOS and JASA, and committees of BMSA of NAS, SAMSI, IPAM and ICERM. Bin will deliver this Rietz Lecture at the World Congress on July .

Theory to gain insight and inform practice Henry L. Rietz, the !rst president of IMS, published his book familiarity with the grammar of mathematics, can secure a satisfactory Mathematical Statistics in $+"-. One reviewer wrote in $+"), understanding of the points involved.” “Professor Rietz has developed this theory so skillfully that the In this lecture, I would like to promote the good tradition of ‘workers in other !elds’, provided only that they have a passing mathematical statistics as expressed in Rietz’s book in order to gain Continues on page  August . 2016 IMS Bulletin . 9

Bin Yu’s Rietz Lecture Preview Continued (om page  insight and inform practice. down the path of seeking con!dence measures of the learned dictio- In particular, I will recount the beginning of our theoretical study nary elements (corresponding to biologically meaningful regions in of dictionary learning (DL) as part of a multi-disciplinary project Drosophila embryo). to “map a cell’s destiny” in Drosophila embryo. I will share insights Finally, I will present preliminary work using our con!dence mea- gained regarding local identi!ability of primal and dual formulations sures to identify potential knockout (or gene editing) experiments in of DL. Furthermore, comparing the two formulations is leading us an iterative interaction between biological and data sciences.

Medallion lecture preview: Nanny Wermuth

Nanny Wermuth is professor emerita at the Division of Mathematical Statistics, Chalmers University of Technology in Gothenburg, Sweden, and affiliated professor at the Medical School of the Gutenberg University in Mainz, Germany. Among her main research interests are multivariate statistical models and their properties, especially graphical Markov models, as well as their applications in the life sciences and in the natural sciences. She has served the profession as, among other positions, president of the International Biometric Society and of the IMS. Nanny’s Medallion Lecture will be at JSM in Chicago, on Monday, August , at :am. (There are two related sessions that Nanny also highlights below, immediately before and after the Medallion session.)

Tracing Pathways of Dependence—How far did we get? Tracing pathways of dependence to under- graph, which may contain directed and undi- In addition to my lecture, there will stand development was a main aim of genet- rected edges and an ordering of the nodes as be two further sessions [on the same day], icist Sewall Wright when he formulated - a given, for instance, by time in a hypothesised concerning graphical Markov models, with century ago - linear data generating processes, generating process. eight speakers who received their PhD within represented them by directed graphs and We now have a subclass, named ‘trace- the last ten years. evaluated the !t to his data. able regressions’, most suitable to model &e !rst is Graphical Markov Models: Path analysis was generalised with development in ordered single and joint Extending, Combining, and Evaluating structural equation models by concentrating responses together with structure in a set of Concepts. &is session is organized by still on linear relations but including joint context variables. &e main di'culties were Giovanni Maria Marchetti (University responses and hidden variables. Intensive to !nd ($) testable properties of generated of Florence), with speakers Po-Ling Loh discussions of special aspects were reported distributions, which needed to concentrate (University of Pennsylvania, Philadelphia), in some of the Berkeley symposia from $+*. on conditional dependences, in addition to Kayvan Sadeghi (University of Cambridge), to $+-". Markov structure, and (") an operator for Piotr Zwiernik (Pompeu Fabra University, A di(erent extension, named graphical graphs which re/ect e(ects of ignoring some Barcelona) and Bala Rajaratnam (Stanford Markov models, started to be developed of the variables or of conditioning on !xed University), and in the late $+-#s by using Andrej Markov’s levels of other variables, or of both combined. &e other one is Recent Enhancements concept of conditional independence to &ese are essential for combining evidence of Graphical Markov Models. &is session de!ne missing edges in graphs, which consist from related, di(erent studies. is organized by me, and features speakers of nodes representing variables and of edges In this lecture, I will try to summarise Monia Lupparelli (University of Bologna), allowing conditional dependences. &ese some of the important insights obtained in Robin Evans (Oxford University), Caroline models permit variables of any type and the last *# years for traceable regressions and Uhler (MIT, Cambridge) and Elias hence more complex than linear relations. to point towards what I regard as important Bareinboim (Purdue University, West Conditioning sets are de!ned by the type of open questions. Lafayette). 10 . IMS Bulletin Volume 45 . Issue 5

Medallion lecture preview: Vanessa Didelez

Vanessa Didelez has just moved to Germany from the University of Bristol, UK, to be Professor of Statistics and Causal Inference at the Leibniz Institute, University of Bremen. Her research focuses on graphical models and causal inference especially in time-dependent settings, and encom- passes aspects of statistics, epidemiology, philosophy and computer science. She obtained her PhD in  from the University of Dortmund under supervision of Iris Pigeot and was appointed Lecturer in Statistics at University College London where she started collaborating with Philip Dawid on a decision theoretic approach to causality. In  she stayed at the Norwegian Centre for Advanced Study establishing an ongoing exchange with Odd Aalen and his group on continuous-time causality. She is known for her contributions to the understanding of Mendelian randomisation as an instrumental variable approach to causal inference in epidemiology. Vanessa’s lecture will be on Thursday, July  at the World Congress in Toronto. Causal Reasoning for Events in Continuous Time We o0en make statements such as event A how causal inference based on inverse theoretic approach for sequential decisions was the cause of event B; most statistical probability weighting (IPW), well known for in longitudinal settings and use a graphical causal inference literature would translate longitudinal data (Robins et al., "###), can representation with in/uence diagrams that this into two binary random variables, and be extended to the continuous-time situation include decision nodes; speci!cally causal use structural equations, causal DAGs and/or using a martingale approach. validity is analogous to the extended stability potential outcomes in order to formalise the In the work that I will present at the of Dawid & Didelez ("#$#). &is provides di(erence between causation and association. Medallion lecture (joint work with Kjetil an explicit representation of the target of An aspect that is o0en only implicit is that Røysland, Odd Aalen and &eis Lange), we inference as well as allowing us to use simple of temporality: event A can only be a cause start by de!ning causal validity of local inde- graphical rules to check identi!ability. While of B if it happens earlier—it therefore seems pendence graphs in terms of interventions, it is common to phrase causal queries in more natural to adopt a stochastic process which in the context of events in time take terms of potential outcomes or counterfactu- approach instead. As a concrete example, a the form of modi!cations to the intensities als, it is worth emphasising that the decision public health authority may want to know of speci!c processes, e.g. a treatment process; theoretic framework and the way we de!ne whether home visits by nurses to elderly causal validity is given if the speci!cation causal validity for events and stochastic pro- patients should be more or less frequent with of the dynamic system is rich enough to cesses do not require these. &e question at view to subsequent hospitalisation events model such an intervention. &is is similar hand is simply phrased in terms of inference and survival. If the frequency of visits is to be to what is known as ‘modularity’ for causal based on a system that is observed under increased, with !nancial implications, then DAGs. We then combine the above previous certain conditions to a system under di(erent we certainly want to know whether they are developments to give graphical rules for conditions, namely such that the intensity for causally a(ecting hospitalisation, and not just the identi!ability of the e(ect of such certain types of events are changed, which whether both are associated. interventions via IPW; these rules can be o(ers greater generality. Dynamic associations among di(erent regarded as characterising ‘unobserved con- types of events in continuous time can be founding’. Re-weighting then simply replaces References: represented by local independence graphs the observed intensity by the one given by Dawid and Didelez (). Identifying the consequences of as developed by Didelez ("##)). Intuitively the intervention of interest. For this to be dynamic treatment strategies: A decision theoretic overview, we say that a process is locally independent meaningful, causal validity and identi!ability Statistics Surveys, , -. of another one if its short-term prediction is are crucial assumptions. As an aside, we !nd Didelez (). Asymmetric Separation for Local Independence not improved by using the past of the other that it is helpful to also use causal reasoning Graphs, in: Proceedings of the nd Annual Conference on process, similar to Granger non-causality; when faced with censoring as the target Uncertainty in Artificial Intelligence, - the graphical representation uses nodes of inference can o0en be regarded as the Didelez (). Graphical models for marked point processes for processes or events and the absence of population in which censoring is prevented, based on local independence, JRSS(B), , -. a directed edge for local independence. i.e. its intensity is set to zero. We apply our Robins, Hernan, Brumback (). Marginal structural models Important independence properties can theoretical results to the example of cancer & causal inference in epidemiology. Epidemiology, :-. be read o( these—possibly cyclic—graphs screening in Norway. Røysland (). A martingale approach to continuous time using delta-separation (Didelez, "##%) which Our approach can be regarded as the marginal structural models. Bernoulli, , -. generalises d-separation from DAGs. In time-continuous version of Dawid & Røysland (). Counterfactual analyses with graphical models related work, Røysland ("#$$, "#$") showed Didelez ("#$#), who develop a decision based on local independence. Ann. Statist., , -. August . 2016 IMS Bulletin . 11

Medallion lecture summary: Peter Diggle

Peter Diggle gave this Medallion Lecture at the ENAR meeting in March. Peter began his academic career in  as Lecturer in Statistics at the University of Newcastle upon Tyne, UK. Between  and  he was Senior Research Scientist, then Principal Research Scientist, then Chief Research Scientist in the CSIRO Division of Mathematics and Statistics in Canberra. Since  he has been at Lancaster University, where his current position is Distinguished University Professor of Statistics in the Faculty of Health and Medicine. He also holds Adjunct positions at Johns Hopkins, Yale and Columbia Universities, and is president of the Royal Statistical Society (–). His research interests are in statistical methods for spatial and longitudinal data analysis and their applications in the biomedical and health sciences, with a particular focus on environmental and tropical disease epidemiology.

Model-Based Geostatistics for Prevalence Mapping in Low-Resource Settings In low-resource settings, prevalence mapping plays an important role predictive probability, at each location, that this threshold is crossed. in determining priority areas for large-scale prevention and treatment &e map e(ectively delineates areas that are “safe”, “unsafe” (predictive programmes. Because disease registries are lacking, prevalence probabilities close to zero or one, respectively) and intermediate areas mapping relies on !eld data collected from prevalence surveys of (in gray) where more information is needed. communities within the region of interest. Only a small fraction of Work is in progress on the following extension to the eyeworm at-risk communities can be included in these surveys, and mapping at problem. It is now known that those at risk of experiencing unsampled locations necessarily involves some form of interpolation severe reaction to Mectizan are people whose blood is heavily or smoothing of the data. &e precision of the interpolated maps can infected with Loa loa parasites, more than ?6,666 parasites per be improved by exploiting the availability of remotely sensed images ml of blood. Determining infection levels routinely in the !eld is that act as proxies for environmental risk factors. di'cult. However the distribution of individual infection levels, A standard geostatistical model for data of this kind is a general- Y, within a community is well described by a Weibull distribution, ized linear mixed model, P(Y>y)=Pexp{−(y/L)κ} : y ≥ 6, where κ≈6.; and (P, L) vary randomly

Yi ~ Bin{mi , P(xi)} between communities. Speci!cally, S2(x) = log{P/(2−P)} and S3(x) = log[P(xi)/{2−P(xi)} = z(xi) β + S(xi) + Ui , log L can be modelled as a bivariate Gaussian process and the correla- where Yi is the number of positives in a sample of mi individuals tion between the two exploited to enable prediction of P(Y>?6,666) at location xi, z(x) is a vector of spatially referenced explanatory in a newly sampled community for which only prevalence data are variables, S(x) is a spatially correlated Gaussian process and the Ui available. Two general conclusions from this work are that in low are uncorrelated Gaussian random variables. &e roles of S(x) and resource settings: geostatistical modelling of prevalence data can

Ui are to account for spatially structured and unstructured variation, deliver practical solutions to problems that would otherwise be intrac- respectively, that is not explained by z(x). table; and that predictive probability mapping is o0en a more useful &is model has been used in particular to assist the operation of inferential paradigm than either testing or estimation. pan-African control programmes for two vector-borne diseases, oncho- cerciasis (river blindness) and lymphatic 0lariasis (elephantiasis). &e References control strategy is based on prophylactic administration of a !laricide, Diggle, P.J., Thomson, M.C., Christensen, O.F., Rowlingson, B., Obsomer, V., Gardon, J., Wanji, S., Mectizan, to whole communities in a(ected areas. In this context, Takougang, I., Enyong, P., Kamgno, J., Remme, H., Boussinesq, M. and Molyneux, D.H. (). estimating prevalence at a particular location is less important than Spatial modelling and prediction of Loa loa risk: decision making under uncertainty. Annals of predicting whether prevalence exceeds a policy-relevant threshold. For Tropical Medicine and Parasitology, , –. example, the operation of the control programmes has been hampered Zoure, H.G.M., Noma, M., Tekle, A.H., Amazigo, U.V., Diggle, P.J., Giorgi, E. and Remme, J.H.F. by the recognition that people heavily infected with a third disease, (). The geographic distribution of onchocerciasis in the  participating countries of the Loa loa (eyeworm), are at risk of experiencing severe, occasionally African Programme for Onchocerciasis Control: . Pre-control endemicity levels and estimated fatal, adverse reactions to Mectizan. &is has resulted in a policy that number infected. Parasites and Vectors, ,  precautionary measures must be taken before Schlueter, D.K., Ndeo-Mbah, M.L., Takougang, I., Ukety, T., Wanji, S., Galvani, A.P. and Diggle, P.J. the drug is administered in a community (). Using community-level prevalence of Loa loa infection to predict the proportion of high- where prevalence of eyeworm is greater than ly-infected individuals: statistical modelling to support lymphatic filariasis elimination programs. 36%. Accordingly, the map (le0) shows the PLoS Neglected Tropical Diseases (submitted). 12 . IMS Bulletin Volume 45 . Issue 5

New IMS Monographs book by Bradley Efron and Trevor Hastie

We are pleased to introduce the latest in the popular IMS Monographs series, published in a cooperative arrangement with Book signing Cambridge University Press. Computer Age Statistical Inference: Algorithms, Evidence, and Data Science is written by Bradley Efron at JSM and Trevor Hastie, both from Stanford University. Published in the UK in July and the USA in September, you can get your copy (with your  IMS member’s discount) from www.cambridge.org/ims If you’re going to JSM you can pick up a copy there: Brad and Trevor will be doing a book signing at the Cambridge University Press stand in the Expo Hall on Tuesday, August  at pm.

&e "$st century has seen a breathtaking expansion of statistical methodology, both in scope and in/uence. “Big data,” “data science,” and “machine learning” have become familiar terms in the news, as statistical methods are brought to bear upon the enormous data sets of modern science and commerce. How did we get here? And where are we going? &is book takes us on an exhilarating journey through the revolution in data analysis following the introduction of electronic computation in the $+.#s. Beginning with classical inferential theories—Bayesian, frequentist, Fisherian—individual chapters take up a series of in/uential topics: survival analysis, logistic regression, empirical Bayes, the jackknife and bootstrap, random forests, neural networks, Markov Chain Monte Carlo, inference a0er model selec- tion, and dozens more. &e distinctly modern approach integrates methodology and algorithms with statistical inference. &e book ends with speculation on the future direction of statistics and data science. Hardback ISBN : US. IMS members .

“How and why is computational statistics “A masterful guide to how the inferential bases “E(on and Hastie… have managed brilliantly taking over the world? In this serious work of of classical statistics can provide a principled to weave the 0ber of "2' years of statistical synthesis that is also fun to read, E(on and disciplinary (ame for the data science of the inference into the more recent historical mech- Hastie, two pioneers in the integration of twenty-0rst century.” anization of computing. $is book provides parametric and nonparametric statistical ideas, — Stephen Stigler, University of Chicago the reader with a mid-level overview of the give their take on the unreasonable e1ectiveness last #'-some years by detailing the nuances of statistics and machine learning in the “$is is a guided tour of modern statistics that of a statistical community that, historically, context of a series of clear, historically informed emphasizes the conceptual and computational has been self-segregated into camps of Bayes, examples.” advances of the last century. Authored by (equentist and Fisher, yet in more recent years, — Andrew Gelman, Columbia University two masters of the 0eld, it o1ers just the right has been uni0ed by advances in computing.” mix of mathematical analysis and insightful — Rebecca Doerge, Carnegie Mellon commentary.” University — Hal Varian, Google August . 2016 IMS Bulletin . 13

Project Euclid News

Two Open Access Probability Journals Join Project Euclid We are excited to announce that both the Electronic Journal of Probability and Electronic Communications in Probability are now available in Project Euclid. Co-published by the Institute of Mathematical Statistics and the Bernoulli Society, these titles join other high-quality publications in probability and statistics available About Project Euclid at http://projecteuclid.org. Project Euclid is a collaborative partnership between Cornell &e Electronic Journal of Probability is an open access journal pub- University Library and Duke University Press, which seeks to lishing full-length research articles in probability theory. Electronic advance scholarly communication in theoretical and applied Communications in Probability is an open access journal publishing mathematics and statistics through partnerships with indepen- short research articles in probability theory. Both publish papers in all dent and society publishers. It was created to provide a platform for small publishers of scholarly journals to move from print to areas of probability, “only accepting papers of high quality in terms of electronic in a cost-effective way. both scienti!c contents and the presentation of the results.” For the past twenty years, the Electronic Journal of Probability Through a combination of support by subscribing libraries and and Electronic Communications in Probability have helped shape the participating publishers, Project Euclid has made  of its jour- !eld of probability theory as two of the longest-published open access nal articles available as open access. Project Euclid provides access journals in this area of mathematics. Both journals’ archives, as well as to over . million pages of open-access content. all future volumes, will be available at http://projecteuclid.org, joining In , Cornell University Library received a grant from the over $.* million pages of openly available scholarship in mathematics Andrew W. Mellon Foundation for the development of an online and statistics on the Project Euclid platform. publishing service designed to support the transition for small, Project Euclid’s ongoing partnership with the Bernoulli Society non-commercial mathematics journals from paper to digital dis- and the IMS is a ful!lment of Project Euclid’s mission to support tribution. Duke University Press, which had experience in putting sustainable independent and society publishing in mathematics and its own math journals online and a similar interest in assisting statistics. “&e long-standing partnership between the IMS, the non-commercial math journals, worked as Cornell’s partner in Bernoulli Society, and Project Euclid is evidence of the viability of developing the grant application and then in developing Project Euclid’s publishing model. sustainable, not-for-pro!t STEM publishing models,” says David Ruddy, Director of Information Technology for Project Euclid. Cornell launched Project Euclid in May  with nineteen jour- “Project Euclid is proud to serve as a new home for the nals. In July , Cornell University Library and Duke University Electronic Journal of Probability and Electronic Communications in Press established a joint venture and began co-managing Project Probability.” Euclid. Duke assumed responsibility for “marketing, financial, and order fulfillment workflows” while Cornell continued to provide Entire CBMS Monograph Series Available in Project Euclid and support Project Euclid’s IT infrastructure. Project Euclid is delighted to announce that the entire NSF–CBMS Currently, Project Euclid hosts both open access journals and Regional Conference Series in Probability and Statistics is now available monographs, as well as its Prime collection of peer-reviewed titles. at http://projecteuclid.org/cbms. At the time of writing, there are  journal titles, from  publish- &e lecture notes in this series cover a variety of essential topics ers from around the world. in mathematical statistics ranging from probability models to process In , Project Euclid received the  Division Award from the theory. Physics-Astronomy-Mathematics Division of the Special Libraries &is series is co-published by the Institute of Mathematical Association. Given annually, this award recognizes significant con- Statistics and the American Statistical Association. It is sponsored by tributions to the literature of physics, mathematics, or astronomy, the Conference Board of the Mathematical Sciences and supported by and honors work that demonstrably improves the exchange of the National Science Foundation. All volumes are openly accessible information within these three disciplines. The award also takes through Project Euclid, with print editions also available for purchase. into consideration projects that benefit libraries. Information from projecteuclid.org and Wikipedia and Information projecteuclid.org from 14 . IMS Bulletin Volume 45 . Issue 5

Recent papers: two IMS-supported journals Bayesian Analysis Volume 11, No 2, June 2016

Bayesian Analysis is an electronic journal of the International Society for Bayesian Analysis. It seeks to publish a wide range of articles that demonstrate or discuss Bayesian methods in some theoretical or applied context. The journal welcomes submissions involving pre- sentation of new computational and statistical methods; critical reviews and discussions of existing approaches; historical perspectives; description of important scientific or policy application areas; case studies; and methods for experimental design, data collection, data sharing, or data mining. Access papers at http://projecteuclid.org/euclid.ba

Exact and Approximate Bayesian Inference for Low Integer-Valued Time Series Models with Intractable Likelihoods ...... CHRISTOPHER C. DROVANDI, ANTHONY N. PETTITT, AND ROY A. MCCUTCHAN;    A Two-Component G-Prior for Variable Selection ...... HONGMEI ZHANG, XIANZHENG HUANG, JIANJUN GAN, WILFRIED KARMAUS, AND TARA SABOATTWOOD;    Flexible Bayesian Survival Modeling with Semiparametric Time-Dependent and Shape-Restricted Covariate Effects ...... THOMAS A. MURRAY, BRIAN P. HOBBS, DANIEL J. SARGENT, AND BRADLEY P. CARLIN;    Sums of Possibly Associated Bernoulli Variables: The Conway–Maxwell-Binomial Distribution...... JOSEPH B. KADANE;    Expert Information and Nonparametric Bayesian Inference of Rare Events ...... HWANSIK CHOI;    Bayesian Registration of Functions and Curves...... WEN CHENG, IAN L. DRYDEN, AND XIANZHENG HUANG;    Posterior Contraction Rates of the Phylogenetic Indian Buffet Processes...... MENGJIE CHEN, CHAO GAO, AND HONGYU ZHAO;    Incorporating Marginal Prior Information in Latent Class Models...... TRACY A. SCHIFELING AND JEROME P. REITER;    A Mixture Model for Rare and Clustered Populations Under Adaptive Cluster Sampling ...... KELLY C. M. GONÇALVES AND FERNANDO A. S. MOURA;    Posterior Propriety for Hierarchical Models with Log-Likelihoods That Have Norm Bounds ...... SARAH E. MICHALAK AND CARL N. MORRIS;    Importance Sampling Schemes for Evidence Approximation in Mixture Models...... JEONG EUN LEE AND CHRISTIAN P. ROBERT;    Pre-surgical fMRI Data Analysis Using a Spatially Adaptive Conditionally Autoregressive Model . ZHUQING LIU, VERONICA J. BERROCAL, ANDREAS J. BARTSCH, AND TIMOTHY D. JOHNSON; 599 - 625

Brazilian Journal of Probability and Statistics Volume 30, No 2, May 2016 The Brazilian Journal of Probability and Statistics is an official publication of the Brazilian Statistical Association and is supported by the IMS. It is published four times a year, in February, May, August, and December. The Journal publishes papers in applied probability, applied statistics, computational statistics, mathematical statistics, probability theory and stochastic processes. Access papers at http://projecteuclid.org/euclid.bjps

A semiparametric Bayesian model for multiple monotonically increasing count sequences...... VALERIA LEIVA-YAMAGUCHI AND FERNANDO A. QUINTANA; 155 - 170 Parametric Stein operators and variance bounds ...... CHRISTOPHE LEY AND YVIK SWAN; 171 - 195 Log-symmetric distributions: Statistical properties and parameter estimation...... LUIS HERNANDO VANEGAS AND GILBERTO A. PAULA; 196 - 220 Local limit theorems for shock models ...... EDWARD OMEY AND REIN VESILO; 221 - 247 A new skew logistic distribution: Properties and applications...... D. V. S. SASTRY AND DEEPESH BHATI; 248 - 271 Fractional absolute moments of heavy tailed distributions ...... MUNEYA MATSUI AND ZBYNĚK PAWLAS; 272 - 298 Limiting behavior of the Jeffreys power-expected-posterior Bayes factor in Gaussian linear models ...... D. FOUSKAKIS AND I. NTZOUFRAS; 299 - 320 On the stability theorem of Lp solutions for multidimensional BSDEs with uniform continuity generators in z...... JIAOJIAO MA, SHENGJUN FAN, AND RUI FANG; 321 - 344 August . 2016 IMS Bulletin . 15

Vlada’s Point: Peer Review II — An Idea

Contributing Editor Vlada Limic has been thinking about ways in board, and !nally which the peer review process could be improved. She writes: ask your questions A young postdoc enters a renowned senior professor’s o'ce one day. directly, without &e master is sighing over a printout of an article, and !nally says : “I having to disclose have read this paper in detail, I checked that the argument is entirely your identity (or that correct line for line. But if you’d ask me what it is, I would not be of the AE handling able to tell… and now I must write a report!” More sighs follow. &e the submission). apprentice is taken aback by this striking piece of news, and thinks, While the AE “Could it be that with all that knowledge, experience, wisdom, creativity will not be able to participate in the discussion (unless they open it and problem-solving capacity, one still faces these challenges?” themselves), they can con!rm (if needed) that your web booth invita- &is scene took place more than sixteen years ago, but it could tion issued to the authors is genuine. have equally been ," years ago, or yesterday. While most of us know Via your secure private channel, you can exchange text messages that this is by far not the saddest ending of a peer review round, it is (posts), pdf, jpeg (and perhaps other format) !les as attachments already signi!cant. with the author team, while staying anonymous. &e entire electronic Now imagine a slightly di(erent world in which, along with all discussion can be printed out at any time by either party. As usual, if a the manuscript-central capacities, a free “self-service web booth” tool login and/or a password is lost, they can be easily recovered. Each par- exists. Each anonymous peer-reviewer , novice and veteran alike, has ticipant is noti!ed each time a new post by the other(s) arrives. &e the option of using it whenever in need of contacting the author(s) discussion stays active as long as something was posted within the last about the article under review. &e attribute anonymous is crucial—if so-many-months (depending on storage space), otherwise (followed the peer review were done openly today, like it was done $## or -# by several email noti!cations to the participants) it gets erased. years ago, we would not need much more than the phone and the Now imagine, for a change, that you are an author of the above internet. Even hand-written “snail mail” would typically work quite article under review, and that you have just received a one or two well. page long report from the journal describing how your Lemma ", Multiple changes in overall scienti!c valorization seem to be then Lemma -, and !nally the main theorem are plainly wrong. And required before anonymity  could be li0ed here. indeed, one of the hypotheses was missing or mistyped, the inclusion So taking the anonymity for granted, I ask you again: imagine or correction of which makes Lemma " and everything else sound. having access today to the above web booth. It’s easy if you can! But hey, the machine already pushed the editors into a decision Consider your most recent peer review assignment. &e topic about your submission: it is requires major revision at best. A letter of is exciting, the paper solves an important/interesting problem, you response with detailed explanations is the least that is now expected know some but not all of the concepts or techniques used. &e task from you in order to put the review back in motion . looks like a great opportunity for learning. Would you not prefer to have been promptly and directly con- However, suppose that once you start reading the preprint in tacted by the reviewer? detail, you realize that either you are misunderstanding the logic, or My !nal point for today: the just described communication tool there is a non-trivial gap in the argument in Lemma ", which (as you already exists, and in more than one incarnation. None of them is alas keep reading) propagates to Lemma -, and !nally undermines the openly and freely available to general academic community. Still, there proof of the main result. is hope on the horizon… Are you missing something, or is it really a gap? How important is it? Can it be !xed?  The words are approximate, the meaning is exact. &inking about it harder, in brief but intensive intervals spaced  …regardless of the journal, or publisher (s)he has committed to over several weeks, and !nally !nding the time to write that one- to serving. two-page report to the editors explaining your concerns, is not the  …which protects the reviewer from potential open or disguised only option. retaliation, in case their report includes reservations or You can go to the web booth platform, open there a secured disapproval. (login/password protected) discussion dedicated to your assignment,  This “ping-pong” can go on for several rounds, the process could simultaneously invite the author(s) to it and notify the editorial well last for years. 16 . IMS Bulletin Volume 45 . Issue 5

Treasurer’s Report 2015

Introduction Membership Data &is report details membership and sub- Total individual paid membership in the Institute as of December ,$, "#$. decreased $.,C scription data for calendar year end "#$.. from December ,$, "#$*. Table $ presents the membership data back to "##). As can be seen, &e "#$. !scal year end audit report will the paid membership reached a high of ,$.% in "##) and has been decreasing since then. &is be published separately in the Fall of "#$% trend is similar to that of other professional societies. Nevertheless, this is clearly an area of a0er the auditors have completed the annual concern, and the IMS Executive Committee continues to look for ways to address this issue. process. TABLE 1: Membership, by Calendar Year In "#$., the total number of IMS paid     *     change members decreased, but the total number Regular , , , , , , , , -. of members increased. Subscriptions by Life/Retired Life         . institutions also decreased this past year. &e Reduced Country/Retired/         . !nancial status of the Institute continues IMS China New Graduate         -. to be stable, and actions have been taken Student , , , , , , , , . to ensure its long-term stability. Details of Total , , , , , , , , . the events of the past year, membership and Total excluding free members , , , , , , , , -. subscription data, sales data are given below. (students, and in – IMS China) * 2012 member figures contain some estimates Publications South/Central America Australia/New Zealand Africa &e following is a list of all current IMS Geographic Distribution of Members. Canada core, co-sponsored, supported and a'liated &e IMS membership is currently distributed journals: as follows (see pie chart, right): %"C United Asia IMS Core Print/Electronic Publications States; $)C Europe; $$C Asia; *C Canada; Annals of Probability; Annals of Applied ,C Australia and New Zealand; <"C South USA Probability; Annals of Statistics; Annals of America, Mexico and the Caribbean; <$C Europe Applied Statistics; Statistical Science; Current Africa. Index to Statistics; IMS Collections; IMS Monographs; IMS Textbooks; IMS Bulletin Subscription Data Co-Sponsored Print/Electronic Publications Selection of Journals by Members: Electronic Communications in Probability; Print subscriptions by members continued to decrease in "#$., as expected, because members Electronic Journal of Probability; Electronic are opting to reduce their use of print while enjoying free electronic access to all journals. Journal of Statistics; Journal of Computational Members are charged actual cost for print copies of journals, so there is no net loss or gain to and Graphical Statistics; NSF–CBMS Series the bottom line from changes in print subscriptions by members. Table " (below) shows the in Probability and Statistics; Probability current selection of print journals by members. Surveys; Statistics Surveys Supported Publications TABLE 2: Member** subscriptions, by calendar year ALEA: Latin American Journal of Probability PRINT (paid)          change and Mathematical Statistics; Annales de l’In- AAP         -. stitut Henri Poincaré (B); Bayesian Analysis; AOP         -. Bernoulli; Bernoulli News; Brazilian Journal AOAS , ,       -. of Probability and Statistics; Stochastic Systems AOS , ,       -. Affiliated Publications STS , , , ,     -. Probability and Mathematical Statistics Total , , , , , , , , -. ** Previously this information was reported as all members (including organizational), however data has been reformatted to show individual members only, to reflect the change in classification and to better view the current status of the data. August . 2016 IMS Bulletin . 17

&e IMS o(ers joint membership oppor- TABLE 3: Institutional paid subscriptions, by calendar year tunities with: the Bernoulli Society (BS); PRINT          change International Statistical Institute/Bernoulli AAP         -. Society (ISI/BS); International Society for AOP         -. Bayesian Analysis (ISBA); and Applied AOAS         . Probability Society/INFORMS (APS/ AOS , , , , , ,   -. INFORMS). STS         -. Bulletin         -. Institutional Subscription Data CIS         -. Table , (right) presents comparative AIHP s         . subscription data for institutions to each of Bernoulli s         . our scienti!c journals for "#$. and previous BJPS s n/a        . years. Almost all journals experienced Total , , , , , , , , -. subscription decreases in "#$.. Overall insti- Total IMS journals , , , , , , , , -. tutional subscriptions decreased by $.%C. &e s denotes IMS-supported journals. Numbers in [brackets] are prior to journal becoming IMS-supported. decrease to IMS journals, speci!cally, was "."C. We are continuing to see increases in our bundled o(erings which are discounted Sales of these volumes are going very well. Conclusion on the whole. Approximately %#C of the &e NSF–CBMS Regional Conference Series &e IMS Executive Committee has reviewed non-member subscribers to IMS journals published no new volumes in "#$.. &e IMS all data in this report. A long term !nancial are in USA and Canada, with the remaining Collections series and the IMS Lecture Notes– plan is already in place and the IMS contin- subscribers distributed throughout the world. Monograph Series have ceased publication. ues to be strong and stable !nancially. &e decreases in institutional subscriptions is Book Sales Data Financial and Audit Report being felt across the market and is not unex- Tables * and . (below) present sales data for &e !scal year ended December ,$, "#$.. pected. &e IMS leadership began planning all IMS book series. &e external audit of the IMS will have been for these decreases over $# years ago and has In "#$#, the IMS published its !rst completed in July "#$%. ensured that IMS resources are shored up to volumes in a cooperative arrangement with &e full audit report will appear in the protect the long-term stability and growth of Cambridge University Press to publish two IMS Bulletin in the Fall. the society. series, IMS Monographs and IMS Textbooks. Jean Opsomer, IMS Treasurer

TABLE 4: Total sales from the NSF-CBMS Regional Conference Series, the Lecture Notes–Monograph Series, and IMS Collections to         TOTAL Total NSF-CBMS sales ( volumes) ,        , Total LNMS sales ( volumes) ,        , Total IMS Collections sales ( volumes) n/a        

TABLE 5: Total sales of IMS Monographs and IMS Textbooks Book to         TOTAL IMS Monographs (four volumes published so far, in , ,  and ) n/a n/a       , IMS Textbooks (six volumes published to date, in , , , ,  and ) n/a n/a       , Total Monographs & Textbooks sales n/a n/a , , , , , , , 18 . IMS Bulletin Volume 45 . Issue 5

IMS meetings around the world At a glance: forthcoming Joint Statistical Meetings: 2016–2020 IMS Annual

IMS sponsored meeting Meeting and JSM 2016 JSM dates July 30–August 4, 2016 Chicago, IL w http://amstat.org/meetings/jsm/2016 &e "#$% Joint Statistical Meetings will be held July ,# to August * at 2016 McCormick Place, ",#$ South Lake Shore Drive, Chicago, IL %#%$%. IMS Annual Meeting/ &e theme of JSM "#$% is “&e Extraordinary Power of Statistics.” 9th World Congress: &e IMS program chair for invited sessions is Jan Hannig, Toronto, Canada, University of North Carolina e [email protected]. &e IMS July 11–15, 2016 contributed program chair is Alexander Aue, University of California, JSM: Chicago, IL, Davis e [email protected] July 30 – August 4 &e program is online: https://www.amstat.org/meetings/jsm/2016/onlineprogram/index.cfm &e early-bird registration deadline was June $. &e "#$. JSM housing reservations went very quickly, so if you are planning to attend, be sure to book your accommodation via the JSM website as soon as possible. 2017 IMS Annual Meeting Workshops for Junior Researchers: &e successful series of writing workshops for junior researchers continues this year, @ JSM: Baltimore, and is complemented for the !rst time by a new Reviewing & Editing workshop. See http://www.amstat.org/meetings/ MD, July 29 – wwjr/ August 3, 2017 Advice for first-timers at JSM: Check out the JSM website for some helpful tips if it is your !rst time at JSM. It’s a 2018 really big meeting, with thousands of people attending from all over the world, so it’s very common to feel a bit over- whelmed, and to have the sense that you’re missing out on sessions. See https://www.amstat.org/meetings/jsm/2016/ IMS Annual Meeting: firsttimeattendees.cfm Vilnius, Lithuania, July 2–6, 2018

ASA President’s Invited Speaker announced: &e ASA President’s Invited Speaker for JSM "#$% is Joe Palca, a science JSM: Vancouver, correspondent for National Public Radio (NPR). Since joining NPR in $++", Palca has covered a range of science Canada, July 28– topics, but is currently focused on the eponymous series, Joe’s Big Idea (http://www.npr.org/series/156490415/joes-big- August 2, 2018 idea), an NPR experiment exploring how ideas become innovations and inventions.

Professional Development at JSM: JSM "#$% will feature *% IMS sponsored meeting 2019 Professional Development o(erings: Continuing Education short 18th Meeting of New Researchers in IMS Annual Meeting courses, Computer Technology Workshops, and Personal Skills Statistics and Probability @ JSM: Denver, CO, Development Workshops. Search the online program for CE, University of Wisconsin–Madison July 27–August 1, CTW and PSD: see http://www.amstat.org/meetings/jsm/2016/ July 28–30, 2016 (immediately before JSM) 2019 professionaldevelopment.cfm w www.stat.wisc.edu/imsnrc18/about.html 2020 IMS sponsored meetings: JSM dates for – IMS Annual Meeting: IMS Annual Meeting JSM 2018 IMS Annual Meeting JSM 2020 IMS Annual Meeting TBD @ JSM 2017: July 28–August 2, @ JSM 2019 August 1–6, 2020 @ JSM 2021 July 29–August 3, 2018 July 27–August 1, Philadelphia, PA August 7–12, 2021, JSM: Philadelphia, 2017, Baltimore, MD Vancouver, Canada 2019, Denver, CO Seattle, WA August 1–6, 2020 August . 2016 IMS Bulletin . 19

More IMS meetings around the world

Peter Hall Memorial Conference UPDATED IMS co-sponsored meeting NEW September 30–October 1, 2016 Random processes and time series: theory Conference Center, University of California, Davis and applications (A conference in honor of w http://www.stat.ucdavis.edu/hallmemorialconference/ Murray Rosenblatt) Please join the UC Davis Department of Statistics for the Peter Hall Memorial Conference in October 21–23, 2016 honor of Distinguished Professor Peter Hall who sadly passed away in January "#$%. La Jolla, California, USA &e con!rmed speakers so far for the Peter Hall Memorial Conference are: w http://www.math.ucsd. t Jeannie Hall, Melbourne, Australia: Memorial Session edu/~rosenblattconf/ t Rudy Beran, UC Davis: “On double bootstrap asymptotics” &is conference will feature research on the t Peter Bickel, UC Berkeley: “$e bootstrap in some novel environments” topic of random processes and time series, t Tony Cai, University of Pennsylvania: “Adaptive estimation of a planar convex set” both theory and t Song Xi Chen, Iowa State University: TBC applications. &e t Ming-Yen Cheng, National Taiwan University: “A simple and adaptive two-sample test in conference celebrates high dimensions” the research of IMS t Aurore Delaigle, University of Melbourne: TBC Fellow Murray t Jianqing Fan, Princeton University: TBC Rosenblatt. Further t Frédéric Ferraty, University of Toulouse: “Variable selection in high-dimensional information about nonparametric regression setting “ Professor Murray t Jiashun Jin, Carnegie Mellon University: “Innovated higher criticism and statisticians’ Rosenblatt is at networks” http://math.ucsd.edu/~williams/mrosenb. t Iain Johnstone, Stanford University: TBC html t Runze Li, Pennsylvania State University: “Projection test for high-dimensional mean vectors &e conference will launch the Murray with optimal direction” and Adylin Rosenblatt Endowed Lecture t Steve Marron, University of North Carolina: “High dimension low sample size asymptotics” Series in Applied Mathematics. &e two inau- t Byeong Park, Seoul National University: “Smooth back0tting in errors-in-variables additive gural lecturers are Cathy Constable, Scripps models” Institution of Oceanography, UC San Diego, t Terry Speed, Melbourne and UC Berkeley: TBC and Robert Engle, New York University. t Matt Wand, University of Technology, Sydney: “Fast approximate inference for arbitrarily &e conference will also feature the large statistical models via message passing “ following plenary speakers: Richard Bradley, t Alan Welsh, Australian National University: “Early work on order statistics; research and Indiana University; David Brillinger, teaching” UC Berkeley; Richard Davis, Columbia t Fang Yao, University of Toronto: “Mixture models and densities for functional data” University; Larry Goldstein, USC; Keh- &e full schedule of events and speaker abstracts will be posted soon. Registration is open Shin Lii, UC Riverside; Magda Peligrad, now: http://www.stat.ucdavis.edu/hallmemorialconference/register.php University of Cincinnati; Dimitris Politis, If you would like to give a short con- UC San Diego; Philip Stark, UC Berkeley; Peter Hall was the IMS Wald lecturer in 2006 tributed talk please complete the form on Murad Taqqu, Boston University; and Wei the website. Space for contributed talks is Biao Wu, University of Chicago. limited and so you may be asked to provide Registration—free but required—is open a poster instead. Talks could be as short as to researchers who have a research interest in !ve minutes depending on the number of the topic of random processes and time series contributors. and are a'liated with Universities or indus- trial or government research institutions. &is includes current postdocs and PhD students. 20 . IMS Bulletin Volume 45 . Issue 5

More IMS meetings around the world

IMS co-sponsored meeting IMS co-sponsored meeting 9th World Congress on Probability and Statistics 39th Conference on Stochastic Processes and their Applications (SPA) July 11–15, 2016. Toronto, Canada July 24–28, 2017 w http://www.fields.utoronto.ca/programs/scientific/16-17/WC2016/ Moscow, Russia &is meeting is jointly sponsored by the Bernoulli Society and the w TBC IMS. &e Scienti!c Programme Chair is Alison Etheridge. &e Local Chair is Tom Salisbury. &e +th World Congress on Probability and IMS co-sponsored meeting Statistics will be hosted by the Fields Institute. Reproducibility of Research: Issues and Proposed Remedies March 8–10, 2017 IMS sponsored meeting Washington DC, USA Joint 2018 IMS Annual Meeting and 12th International Vilnius w http://www.nasonline.org/programs/sackler-colloquia/upcoming- Conference on Probability Theory & Mathematical Statistics colloquia/ July 2–6, 2018 &is meeting is one of the Arthur M. Sackler Colloquia, which Vilnius, Lithuania address scienti!c topics of broad and current interest that cut across w TBC the boundaries of traditional disciplines. We are please to announce that the "#$) IMS Annual Meeting will be held in beautiful Vilnius, the capital of Lithuania, in conjunction IMS co-sponsored meeting with the $"th Vilnius Conference on Probability &eory and 6th IMS-FIPS (Finance, Insurance, Probability & Statistics) Workshop Mathematical Statistics. &e Program Co-chairs are Peter Bühlmann July 7–9, 2016. Edmonton, Alberta, Canada (IMS) and Vygantas Paulauskas (Vilnius). &e Local Chair is w http://www.mathfinance2016.com Remigijus Leipus. Details to follow, but mark your calendars! &e primary purpose of the workshop is to bring together a global cast of leading academic experts, practitioners and junior researchers IMS co-sponsored meeting to share research that underscores the contributions of probability 2017 IMS-China International Conference on Statistics and statistics to the development of quantitative models, methods, and Probability techniques and technologies in the !elds of !nance and insurance. June 28–July 1, 2017 Nanning, Guangxi Province, China w TBC Local organizing committee chair: Zijia Peng, Guangxi University for Nationalities, China e [email protected]. Scienti!c program com- ENAR 2017–2019 dates mittee chair: Ming Yuan, University of Wisconsin–Madison, USA IMS sponsored meetings e [email protected]. &e website is under construction. March 12–15, 2017: in Washington DC March 25–28, 2018: in Atlanta, GA IMS co-sponsored meeting March 24–27, 2019: in Philadelphia, PA WNAR Annual Meeting in conjunction with the XXVIII International w http://www.enar.org/meetings/future.cfm Biometric Conference July 10–15, 2016, Victoria, BC, Canada IMS co-sponsored meeting w http://biometricconference.org/ The 10th ICSA International Conference &e next WNAR Annual Meeting, in conjunction with the XXVIII December 19–22, 2016 International Biometric Conference (IBC"#$%), will be held Shanghai Jiao Tong University, China July $#–$., "#$% at the Victoria Conference Centre in Victoria, w http://www.math.sjtu.edu.cn/conference/2016icsa/ British Columbia, Canada. A list of invited sessions is at http:// &e theme is Global Growth of Modern Statistics in the "%st Century. biometricconference.org/invited-sessions/. &ere will also be four full Plenary speakers are: Jim Berger, Tony Cai, Kai-Tai Fang, Zhiming day short courses. Ma, Marc A. Suchard, Lee-Jen Wei and C.F. Je( Wu. August . 2016 IMS Bulletin . 21

Other meetings and events around the world

Workshop on Fractals, Ergodic Theory NEW 7th CEQURA Conference on Advances in Financial NEW and Number Expansions and Insurance Risk Management August 30–September 2, 2016 September 26–27, 2016 Utrecht, The Netherlands Munich, Germany w http://www.math.leidenuniv.nl/~kallecccj/uu/workshopuu.html w http://www.cequra.uni-muenchen.de/conference2016 A workshop on “Fractals, Ergodic &eory and Number Expansions” Contact: Stefan Mittnik [email protected] that will take place at Utrecht University from August ,# to &e -th CEDURA Conference on Advances in Financial and September ", "#$%. Insurance Risk Management, organized by the Society for Financial Con!rmed invited speakers are Henk Bruin, Vilmos Komornik, and Insurance Econometrics and in collaboration with the Montreal Derong Kong, Wenxia Li, Joerg &uswaldner and Reem Yassawi. Institute of Structured Products and Derivatives and the Bayerisches For further information please visit the website. Finanz Zentrum, provides a platform for presenting and discussing current developments in research and industry, and fosters the exchange between academics and practitioners from the risk manage- ment community.

Women in Statistics and Data Science Conference 2016 NEW International Conference on NEW October 20–22, 2016 Computer Systems & Mathematical Sciences Charlotte, NC, USA November 18–19, 2016 w https://ww2.amstat.org/meetings/wsds/2016/index.cfm Ghaziabad (Delhi NCR), India Early registration is open until September %. w http://www.its.edu.in/iccsms-2016 Join us for the "#$% Women in Statistics and Data Science Conference—the only conference &e conference aims at creating a competent in the !eld tailored speci!cally for women—October "#–"" in Charlotte, NC. WSDS will forum for academicians, researchers and bring together academic, industry, and government professionals and students working in industry professionals from all across the statistics, biostatistics, and data science. world to bring together and share their Want more info? Listen to this half-hour podcast to get the inside scoop on how the experience, work, research outcomes from conference came to be, what to expect this year, and plans for the future: http://community. various parts of the world on a global forum. amstat.org/biop/podcast &is conference includes all major topics of Register online and stay tuned for the full program, coming soon. Mathematical Sciences, Statistical Sciences & Computer Sciences.

IEEE International Conference on Computing, Analytics and Security Trends NEW December 19–21, 2016 College of Engineering Pune, Maharashtra, India w http://cast2016.coep.org.in/ Potential authors are invited to submit manuscripts of original unpublished research and recent developments in the topics related to the CAST-"#$% conference. Accepted papers will be published in the conference proceedings, which will be submitted to IEEE Xplore Digital Library. Call for original contributions from wide area listed below but are not limited to: t
Data Analytics, Big Data and Bioinformatics t
Information Security and Networking t
Distributed,Parallel and Cloud Computing t
Natural Language Processing and Information Retrieval t
Signal Processing, Multimedia and Embedded Systems t
Green Computing and Sustainable Energy Systems For details, please visit the conference website above or email [email protected] for any further query. 22 . IMS Bulletin Volume 45 . Issue 5

Other meetings and events around the world

Latent Variables 2016 Conference NEW Two-day workshop on Post-Model Selection NEW October 12–14, 2016 August 22–23, 2016 Columbia, South Carolina, USA Leuven, Belgium w http://www.stat.sc.edu/latent-variables-2016 w http://feb.kuleuven.be/public/NDBAF45//postselection.htm &e Department of Statistics of the University of South Carolina, in &e main theme is how to deal with the e(ects of model selection partnership with NISS (National Institute of Statistical Sciences), is on further inference, this in various settings, using several estimation pleased to announce that it is organizing the Latent Variables "#$% and selection methods. Speakers include: F. Bachoc (Toulouse), A. Conference. &is conference will be held October $"–$*, "#$% at the Charkhi (KU Leuven), N.L. Hjort (Oslo), D. Kozbur (ETH Zurich), Alumni Center on the campus of the University of South Carolina, D. Rothenhäusler (ETH Zurich), U. Schneider (Vienna), J. Taylor Columbia SC. (Stanford), X. Tian (Stanford). &e conference will honor Prof. Jayaram Sethuraman’s contri- Registration is open until July ). bution to foundational issues in latent variables research. In addition to a plenary talk by Prof. Sethuraman, there will be plenary talks to XIV CLAPEM NEW be given by Prof. David Blei (Columbia University), Prof. Xuming December 5–9, 2016 He (University of Michigan) and Prof. Sonia Petrone (Universita’ San José, Costa Rica Boccone). &e conference will also feature six pairs of concurrent ses- w http://www.clapem.emate.ucr.ac.cr/ sions, each with three speakers, with each speaker giving ".-,# minute Contact: Luis Barboza Chinchilla [email protected] talks. &e Department of Statistics has research interests in several &e Latin American Congress of Probability and Mathematical areas related to latent variables theory and analysis. &e concurrent Statistics (CLAPEM) is the main event in probability and statistics sessions will feature research in these areas, including multiple testing, in the region, having been held roughly every two or three years for frailty models, objective Bayes analysis, factor models, diagnostic almost ,# years. It is organized under the auspices of the Bernoulli screening, model misspeci!cation, group testing and biomarker Society for Mathematical Statistics and Probability and the SLAPEM pooling, item response theory, model-based clustering, generalized (Latin-American Society on Probability and Mathematical Statistics). linear latent variable models, and measurement error models. &ere On this occasion the event is organized by the Universidad de Costa will also be poster sessions for graduate students and junior investiga- Rica, the Universidad Nacional and the Instituto Tecnológico de tors, with some travel support likely for graduate students and junior Costa Rica. &ere will be !ve short courses: Stochastic analysis by investigators. David Nualart, University of Kansas, USA; Large Deviations and For information, registration and updates about the conference, Sum Rules for Orthogonal Polynomials by Barry Simon, Caltech, USA; go to http://www.stat.sc.edu/latent-variables-2016 Bayesian hierarchical models by Gavin Shaddick, University of Bath, UK; Big data by Jean-Michel Loubes, Université Toulouse III – Paul Sabatier, France; Causal Inference by Michael Hudgens, University of North Carolina, USA. Invited speakers: Graciela Boente, Universidad de Buenos Are you organizing a meeting? It’s Aires, Argentina; Alexei Borodin, MIT, USA; Pietro Caputo, free, and easy, to get it listed here, Università Roma Tre, Italy; Rick Durrett, Duke University, USA; Onésimo Hernández, CINVESTAV, México; Jean-Michel Loubes, and also at the online calendar, Univ. Toulouse III – Paul Sabatier, France; Eric Moulines, Ecole www.imstat.org/meetings. Polytechnique, France; Susan Murphy, University of Michigan, USA; David Nualart, University of Kansas, USA; Gavin Shaddick, Submit the details at University of Bath, UK; Barry Simon, Caltech, USA; Michael www.imstat.org/submit-meeting Hudgens, University of North Carolina; &e deadline for contributed talks is September ., "#$%. &e deadline for poster contributions is October ,, "#$%. Registration is open. August . 2016 IMS Bulletin . 23

10th Conference on Extreme Value Analysis: EVA 2017 NEW CEN-ISBS Vienna 2017 Joint Conference on Biometrics & NEW June 26–30, 2017 Biopharmaceutical Statistics Delft, The Netherlands August 28–September 1, 2017 w www.eva2017.nl Vienna, Austria &e $#th International Conference on Extreme Value Analysis will w www.cenisbs2017.org take place at Del0 University of Technology in &e Netherlands. Contact: Alexandra Seppi e [email protected] It will schedule presentations on all Probabilistic and Statistical Joint conference of the International Biometric Society- aspects of Extreme Value Analysis and applications in Climate and Central European Network and the International Society for Atmospheric Science, Industrial Risks, Geosciences, Hydrology, Biopharmaceutical Statistics. CEN-ISBS Vienna "#$- is a unique Finance, Economics and Insurance, Biosciences, Physics, and joint conference integrating multiple perspectives on current Telecommunications and Stochastic Networks. Topics of interest biostatistical research. Get in touch with your colleagues from include: academia, industry and regulators, and join us in Vienna! Call for t Univariate, multivariate, in!nite dimensional extreme value topic-contributed sessions opens in October "#$%, abstracts can be theory submitted from January to March "#$-. t Heavy tails/light tails t Order statistics and records t Rare events and risk analysis, and t Spatial/spatio-temporal extremes More details of the conference can be found at the conference website.

Employment Opportunities around the world

New Zealand: Wellington United States: Orlando, FL Victoria University of Wellington University of Central Florida College of Medicine Professor in Actuarial Science Assistant, Associate or Full Professor of Medicine (Biostatistician) http://jobs.imstat.org/c/job.cfm?site_id=1847&jb=28102914 http://jobs.imstat.org/c/job.cfm?site_id=1847&jb=27071868

Switzerland: Lausanne United States: Princeton, NJ Ecole Polytechnique Federale de Lausanne Princeton University Instructorships in Mathematics Lecturer Position in Operations Research and Financial Engineering http://jobs.imstat.org/c/job.cfm?site_id=1847&jb=27571861 http://jobs.imstat.org/c/job.cfm?site_id=1847&jb=28118789

United States: Riverside, CA University of California, Riverside Multiple Ladder-Rank Faculty Positions in Business Analytics including Endowed Chairs (open rank) http://jobs.imstat.org/c/job.cfm?site_id=1847&jb=27285841

::: Search our online database of the latest jobs around the world for free at http://jobs.imstat.org ::: 24 . IMS Bulletin Volume 45 . Issue 5 International Calendar of Statistical Events IMS meetings are highlighted in maroon with the logo, and new or updated entries have the NEW or UPDATED symbol. Please submit your meeting details and any corrections to Elyse Gustafson: [email protected]

August 2016 August –September : London, UK. Population-based Time-to- event Analyses w http://csg.lshtm.ac.uk/pta2016/ August –: Ilulissat, Greenland. Applied Probability Symposium w http://thiele.au.dk/events/conferences/2016/ilulissat/ September 2016 August –: Hamilton, ON, Canada. Ordered Data and their Applications in Reliability and Survival Analysis: An September –: Manchester, UK. RSS #"$! International International Conference in Honour of N. Balakrishnan for his Conference w www.rss.org.uk/conference2016 !"th Birthday (ODRS #"$!) w http://mathandstats.mcmaster.ca/ odreliabilityandsurvival/ September –: Minsk, Republic of Belarus. Computer Data Analysis and Modeling: (eoretical and Applied Stochastics August –: University of Tennessee, Knoxville, USA. NIMBioS (CDAM #"$!) w http://www.cdam.bsu.by Tutorial: Evolutionary %uantitative Genetics #"$! w http://www. nimbios.org/tutorials/TT_eqg2016 September –: Almaty, Kazakhstan. (ird International Conference on Analysis and Applied Mathematics w http://www. August –: Stanford, CA, USA. MCQMC #"$!: $#th icaam-online.org International conference on Monte Carlo and quasi-Monte Carlo methods in scienti&c computing w http://mcqmc2016.stanford.edu September –: Ghent, Belgium. Flexible Statistical Modeling past, present and future w http://www.fsm16.ugent.be/ August –: Rønne, Bornholm, Denmark. Workshop on Geometry and Stochastics of Nonlinear, Functional and Graph September –: NCAR, Boulder, CO, USA. !th International Data w http://csgb.dk/activities/2016/geometry/ Workshop on Climate Informatics (CI #"$!) w https://www2.cisl. ucar.edu/events/workshops/climate-informatics/2016/home August –: Maastricht, &e Netherlands. Small Area Estimation Conference #"$! w http://www.sae2016.nl NEW September –: Munich, Germany. )th CE%URA Conference on Advances in Financial and Insurance Risk August –: Corvallis, Oregon, USA. #"$! IISA Conference on Management w http://www.cequra.uni-muenchen.de/ Statistics w http://iisaconference.org/ conference2016

August –: Birmingham, UK. International Society for Clinical September –: Washington DC. #"$! ASA Biopharmaceutical Biostatistics #"$! Conference w http://www.iscb2016.info/ Section Regulatory-Industry Statistics Workshop w http://www. amstat.org/meetings/biopharmworkshop/2016/ NEW August –: Leuven, Belgium. Two-day workshop on Post-Model Selection w http://feb.kuleuven.be/public/NDBAF45// September –October : University of California, Davis, postselection.htm USA. Peter Hall Memorial Conference w http://www.stat.ucdavis. edu/hallmemorialconference/ August –: Kerman, Iran. $'th Iranian Statistical Conference w http://isc13.uk.ac.ir/index.php?slc_lang=en&sid=1 September –October : St Louis, MO, USA. Workshop on Higher-Order Asymptotics and Post-Selection Inference (WHOA- NEW August –September : Utrecht, &e Netherlands. PSI) w http://www.math.wustl.edu/~kuffner/WHOA-PSI.html Workshop on Fractals, Ergodic (eory and Number Expansions w http://www.math.leidenuniv.nl/~kallecccj/uu/workshopuu.html August . 2016 IMS Bulletin . 25

October 2016 NEW December –: San José, Costa Rica. XIV CLAPEM w http://www.clapem.emate.ucr.ac.cr/ NEW October –: Columbia, South Carolina, USA. Latent Variables #"$! Conference w http://www.stat.sc.edu/latent- December –: Canberra, Australia. Australian Statistical variables-2016 Conference, $*th Australasian Data Mining Conference, +th Conference on Teaching Statistics w www.asc2016.com.au October –: Poznan, Poland. eRum #"$! (European R users meeting) w http://erum.ue.poznan.pl/ December –: Taipei, Taiwan. Conference on Experimental Designs and Analysis (CEDA) #"$! w http://www3.stat.sinica.edu. October –: Niagara Falls, Canada. International Conference tw/ceda2016/ on Statistical Distributions and Applications (ICOSDA #"$!) w http://people.cst.cmich.edu/lee1c/icosda2016/ NEW December –: College of Engineering Pune, Maharashtra, India. IEEE International Conference on Computing, Analytics October –: Charlotte, NC, USA. #"$! Women in Statistics and Security Trends w http://cast2016.coep.org.in/ and Data Science Conference w http://ww2.amstat.org/meetings/ wsds/2016/index.cfm December –: Chennai, India. Statistical Methods in Finance #"$! w http://www.cmi.ac.in/~sourish/StatFin2016/ NEW October –: La Jolla, California, USA. Random pro- cesses and time series: theory and applications (A conference in honor December –: Shanghai, China. $"th ICSA International of Murray Rosenblatt) w http://www.math.ucsd.edu/~rosenblattconf/ Conference w http://www.math.sjtu.edu.cn/conference/2016icsa/

December –: Kolkata, India. Platinum Jubilee International November 2016 Conference on Applications of Statistics w http://stat.caluniv.in/ platinum/ November –: University of Tennessee, Knoxville, USA. NIMBioS Workshop: Next Generation Genetic Monitoring w http://www. nimbios.org/workshops/WS_nextgen March 2017

November –: Miami, FL. International Conference on March –: Washington DC, USA. Reproducibility of %uestionnaire Design, Development, Evaluation, and Testing w Research: Issues and Proposed Remedies w http://www.nasonline. http://www.amstat.org/meetings/qdet2/index.cfm org/programs/sackler-colloquia/upcoming-colloquia/

NEW November –: Ghaziabad (Delhi NCR), India. International Conference on Computer Systems & Mathematical June 2017 Sciences w http://www.its.edu.in/iccsms-2016 June –: Riverside, CA, USA. $"th International Conference on Multiple Comparison Procedures w http://www.mcp-conference. December 2016 org/hp/2017

December –: Atlantic City, NJ, USA. )#nd Annual Deming NEW June –: Del0, &e Netherlands. $"th Conference on Conference on Applied Statistics w www.demingconference.com Extreme Value Analysis: EVA #"$) w www.eva2017.nl

Continues on page 26 26 . IMS Bulletin Volume 45 . Issue 5

International Calendar continued

June 2017 continued July –: Edinburgh, UK. ISBA #"$, World Meeting w TBC

June –July : Nanning, Guangxi Province, China. #"$) IMS- July  – August : Vancouver, Canada. JSM #"$, China International Conference on Statistics and Probability w TBC w http://amstat.org/meetings/jsm/

July 2017 July 2019

July –: Groningen, &e Netherlands. IWSM #"$) w http:// July –August : Denver, CO, USA. IMS Annual Meeting at iwsm2017.webhosting.rug.nl/ JSM #"$+ w http://amstat.org/meetings/jsm/

July –: Vigo, Spain. ',th Annual Conference of the International Society for Clinical Biostatistics w TBC August 2020

July –: Marrakech, Morocco. !$st ISI World Statistics Congress August –: Philadelphia, PA, USA. JSM #"#" #"$) w http://www.isi2017.org/ w http://amstat.org/meetings/jsm/

July –: Moscow, Russia. '+th Conference on Stochastic Processes and their Applications (SPA) w TBC August 2021

July  – August : Baltimore, USA. IMS Annual Meeting at August –: Seattle, WA, USA. IMS Annual Meeting at JSM JSM #"$) w http://amstat.org/meetings/jsm/ #"#$ w http://amstat.org/meetings/jsm/ Come to JSM 2017: this is Baltimore Inner Harbor at night (photo by Mitch Lebovic)

August 2022

August –: Washington DC, USA. JSM #"## w http://amstat.org/meetings/jsm/

Are we missing something? If you know of August 2017 any statistics or probability meetings which aren’t listed here, please let us know. NEW August –September : Vienna, Austria. CEN-ISBS Vienna #"$) Joint Conference on Biometrics & Biopharmaceutical You can email the details to Elyse Gustafson at Statistics w www.cenisbs2017.org [email protected], or you can submit the details yourself at http://www.imstat.org/ July 2018 submit-meeting.html We’ll list them here in the Bulletin, July –: Vilnius, Lithuania. Joint #"$, IMS Annual Meeting and on the IMS website too, at and $#th International Vilnius Conference on Probability (eory www.imstat.org/meetings & Mathematical Statistics w TBC August . 2016 IMS Bulletin . 27

Membership and Subscription Information Information for Advertisers Journals General information: The IMS Bulletin and webpages are the official The scienti!c journals of the Institute of Mathematical Statistics news organs of the Institute of Mathematical Statistics. The IMS are The Annals of Statistics, The Annals of Probability, The Bulletin, established in 1972, is published 8 times per year. Print Annals of Applied Statistics, The Annals of Applied Probability, and Statistical Science. The IMS Bulletin is the news organ of the circulation is around 2,000 paper copies, and it is also free online in Institute. PDF format at http://bulletin.imstat.org, posted online about two weeks before mailout (average downloads over 8,000). Subscription Individual Memberships Each individual member receives the IMS Bulletin (print and/or to the IMS Bulletin costs $125. To subscribe, call 877-557-4674 (US electronic) and may elect to receive one or more of the !ve scien- toll-free) or +1 216 295 2340 (international), or email staff@imstat. ti!c journals. Members pay annual dues of $115. An additional $74 org. The IMS website, http://imstat.org, established in 1996, receives is added to the dues of members for each scientific journal selected over 30,000 visits per month. Public access is free. ($49 for Stat Sci). Reduced membership dues are available to full-time students, new graduates, permanent residents of countries designated by the IMS Council, and retired members. Advertising job vacancies A single 60-day online job posting costs just $285.00. We will also Individual and General Subscriptions Subscriptions are available on a calendar-year basis. Individual include the basic information about your job ad (position title, subscriptions are for the personal use of the subscriber and must location, company name, job function and a link to the full ad) in the be in the name of, paid directly by, and mailed to an individual. IMS Bulletin at no extra charge. See http://jobs.imstat.org Individual subscriptions for 2016 are available to The Annals of Applied Probability ($199), The Annals of Applied Statistics ($199), Advertising meetings, workshops and conferences The Annals of Probability ($199), The Annals of Statistics ($199), Statistical Science ($174), and IMS Bulletin ($125). General Meeting announcements in the Bulletin and on the IMS website at subscriptions are for libraries, institutions, and any multiple- http://imstat.org/meetings are free. Send them to Elyse Gustafson; see readership use. Institutional subscriptions for 2016 are available to http://www.imstat.org/program/prog_announce.htm The Annals of Applied Probability ($475), The Annals of Applied Statistics ($475), The Annals of Probability ($475), The Annals Rates and requirements for display advertising of Statistics ($475), Statistical Science ($270), and IMS Bulletin ($118). Airmail rates for delivery outside North America are $135 Display advertising allows for placement of camera-ready ads for per title. journals, books, software, etc. A camera-ready ad should be sent as a IMS Bulletin grayscale PDF/EPS with all fonts embedded. Email your advert to The IMS Bulletin publishes articles and news of interest to IMS Audrey Weiss, IMS Advertising Coordinator [email protected] or members and to statisticians and probabilists in general, as well as see http://bulletin.imstat.org/advertise details of IMS meetings and an international calendar of statistical Dimensions: width x height Rate events. Views and opinions in editorials and articles are not to be / page ." wide x " high ( x  mm)  understood as official expressions of the Institute’s policy unless so stated; publication does not necessarily imply endorsement in any / page ." wide x " high ( x  mm)  way of the opinions expressed therein, and the IMS Bulletin and / page ." wide x " high ( x  mm)  its publisher do not accept any responsibility for them. The IMS Full page (to edge, ." wide x ." high  Bulletin is copyrighted and authors of individual articles may be including /" bleed) ( mm x  mm) asked to sign a copyright transfer to the IMS before publication. Full page (within usual ." wide x ." high  Bulletin margins) ( mm x  mm) The IMS Bulletin (ISSN 1544-1881) is published eight times per year in January/February, March, April/May, June/July, August, Deadlines and Mail Dates for IMS Bulletin September, October/November and December, by the Institute of Mathematical Statistics, 3163 Somerset Dr, Cleveland, Ohio Issue Deadline Online by Mailed 44122, USA. Periodicals postage paid at Cleveland, Ohio, and at : January/February December  December  January  additional mailing offices. Postmaster: Send address changes to : March February  February  March  Institute of Mathematical Statistics, 9650 Rockville Pike, Suite : April/May March  April  April  L3503A, Bethesda, MD 20814-3998. : June/July May  May  June  : August July  July  August  Copyright © 2016 by the Institute of Mathematical Statistics. : September August  September  September  Printed by The Sheridan Press, 450 Fame Avenue, Hanover, PA : Oct/Nov September  October  October  17331, USA. : December November  November  December  Volume 45 . Issue 5 the The purpose of the Institute is to foster the nextissue is development and dissemination September of the theory and applications of 2016 statistics and probability IMS: Organized September 12, 1935

ISSN 1050-5164 (print) ISSNhttp://projecteuclid.org/aoap 2168-8737 (online) Read IMS Bulletin THE ANNALS AAP June 2016 articles online at of http://bulletin.imstat.org APPLIED PROBABILITY

AN OFFICIAL JOURNAL OF THE DEADLINES INSTITUTE OF MATHEMATICAL STATISTICS submissionsfor Articles August 15, then Social contact processes and the partner model ERIC FOXALL,RODERICK EDWARDS AND P. VA N D E N DRIESSCHE 1297 Diverse market models of competing Brownian particles with splits September 15 and mergers ...... IOANNIS KARATZAS AND ANDREY SARANTSEV 1329 A positive temperature phase transition in random hypergraph 2-coloring VICTOR BAPST,AMIN COJA-OGHLAN AND FELICIA RASSMANN 1362 Please see inside Propagation of chaos for interacting particles subject to environmental noise MICHELE COGHI AND FRANCO FLANDOLI 1407 Approximations of stochastic partial differential equations the back cover for GIULIA DI NUNNO AND TUSHENG ZHANG 1443 Local asymptotics for controlled martingales subscription details SCOTT N. ARMSTRONG AND OFER ZEITOUNI 1467 Stein estimation of the intensity of a spatial homogeneous Poisson point process MARIANNE CLAUSEL,JEAN-FRANÇOIS COEURJOLLY AND JÉRÔME LELONG 1495 and information for A probabilistic approach to mean field games with major and minor players RENÉ CARMONA AND XIUNENG ZHU 1535 advertisers, including Estimation for stochastic damping Hamiltonian systems under partial observation. III. Diffusion term . . . . PATRICK CATTIAUX,JOSÉ R. LEÓN AND CLÉMENTINE PRIEUR 1581 From transience to recurrence with Poisson tree frogs all our deadlines and CHRISTOPHER HOFFMAN,TOBIAS JOHNSON AND MATTHEW JUNGE 1620 Bernoulli and tail-dependence compatibility requirements PAUL EMBRECHTS,MARIUS HOFERT AND RUODU WANG 1636 Beyond universality in random matrix theory ALAN EDELMAN,A.GUIONNET AND S. PÉCHÉ 1659 Super-replication with nonlinear transaction costs and volatility uncertainty PETER BANK,YAN DOLINSKY AND SELIM GÖKAY 1698 The snapping out Brownian motion ...... ANTOINE LEJAY 1727 Backward stochastic differential equation driven by a marked point process: An elementary approach with an application to optimal control Journal FULVIA CONFORTOLA,MARCO FUHRMAN AND JEAN JACOD 1743 Entropic Ricci curvature bounds for discrete interacting systems alerts MAX FATHI AND JAN MAAS 1774 Hack’s law in a drainage network model: A Brownian web approach For alerts and special RAHUL ROY,KUMARJIT SAHA AND ANISH SARKAR 1807 Gaussian fluctuations for linear spectral statistics of large random covariance matrices ...... JAMAL NAJIM AND JIANFENG YAO 1837 information on all the Duality theory for portfolio optimisation under transaction costs CHRISTOPH CZICHOWSKY AND WALTER SCHACHERMAYER 1888 IMS journals, sign up A note on the expansion of the smallest eigenvalue distribution of the LUE at the hard edge...... FOLKMAR BORNEMANN 1942 at the IMS Groups site http://lists.imstat.org Vol. 26, No. 3—June 2016