ims bulletin

8/10/2019 IMS Bulletin

1/36

September 2013

CONTENTS

1 COPSS Awards

2 Members News: Sourav

Chatterjee; Mike West; Sastry

Pantula;IMS Lecturers ;

Shahjahan Khan; Eswar

Phadia

4 X-L Files: Rejection Pursuit

6 Meeting reports: SPA;NRC

8 Obituaries: MauricePriestley;Wenbo Li

10 IMS Presidential Address

13 Travel grants

15 Statistics2013 workshop;NSF update

16 JSM 2013 in pictures

19 Recent papers:Annals ofProbability;Annals of Applied

Probability; IMS Textbooks

21 Terences Stuff:Statisticsand The War

22 IMS meetings

26 Other meetings

28 EmploymentOpportunities

31 International Calendar of

Statistical Events

35 Information for Advertisers

Volume 42 Issue 6

IMS BulletinCOPSS Awards at JSM

COPSS, the Committee of Presidents of Statistical Societies (IMS, ASA, SSC and

ENAR/WNAR) presented its awards to four IMS members, three of whom are

Fellows, at the Joint Statistical Meetings in Montreal on August .

Te biannual George W. Snedecor Award honors an individual who was instrumental

in the development of statistical theory in biometry. Te Snedecor Award was

presented to John David (Jack) Kalbfleisch, University of Michigan, For oundational

contribution to the field o biometry, especially or innovative analysis methods or ailure time

data, event history analysis, mixture models and likelihood theory. For influential collabora-

tive research, especially in the area o solid organ transplantation. For exceptional mentoring

o junior researchers, exemplary senior leadership o statistical groups, and steadast service to

the proession.

Te F. N. David Award recognizes a female statistician who exemplifies the contribu-

tions of Florence Nightingale David, an accomplished statistician in combinatorial prob-

ability theory. Te F. N. David Award was presented to Lynne Billard, University

of Georgia, For her world-leading research in sequential analysis, stochastic processes,

epidemiology and symbolic data analysis; or her extensive collaborations, especially in poultry

science and computer science; or exceptional contributions to leadership in the proession; and

or outstanding contributions as a role model, ostering opportunities especially or academic

women as leaders, researchers, administrators, and educators.Te COPSS Presidents Award goes annually to a young member of the statistical

community in recognition of outstanding contributions to the profession of statistics.

Tis years recipient is Marc A. Suchard, University of California, Los Angeles, For

extraordinarily wide-ranging and insightul contributions to the statistical sciences in the

areas o Bayesian inerence, computation and stochastic processes; or innovative statistical

modeling and novel computational techniques or ormerly intractable problems in molecular

epidemiology, evolutionary medicine, phylogenetics, phylogeography, computational biology

and emerging massive public health data sets; or a panoply o flexible, well-regarded and

widely used statistical sofware tools or the applied sciences; or extensive and creative statisti-

cal applications across an impressive range o human inquiry; and or service to the proession

as a prolific and dedicated mentor o the next generation o statisticians.Finally, the recipient of the R .A. Fisher Award & Lectureship, established in

to honor the contributions of Sir Ronald Aylmer Fisher and the work of a present-day

statistician, was Peter J. Bickel, University of California, Berkeley, For groundbreaking

contributions to semiparametric and nonparametric methods, adaptive estimation, and robust

statistics; or applying in-depth and intricate theoretical analysis to realistic problems in the

biological sciences; or penetrating and insightul analysis o scientific methodology which has

yielded a lasting impact on our understanding o both theory and methods; and or excep-

tional training and mentoring o students, leadership o proessional societies, and leadership

o his academic department.

Read it online at

http://bulletin.imstat.org
http://bulletin.imstat.org/http://bulletin.imstat.org/http://bulletin.imstat.org/


2/36

Executive CommitteePresident: Bin Yu

[email protected]

President-Elect: Erwin [email protected]

Past President: Hans R. [email protected]

Treasurer: Jean [email protected]

Program Secretary:Judith [email protected]

Executive Secretary:Aurore [email protected]

IMS Bulletin Editor: Dimitris Politis

Assistant Editor: Tati Howell

Contributing Editors:Peter Bickel, Anirban DasGupta,Nicole Lazar, Xiao-Li Meng, Terry Speed

Contact the IMS Bulletin by email:e [email protected] http://bulletin.imstat.org

https://www.facebook.com/IMSTATI

Contact the IMS regarding your dues,

membership, subscriptions, orders orchange o address:

IMS Dues and Subscriptions Office9650 Rockville Pike, Suite L3503A

Bethesda, MD 20814-3998

USA

t 877-557-4674 [toll-free in USA]t +1 216 295 5661[international]f +1 301 634 7099

e [email protected]

Contact the IMS regarding any othermatter, including advertising, copyright

permission, oprint orders, copyrighttranser, societal matters, meetings, ellowsnominations and content o publications:

Executive Director, Elyse GustafsonIMS Business OfficePO Box 22718, Beachwood

OH 44122, USA

t 877-557-4674 [toll-free in USA]t +1 216 295 5661[international]f +1 216 295 5661

e [email protected]

2.IMSBulletin Volume 42 .Issue 6

IMS Members News

Contact information

IMSBulletinVolume 42 Issue 6September 2013

ISSN 1544-1881

Sourav Chatterjee

Sourav Chatterjee awarded 2013 Love Prize

Te Line and Michel Love International Prize in Probabilityhas been awarded to Sourav Chatterjeeo the Courant Institute. Te

prize, which carries a monetary award o ,, will be presented ata ceremony in Berkeley to be held in Fall .

Sourav Chatterjee received his PhD in , advised by PersiDiaconis at Stanord University. His work has extraordinary breadth.On one side he has brought new ideas to bear upon classical topicsan extension o Lindebergs proo o the central limit theorem to aninvariance principle or arbitrary smooth unctions o weakly dependent random variables,a simpler proo o the amous KM theorem on strong approximation o a random walkby Brownian motion, and a new version o Steins method, reducing a large class o normalapproximation problems to variance bounding exercises. In other work, he has taken upalagrands Challenge to Mathematicians (to give rigorous analysis o spin glass models romstatistical physics) by providing analyses o random overlap structures and showing that theSherrington-Kirkpatrick model is chaotic under small perturbations o the couplings at anytemperature in the absence o an external field. Other topics to which he has made substantialcontributions include large deviations or random graphs and random matrices, first-passage

percolation, and probabilistic methods or discrete nonlinear Schrdinger equations.Among his other honors, Sourav received the IMS weedie New Researcher Award in

, and was a Medallion Lecturer in .Te prize commemorates Michel Love, Proessor at the University o Caliornia,

Berkeley, rom until his untimely death in . Te Prize was established by his widow,

Line, shortly beore her death in . Awarded every two years, it is intended to recognizeoutstanding contributions by researchers in probability who are under years old.

Sastry Pantula Dean

Oregon State Univ

Sastry Pantulahasmoved to OregonState University asDean o the Collegeo Science. Sastryserved on the aculty

o North CarolinaState Universitysince , and isa ormer directoro the NationalScience FoundationsDivision oMathematicalSciences.

Mitchell Prize awarded to Mike West and co-authors

Te Mitchell Prize is awarded in recognition o an outstanding paperthat describes how a Bayesian analysis has solved an important applied

problem. Te Prize is jointly sponsored by the Section on BayesianStatistical Science (SBSS) o the ASA, the International Societyor Bayesian Analysis (ISBA), and the Mitchell Prize FoundersCommittee.

Mike West, the Arts & Sciences Proessor o Statistics & Decision

Sciences at Duke University, received the award at JSM, with hisco-author Ioanna Manolopoulou (University College London), onbehal o their co-authors Melanie Matheu, Mike Cahalan and omKepler. Te award was or their studies o spatio-dynamic modellingin systems biology/immunology in their JASApaper entitledBayesian Spatio-Dynamic Modelling in Cell Motility Studies:Learning Nonlinear axic Fields Guiding Immune Response.

Tis award winning paper was cited, or its sophisticateddevelopment o model-based inerence methods in the expandingfield o immune-cell dynamics.
mailto:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]://bulletin.imstat.org/https://www.facebook.com/IMSTATImailto:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]://www.facebook.com/IMSTATIhttp://bulletin.imstat.org/mailto:[email protected]:[email protected]:[email protected]


3/36

= access published papers online

IMS Journals and Publications

Annals of Statistics:Peter Hall and Runze Li

http://imstat.org/aoshttp://projecteuclid.org/aos

Annals of Applied Statistics:Stephen Fienberghttp://imstat.org/aoas

http://projecteuclid.org/aoas

Annals of Probability:Krzysztof Burdzyhttp://imstat.org/aop

http://projecteuclid.org/aop

Annals of Applied Probability:Timo Sepplinenhttp://imstat.org/aap

http://projecteuclid.org/aoap

Statistical Science:Jon Wellnerhttp://imstat.org/sts

http://projecteuclid.org/ss

IMS Collectionshttp://imstat.org/publications/imscollectio

http://projecteuclid.org/imsc

IMS Monographs andIMS Textbooks:David Chttp://imstat.org/cup/

IMS Co-sponsored Journals andPublications

Electronic Journal of Statistics:George Michailidhttp://imstat.org/ejs

http://projecteuclid.org/ejs

Electronic Journal of Probability:Michel Ledouxhttp://ejp.ejpecp.org

Electronic Communications in Probability:Anton Bovier

http://ecp.ejpecp.org

Current Index to Statistics:George Styanhttp://www.statindex.org

log into members' area at imstat.org

Journal of Computational and Graphical StThomas Lee

http://www.amstat.org/publications/jcgslog into members' area at imstat.org

Statistics Surveys:Donald Richards

http://imstat.org/sshttp://projecteuclid.org/ssu

Probability Surveys:Laurent Saloff-Costehttp://imstat.org/ps

http://www.i-journals.org/ps/

IMS-Supported Journals

Annales de l'Institut Henri Poincar (B):ThierrBodineau & Lorenzo Zambotti http://imstat.org

http://projecteuclid.org/aihp

Bayesian Analysis:Marina Vannuccihttp://ba.stat.cmu.edu

Bernoulli:Eric Moulineshttp://www.bernoulli-society.org/

http://projecteuclid.org/bj

Brazilian Journal of Probability and StatisticNancy Lopes Garcia http://imstat.org/bjps

http://projecteuclid.org/bjps

Stochastic Systems: Peter W Glynnhttp://www.i-journals.org/ssy/

IMS-Affiliated Journals

ALEA: Latin American Journal of ProbabilityStatistics: Servet Martinez

http://alea.impa.br/english

Probability and Mathematical Statistics: K. BM. Musiela, J. Rosiski, W. Szczotka, & W.A. Woyczys

http://www.math.uni.wroc.pl/~pms

IMSBulletin.3September .2013

IMS Members News

If you hear of the

passing of an IMS

member, please

let us know so we

can organize an

obituary. Email

[email protected]

Eswar Phadia retires

Eswar Phadia hasretired rom the

William Paterson

University o NewJersey, and joinedthe Departmento Statistics at

Wharton School oBusiness, Universityo Pennsylvania, asa Visiting Scholar.His email address isephadia@wharton.

upenn.edu.

IMS Special Invited Lectures in 2014

Each year the IMS Committee on Special Lectures selects a number o leading statisticiansand probabilists to give invited lectures at IMS sponsored and co-sponsored meetings aroundthe world. In 2014, there will be a Wald, Neyman and Blackwell lecture, and eight Medallionlectures.

Te Wald Lectures will be given byThomas G. Kurtz, University o WisconsinMadison at the IMS Annual Meeting in conjunction with the Australian StatisticalConerence, held July , , in Sydney. At the same meeting will be a Neyman Lecturerom Peter Donnelly, University o Oxord; Schramm Lecture romTerry Lyons, Universityo Oxord; and five Medallion Lectures, given by Nina Gantert, echnische UniversittMnchen; Martin Hairer, University o Warwick;Timo Sepplinen, University o

WisconsinMadison; Matthew Stephens, University o Chicago; and Harrison Zhou, YaleUniversity.

At the WNAR/IMS meeting in Honolulu, June , , there will be a MedallionLecture romTilmann Gneiting, Universitt Heidelberg.

Te 37th Conerence on Stochastic Process and their Applications, in Buenos Aires ( July28August 1, 2014) will eature a Medallion Lecture rom David Nualart, University oKansas.

Last but not least, at the Joint Statistical Meetings in Boston (August , ), therewill be the inaugural Blackwell Lecture, by Gareth Roberts, University o Warwick, as well asa Medallion Lecture rom Mathias Drton, University o Chicago.

Shahjahan Khan elected to Bangladesh Academy of Science

Shahjahan Khan, University o Southern ueensland, Australia,has recently been elected as an Expatriate Fellow o the Bangladesh

Academy o Science, the leading scientific organization in Bangladeshand a leading representative to the government in matters o scienceand research.

Its an excellent eeling to be recognised or your work bymembers o the scientific community in the country o your birth,Proessor Khan said. Te academy engages with the government inthe development and promotion o science within Bangladesh andthe academys input is highly valued by the policy makers. I can alsohelp with organising seminars, conerences and workshops such as the

International Statistics Conerences that I helped organize in and in Bangladesh.Proessor Khan has served as the President o ISOSS (),

an international proessional organization o statisticians, organisedour international conerences in Malaysia, Egypt, Pakistan and Qatar,and has served as the Founding Chie Editor o theJournal o Applied

Probability and Statistics since .Te Bangladesh Statistical Association awarded Proessor Khan

their prestigious Qazi Motahar Hossain Gold Medal in inrecognition o his scientific contribution and promotion o statistics.
http://imstat.org/aoshttp://projecteuclid.org/aoshttp://imstat.org/aoashttp://projecteuclid.org/aoashttp://imstat.org/aophttp://projecteuclid.org/aophttp://imstat.org/aaphttp://projecteuclid.org/aoaphttp://imstat.org/stshttp://projecteuclid.org/sshttp://imstat.org/publications/imscollections.htmhttp://projecteuclid.org/imschttp://imstat.org/cup/http://imstat.org/ejshttp://projecteuclid.org/ejshttp://ejp.ejpecp.org/http://ecp.ejpecp.org/http://www.statindex.org/http://www.amstat.org/publications/jcgshttp://imstat.org/sshttp://projecteuclid.org/ssuhttp://imstat.org/pshttp://www.i-journals.org/ps/http://imstat.org/aihphttp://projecteuclid.org/aihphttp://ba.stat.cmu.edu/http://www.bernoulli-society.org/index.php/publications/bernoulli-journal/bernoulli-journalhttp://projecteuclid.org/bjhttp://imstat.org/bjpshttp://projecteuclid.org/bjpshttp://www.i-journals.org/ssy/http://alea.impa.br/englishhttp://www.math.uni.wroc.pl/~pmsmailto:[email protected]:[email protected]://www.math.uni.wroc.pl/~pmshttp://alea.impa.br/englishhttp://www.i-journals.org/ssy/http://projecteuclid.org/bjpshttp://imstat.org/bjpshttp://projecteuclid.org/bjhttp://ba.stat.cmu.edu/http://projecteuclid.org/aihphttp://imstat.org/aihphttp://www.i-journals.org/ps/http://imstat.org/pshttp://projecteuclid.org/ssuhttp://imstat.org/sshttp://www.amstat.org/publications/jcgshttp://www.statindex.org/http://ecp.ejpecp.org/http://ejp.ejpecp.org/http://projecteuclid.org/ejshttp://imstat.org/ejshttp://imstat.org/cup/http://projecteuclid.org/imschttp://projecteuclid.org/sshttp://imstat.org/stshttp://projecteuclid.org/aoaphttp://imstat.org/aaphttp://projecteuclid.org/aophttp://imstat.org/aophttp://projecteuclid.org/aoashttp://imstat.org/aoashttp://projecteuclid.org/aoshttp://imstat.org/aoshttp://www.bernoulli-society.org/index.php/publications/bernoulli-journal/bernoulli-journalhttp://imstat.org/publications/imscollections.htm


4/36


The XL-Files: Rejection Pursuit

Photo: Harvard University

Contributing Editor Xiao-Li Meng writes:

As an assistant proessor in late ,working on a unit-root AR() model,I came across a cute way o calculatingmoments with negative or ractional powersby integratingnot differentiatingamoment-generating unction (MGF). Likemany junior researchers, I was eager to turnmy findings into a publication. By January o, A Note on Finite-Sample Evaluationso Moments o Ratios was ready or col-leagues comments.

I love creative titles. But thatwasanything but creative! o make mattersworse, my note started with A Few UseulLemmas, and I submitted it toJASA.Clearly, I had little clue aboutJASAs scopeor its readership.

Te rejection letter arrived in mid-April,three weeks afer my submission. Te letterreflected well the care that Editor GeorgeCasella and an Associate Editor (AE) hadtaken towards an inexperienced and impa-

tient junior researcher:We both eel that the paper is really anice piece o work, and the real task will befinding it the right home. (Te associate edi-tors letter to me was very complimentary othe paper.) My eeling is that your best shotmight be Statistical Science. You probably willhave to do a major revision, but it may be

worth it. Good luck.I couldnt possibly reject this rejection.

Tey acted very promptly, and the AEs

report complimented appropriately the schol-arship o the paper, but reasoned effectivelythat I chose a wrong journal: ManyJASAreaders will stop right at the first line: SomeUseul Lemmas.

Did I learn the lesson? Well, not really. Itook the suggestion to try Statistical Science(SS)but did not bother to revise. In mydeense, I did not ormally submit but rathersent it with a letter o inquiry, expressingmy willingness to revise should the topicbe ound suitable or SS. Inquiring aboutsuitability can be effective, but in this caseit sent the rather unsubtle message that I

was interested more ingetting the paperpublishedthan actually improving the paper, adistinction that unortunately I was not theonly (junior) researcher to ail to appreciate.Consequently, SSs rejection letter came

within a week, together with a Desiderataor Statistical Science, a rather unsubtle mes-sage in returndo some homework beore

submitting!Did I learn the lesson this time? Well,read on. During the next ew months Iextended the results rom ratios to loga-rithms, and I elt the paper was substantialenough to drop the note distinction andsubmit it to theAnnals o Statistics(AOS).Te title was equally plain, On AnalyticEvaluations o Moments o Ratios andLogarithms, and the paper still started withA Few Useul Lemmasthat may turn off

JASA or SSreaders, but surely it should turnon theAOSreviewers shouldnt it?AOSs rejection arrived in November o

, a year afer my submission (perhaps asign o serious interest byAOSs reviewers?).Te main reason or rejection was although

you have provided a very nicely writtenreview and exposition, there is not enoughthat is new. Again, I couldnt declare this wasan unjustified decision (orAOS). Whereasthe logarithm results were new to the best

o my knowledge, my initial discoveryturned out to be known in the econometricsliterature, though my article connected manydots and ound some new applications.

My rejection pursuit thereore continuedI submitted it toInternational Statistical

Review(ISR), in February o , afer acosmetic revision. A sof rejection came inApril , suggesting reducing it to a crispnote ocusing on the AR() application. Soafer all the trouble and time, I was back atsquare one!

TeISRs reviewer report summarizedwell the reason or my ongoing rejectionpursuit: Although the author really does tryhard, there is no clear purpose o this article.Indeed, without a clear purpose, more effortcan only lead to more rejections! Finally themessage sank in: stop publication hunting,but think hard about what message I should,and could, get across.

Deep thinking takes time, and in this case

it took five years. Te next version, FromUnit Root to Steins Estimator to Fishers kStatistics: I You Have a Moment, I Can ellYou More was completed in September, and by then it had a clear aim to fill ina missing chapter in textbooks treatments oMGFs. I was confident enough to re-submitit to SS(partly because George Casella hadbecome its editor, so I was merely ollowinghis suggestion made seven years earlier!).

aking my time paid offthe AE sum-

marized the reviewers reaction as ollows:Tey called it lively, interesting, entertainingand cute.

Te article appeared in print in , adecade afer its conception.

Not every paper or project can affordsuch a long period o contemplation,nor should it. Perectionism can turn to

paralyzing procrastination. Te five-yearhiatus would be extreme or topics that aretime-sensitive (and indeed I had plenty o


5/36

September .2013 IMSBulletin.5

those during those five years), but in this caseit was a therapeutic reinorcement to resistthe seduction o increased quantity at therisk o a reduction in quality. Indeed, thelessons I have learned as an author, reviewerand editor over two decades have convinced

me that both our proession and we ourselvesare better served i we all spend a bit moretime on improving a rejected paper than ongetting it published somewhere.

Te Seven Sins course handoutreproduced below was designed with that

goal in mind (and it was distributed duringthe Starting Research Career Panel at JSM).

So please, take your time. Publishing apaper may be hard, but un-publishing one isimpossible.

Seven Deadly Sins of Research Papers, Seven Virtues to CultivateFrom Stat366 Research Cultivation & Culmination at Harvard Statistics, taught by Joseph Blitzstein and Xiao-Li Meng

Sins Leading to Rejection Whats in the Referees Mind Virtues Leading to Acceptance

Insufficient Innovation:minor varia-tions; routine exercise; tedious deriva-tion of something not exciting; obviousgeneralizations

Theres not much hereWell, this looks like a term project!

I dont learn anything new or interesting.

TAKE TIME to develop more elegant andsubstantial methods and theory; workout more applications; include morecomparisons with existing methods

Narrow Scope:possibly good or evendeep results presented in a very narrowor incomprehensible way; very limitedreadership

How many people would ever care aboutthis?

This is really too specialized to deserve pagesin this journal.

TAKE TIMEto expand the scope;explain better how the work connectsto a larger area; if not possible, considersubmitting to a specialized journal

Premature Development: containsonly half-baked ideas or immaturemethods; very loose organization; lackof a coherent theme

Well, there could be something interestinghere, but the author really needs to do a lotmore work to convince me!

TAKE TIME to think through what canbe fully developed and what should besaved for a future paper; develop a lucidflow; write an exploratory overview

paperSloppy Research:hand-waving/incor-rect arguments presented as theoreticalproof; misguided reasoning or simula-tion configurations

This person is not a very good researcher

Whats new is not correct, and whats correctis not new!

TAKE TIME to double-check each der-ivation; think twice, deeper, and fromdifferent angles; look for independentways to validate results; try out simplecases

or to Tentative Rejection

Inadequate Literature (Re)Search:missing major citations; insufficientcomparison with existing methods;reinventing wheel

Clearly the author has not done his/herhomework!

How could the author not know this???

What about mywork?

TAKE TIME to do a thorough literaturesearch, possibly outside of statistics;provide enough honest comparisonswith major competitors; give due credit!

Unwarranted Claims: exaggerates theadvantage or applicability of proposedmethods; inappropriately downplaysothers work

Gosh, this author is full of himself/herself!This is just too much, and too good to betrue!

This really p***** me off!

TAKE TIME to support each claim withtheoretical and/or empirical evidence;separate facts from speculations orhope; show rather than tell the advan-tages; give credit

Sloppy Presentation:reads likeresearch notes, abrupt with bad flow;lots of typos and/or grammatical errors;inconsistent notation, captions, andlabeling; missing/incorrect references

Well, Im not going to waste my time whenthe author does not spend time proofreadinghis/her own paper!

Who is going to read such a badly writtenpaper??

TAKE TIME to proofread at least fivetimes; critically examine the flow andthe order of presentation; ensuresmooth transitions between segments;meticulously check for correctnessand consistency; find a native writer toproofread/comment


6/36


Jnos Englnder reports: Te thConerence on Stochastic Processes andTeir Applications was held on the beautiulcampus o the University o Colorado,Boulder, rom July to August , .Tere were participants. Te interna-tional scientific committee was chaired bySteve Lalley (Chicago), and the local organiz-ers were Anne Dougherty, Vanja Dukic, JnosEnglnder, Sergei Kuznetsov, Manuel Lladserand Brian Rider (chair).

At the beginning o the conerenceDavid Aldous announced that the Line and Michel Love International Prizein Probability was awarded to SouravChatterjee o the Courant Institute, NewYork (see Members News on page ).

Named lectures were given by GrardBen Arous (Courant, New York University),

who gave the Lvy Lecture; Itai Benjamini(Weizmann Institute o Science), who

was the Inaugural Schramm Lecturer;

Ken Golden (University o Utah), theBernoulli Society Open Lecture; NeilOConnell (University o Warwick) theDoob Lecture; Hiroumi Osada (KyushuUniversity) the It Prize Lecture; andBlint Virg (University o oronto), who

was an IMS Medallion Lecturer. Te otherplenary speakers were: Zhen-Qing Chen(University o Washington); Ron Doney

(University o Manchester); Hugo Duminil-Copin (Universit de Genve); adahisaFunaki (University o okyo); Niels Jacob(Swansea University); Vadim Kaimanovich(University o Ottawa); Jeremy uastel(University o oronto); Kavita Ramanan(Brown University); Qi-Man Shao (ChineseUniversity o Hong Kong); Amandine Vber(cole Polytechnique); and Oer Zeitouni(University o Minnesota & WeizmannInstitute o Science). Pablo Ferrari unortu-nately had to cancel his plenary address.

Te inaugural Schramm lecture, by I.Benjamini, was in memory o the outstand-ing Israeli-American mathematician, OdedSchramm, known or the invention o theSchrammLoewner evolution, who lost hislie in a hiking accident in his orties. TeIt Prize Lecture was delivered by the recent

winner o the prize, H. Osada. Te BernoulliSociety Open Lecture, by K. Golden, was apublic lecture, aimed at a broader audience;

it addressed environmental as well as mathe-matical problems.A large number o invited and contrib-

uted sessions complemented the plenarytalks, and a successul Early ResearcherPanel Discussion took place on the uesday.

According to eedback rom the partic-ipants, the scientific program was very wellreceived, and those who participated either in

the organized Wednesday hike in the Rockiesor in one o the more privately organizedones, were ascinated by the natural beauties

o the conerence venue. Te conerencedinner took place at Hotel Boulderado, alandmark hotel in Boulder.

Te meeting was sponsored by IMS, theBernoulli Society, International Society orBayesian Analysis, Elsevier, Illinois Journal oMathematics, National Science Foundation,National Security Agency, US ArmyResearch Office, Department o the US Navy

Microsof Research,and urthermore

by the Universityo Colorado as wellas its Departmentso Mathematicsand AppliedMathematics. Allsupport receivedrom the aboveagencies is grateullyacknowledged by theorganizers.

Stochastic Processes and their Applications

Photo:AdrianaOcejo

Above: Itai Benjamini delivered the Inaugural SchrammLecture. Below: Medallion Lecturer Blint Virg


7/36


New Researchers Conference

Jingchen Liu (above right, receiving his plaque from then IMS President Hans Knsch)delivered this years Tweedie New Researcher Lecture on rare-event analysis for nonlin-ear functionals of Gaussian random fields

The 15th IMSNew Researchers

Conference in

Montral featuredtwo sessions ofcreative five-minute

spotlight talks byposter presenters,

which were followedby poster sessions

Johanna G. Nelehov and Aarti Singh report on the th IMS New

Researchers Conference, jointly sponsored by the IMS and the SSC:

Te th edition o the IMS New Researchers Conerence took placeat the Centre de recherches mathmatiques (CRM) in Montral,rom August to August , . Because o its Canadian location,this years meeting was co-sponsored by the Statistical Society oCanada. In total, young researchers participated in the coner-ence women and menrom the United States, Canada,France, Switzerland, and Japan.

Te first two days o the meeting were tightly packed with talksand posters by the new researchers, and our invited lectures. Afera cordial welcome by Hans Knsch and Christian Genest (ISM &CRM), erry Speed opened the meeting with a thought-provokingaddress entitled Mathematical Statistics: How long can and should

you stay in it? Te lively discussion that ollowed got us all warmedup or a packed program o short talks on mathematical statistics, bio-statistics, high-dimensional issues, spatio-temporal modeling, surveysampling, and genetics. Te short talks were alternated with threeinvited lectures: on Tursday afernoon, Aurore Delaigle exploredclassification using censored unctional data; on Friday morning, JeffRosenthal introduced us to the puzzles o MCMC; Steve Fienbergdiscussed the intricacies o causes o effects in the afernoon. Both

Tursday and Friday were crowned by creative five-minute spotlighttalks by poster presenters, ollowed by poster sessions. On Fridaynight, we could all finally enjoy a rich dinner and admire the magnifi-cent view rom the roofop restaurant Le Cercle o HEC Montral.

Saturday started with the weedie Award Invited Lecture onrare-event analysis or nonlinear unctionals o Gaussian randomfields, given by this years winner Jingchen Liu. Aurore Delaigle andSteve Fienberg then offered advice on publishing, while erry Speedand Hans Knsch shared their experiences with the mentoring o

graduate students. Afer lunch, Hans Knsch explained why we needscholarly societies like the IMS. Jeff Rosenthal and Larry Wassermanthen offered tips on how to ace the challenges in teaching probabilityand statistics. Te meeting ended with a longer panel on unding,led by Dave Stephens and various representatives rom undingagencies: Nandini Kannan (NSF), Madeleine Bastien (NSERC), Joe

McCloskey (NSA), Michelle Dunn (NIH), and Pierre des Lierres(Mitacs).Te meeting was sponsored by the National Science Foundation,

National Institutes o Health, Centre de recherches mathmatiques,the Institut des sciences mathmatiques, and the Statistical Society oCanada. We also obtained generous assistance in all administrativematters rom the Centre de recherches mathmatiques.

More details about the meeting, as well as photographs, can beound at http://www.math.mcgill.ca/nrc2013/.
http://www.math.mcgill.ca/nrc2013/http://www.math.mcgill.ca/nrc2013/


8/36


OBITUARY:Maurice Priestley

19332013

Maurice Bertram Priestley

M B P, Proessoremeritus o the University o Manchester,UK, was an outstanding, influential andhighly respected figure in the field o timeseries analysis. His book on Spectral Analysis

and ime Series, first published in and reprinted several times, has become astandard reerence on the subject. In the

preace o the book, Maurice Priestley statedthat, or the most part, he ollowed the styleo theappliedmathematician, but there arecertain basic results (such as the spectralrepresentation theorem) which are crucial ora proper understanding o the subject, and bytheir nature, require a careul and precise pre-sentation. In his book one can see his subtleuse o real analysis and generalized unctions,subjects o which he was very ond, in deriv-ing the results. At the same time, one cansee his skill in relating the results to applied

problems arising in fields such as physics,engineering and economics. Besides beingan outstanding research worker, he was wellknown to be an excellent teacher. He wasone time associate editor o theJournal o the

Royal Statistical Society Series B (JRSS B) androm its oundation in until the end o he was editor-in-chie o theJournal oime Series Analysis (JSA), a leading journalin the field. He was elected as a member othe International Statistical Institute, and also

a Fellow o the IMS.Maurice Priestley had a lie-long loyaltyto Manchester, the city o his birth on thMarch . He attended the -year-oldrenowned Manchester Grammar Schooland went on to Jesus College, Cambridge,

where he was a wrangler in mathematics.At Cambridge he stayed on to take theDiploma in Mathematical Statistics in theStatistical Laboratory where his interests

were influenced by Dennis Lindley and

Henry Daniels among many others. From to he was Scientific Officer atthe Royal Aircraf Establishment, gaining

practical experience o the relatively newfield o spectral analysis and working or ashort time with Gwilym Jenkins, who was a

junior ellow there, and who would also goon to make his reputation in this area. Tey

published a joint paper inJRSS Bin and Maurice decided to study or a PhD. Onthe advice o George Barnard he returned toManchester to work under the guidance oMaurice Bartlett and at the same time tookup an assistant lectureship at the Universityo Manchester. He carried on his work onspectral analysis that he started with Gwilym

Jenkins and published two very influentialpapers on spectral estimation (echnometrics,) and the role and choice o bandwidthin the estimation o spectral density unction

(Applied Statistics, ), a topic whichreceived a great deal o attention in thecontext o probability density estimation andnonparametric regression. His reputation

was firmly established with a urther twopapers inJRSS B based on his PhD topic oninerence or mixed spectrathe detection orequencies in the presence o coloured noise.He published in a seminal paper (a read

paper to the Royal Statistical Society) onevolutionary spectra and non-stationary pro-

cesses. Tis paper stimulated urther researchin this area by others and led to the develop-ment o statistical tests or non-stationarity(Priestley and Subba Rao,JRSS B, ), todevelopment o time varying parameter linearARMA models and to the definition o the

weighted likelihood unction as a tool orthe estimation o time dependent parameters(Subba Rao,JRSS B, ). More recently,Dahlhaus (Annals o Statistics, ) and hiscolleagues extended these ideas urther or

the analysis o locally stationary processessatisying ARCH models, a topic which isnow receiving considerable attention.

During the s, Priestleys departmentbecame an exciting centre or time seriesresearch with the appointment, amongothers, o ong and Subba Rao whorapidly established their own reputationsin non-linear time series modelling. Onecan describe the s as a golden decadeor rapid developments in nonlinear andnon-stationary processes, many o theseemanating rom the Manchester School

under the leadership o Maurice Priestley(Bilinear Models o Subba Rao,JRSS B Treshold Models o ong and Lim,JRSS B, Amplitude Dependent AR models oHaggan and Ozaki,Biometrika , StateDependent Models o Priestley,JSA ).During the same period, Clive Granger andhis colleagues at Nottingham were publishing

papers in similar and related areas. ogether,Maurice Priestley and Clive Granger ormedthe ManchesterNottingham ime Series

group, which held requent meetings todiscuss new developments and stimulatedurther research in these areas. Tese meet-ings continued until Clive Granger movedin to the University o Caliornia, SanDiego. Priestley had been visiting proessorat both Princeton and Stanord Universitiesin the USA in the early s and through themany contacts he ostered and his personalstanding he attracted to his department asuccession o visitors o international repute


9/36


OBITUARY:Wenbo Li

19632013

Wenbo Li

W V. L, proessor o mathematicalsciences at the University o Delaware, diedsuddenly o heart ailure on January , ,near his home in Newark, Delaware. He was

years old.Wenbo Li was born on October , in Harbin, China. Afer obtaining hisBachelors degree in applied mathematicsrom Jilin University in Changchun, China,he came to the United States in andstudied probability theory at the Universityo WisconsinMadison, under the supervi-sion o James Kuelbs. Wenbo Li obtained hisPhD in mathematics in and accepteda position at the University o Delaware as

an assistant proessor, becoming an associateproessor in and a ull proessor in. At the time o his death, he was also anadjunct proessor in Delawares Departmento Electrical and Computer Engineering(since ) and Department o AppliedMathematics and Teoretical Physics (since); a visiting proessor at the Instituteo Applied Mathematics o the ChineseAcademy o Sciences (since ); and anadjunct proessor at Harbin Institute o

echnology, China (since ).During his career as a research

mathematician, Wenbo Li had also heldvisiting and other academic positions at

various universities and research institutions,including Hong Kong University o Scienceand echnology, Peking University, theUniversity o Pennsylvania, and exas A&MUniversity. He also served as an associate edi-tor or several probability journals, including

Annals o Probability, Journal o Teoretical

Probability, Journal o Mathematical Research

and Exposition, andInternational Journal oStochastic Processes.

Wenbo Li was a probabilist who studied

Gaussian measures and processes and pub-lished more than research papers in theseand related areas. He was a leading expert onsmall ball probability estimates: he contrib-uted significantly to our understanding othe connection between the small ball prob-ability and the entropy number. He workedor many years on the Gaussian correlationconjecture concerning a lower bound o theGaussian measure o the intersection o twosymmetric convex sets and was especially

proud o a weak orm o this conjecture thathe proved. In , he was elected a ellowo the IMS, or his distinguished researchin the theory o Gaussian processes and inusing this theory to solve many important

problems in diverse areas o probability.Besides his own research, Wenbo Li was

very helpul to young researchers rom Chinaby hosting them in his department and work-ing jointly with them to start their careers asresearch mathematicians.

Wenbo Li was a nature lover and enjoyedhiking in mountains in China and on theGreat Wall. All his riends and acquaintanceselt the immensity o his energy and hiseagerness or activities. Wenbo Li is survivedby his wie Sunny and his son James.

Elton P. Hsu, Northwestern University

in time series research, including H. Akaike,. W. Anderson, E. J. Hannan, E. A. Parzen,M Rosenblatt, ohru Ozaki and G. Wahba.

Maurice Priestley was appointed to ullProessorship in and served as Heado the department o Mathematics at theUniversity o Manchester Institute o Scienceand echnology (UMIS) or several peri-ods, in total over years. He figured prom-inently in the UMIS administration, andserved with distinction on many committees.He was instrumental in the ormation othe joint ManchesterSheffield School oProbability and Statistics, being appointedHonorary Proessor at the University oSheffield. In , ollowing his retirement

rom UMIS, he was appointed EmeritusProessor in the University o Manchester.

Maurice was a Hi-Fi enthusiast and keenamateur radio operator (related to signal

processing interests and not unconnectedwith time series?). He was more interestedin the technical side o Hi-Fi equipment (orexample the requency response unctionso the sound systems such as speakers andamplifiers rather than actual sound itsel ).Similarly his interest in gol, where he wasmore interested in flight dynamics o the golball and optimum position to hold the golclub or a specific shot. He was a dedicatedsupporter o Manchester United ootball(soccer) club and any discussion related to

the club, its manager, or the game couldbecome interesting. He also had clear and

well thought-out views on his own specialitybut in his leadership o his department andeditorship o theJournal o ime Series

Analysiswas moderate and o a generousleaning. He earned widespread respect, both

personal and proessional, rom academiccolleagues.

He married Nancy Nelson in andthey have two children, Michael and Ruth,and our grand children. He died on th

June ollowing a long illness.ata Subba Rao, University o Manchester

and Granville unnicliffe-Wilson

Lancaster University


10/36


IMS Presidential Address:Ars Conjectandi300 Years Later

Hans R. Knsch, ETH Zurich, completed his term as IMS President

at the Joint Statistical Meetings in Montreal. He delivered this

Presidential Address:Tis year we are celebrating the anniversary o a book published years ago, and on this occasion we have declared to be theInternational Year o Statistics. So the book gives us the opportunityto attract the attention o a larger public to statistics, but what rele-

vance does it have or us statisticians and probabilists today?Some might argue, very little: First, the title,Ars Conjectandi, the

Art o Conjecturing, seems odd or a book about scienceafer all,science and arts have very different standards. And isnt the booksmain result just the weak law o large numbers or the binomialdistribution? We can prove that theorem in much greater generality ina ew lines. Such arguments miss however important points. With his

Ars Conjectandi, Jacob Bernoulli has laid an important oundation orour fields, and I want to use it tonight or the ollowing purposes: to look back how our subject evolved, to appreciate how ar we have come, to better understand the conceptual difficulties with probability

and statistics applied to real world problems, to compare how science was done in the past with how it is done

nowadays, to speculate about possible directions o our field.I believe that such thoughts also fit well with the efforts o IMS tohighlight the rich history o probability and statistics through ourScientific Legacy project, although that project will not go back so arin the past.

Let me first recall briefly the historical context oArs Conjectandi:Te scientific approach to randomness starts in with anunpublished correspondence between Pascal and Fermat, ollowed in by the first book, written by Huygens. Tey dealt with gameso chance where probabilities could be determined by symmetry

arguments, without relying on observations.In Graunt published his Observations made upon the billso mortality, and in Halley publishedLie tables with seven uses.Tese works compute relative requencies o survival given a certainage, ignoring the underlying random variation.

Jacob Bernoullis bookArs Conjectandibrought these twodevelopments together by giving a mathematical argument or theconvergence o relative requencies to probabilities.

It laid thereore the oundation or the application o probabilitytheory in situations outside o pure games o chance. More impor-tantly, because Bernoulli gave specific bounds on the required number

o observations or a desired precision, his results made it possible toquantiy the uncertainty in the estimation o probabilities by relativerequencies. In a wider perspective, the book gives a justification orinductive reasoning: How can we derive mathematical models romdata?

In a less known ourth part, Jacob Bernoulli also discussesprinciples or the application o probability theory in political,judicial, moral and business matters. He was convinced that a preciseknowledge o probabilities was the basis or predictions, judgmentsand actions. He could not have imagined the many other applicationso probability in the natural sciencesstatistical physics, quantummechanics and genetics were completely unknown in his time.

Beore we look more closely at some o the conceptual issues, Iwould like briefly to describe the personality o the author and somecircumstances regarding how the book was written.

Te Bernoulli amily was o Dutch origin and had moved toBasel in Switzerland or religious reasons. Jacob was the first o manyscientists that this amily produced over generationsall o themmale; the first emale member o the Bernoulli amily that I couldfind on the web is Eva Bernoulli, , who was active in thetemperance movement against alcohol. Presumably there were other

emale Bernoullis, and we can only wonder what they might havecontributed to science i they had had the same access to education asthe men.

Jacob originally studied theology at his athers wishes, but laterbecame a mathematician. We can guess at his relationship with hisather rom the motto that he adopted: In spite o my ather I amamong the stars. His younger brother Johann studied first medicine,but then also switched to mathematics. At the beginning he wastaught by Jacob, and together they became leaders in the applicationo the new infinitesimal calculus, developed by Newton and Leibniz,to problems in geometry and physics. Te two brothers solved a num-

ber o the most difficult problems o their time, but their cooperationturned quickly to a bitter rivalry which they ought out ruthlessly intheir publications.

I will not go into the details o this clash, but just quote a ewcharacterizations o the two brothers used by their biographers: abilious and melancholic temperament, mostly quarrelsome and

jealous, or violent, abusive and, when necessary, dishonest. Wemay ask i the way we deal with competition in research has changedin the past years? Difficult characters are presumably unavoidablein science, but this is no excuse or personal attacks or polemics. Westill have to continue our efforts to create an environment or science


11/36


which is characterized by air play and mutual respect.Jacob Bernoulli did the main work on hisArs Conjectandirom

to , but he did not publish it during his lietime. He diedin , and the book appeared only years later. Te delay between

Jacobs death and the publication o the book is mainly a consequenceo the rivalry with his brother, but why did he not publish it duringhis lietime? Historians believe that this is primarily because he wasnot satisfied with it: Te bounds that he had are not tight and thusthe required sample sizes or a precise estimation are huge. Moreover,he wanted to obtain some data to illustrate the arguments in hisourth part. Who would still do this nowadays and resist the pressureto publish quickly despite the eeling that he or she hasnt yet ullyunderstood the problem?

Now let me discuss some o the conceptual difficulties relatedtoArs Conjectandi. Te first one is the meaning o probability.Bernoullis definition is still my preerred one: Probability is a degreeo certainty, and differs rom certainty as a part rom the whole.In the weak law o large numbers, uncertainty is again expressed asa probability. Tus or the interpretation o this law, we need theconcept that Events with low probability do not occur in a singleoccasion. Bernoulli ormulated this idea using the term morally

impossible. But how small should the probability o an event be sothat we are morally certain that it will not happen? Tis questionis still with us today in the assessment and communication o risk.For instance, during my presidency I had to decide whether IMSshould take a position about the conviction o scientists because otheir statements beore the earthquake in lAquila, Italy. Te issue wasnot whether earthquakes can be predicted, but rather whether thescientists had weighed and communicated the evidence according tothe most recent scientific knowledge. [Hans recommended attending

David Spiegelhalters public lecture given a couple o days later,

From Gambling to Global Catastrophe: Metaphors and Images or

Communicating Numerical Risks]Let me next turn to the issue o inductive reasoning. Since

Bernoulli there have been many attempts to understand ar morecomplex systems than a simple urn by analyzing empirical data. Inthe era o Big Data, data are available almost without limitations, andexpectations are high.

For instance the European Journal o Physics published last yeara Maniesto o computational social science which contains state-ments like the ollowing: Inormation and communication technologies (IC) produce a flood

o data. Tese data represent traces o almost all kinds o activitieso individuals enabling an entirely new scientific approach or social

analysis.

Te analysis o huge data sets as obtained, say, fom mobile phone

calls, social networks, or commercial activities provides insight into

phenomena and processes at the societal level.

IC can greatly enhance the possibility to uncover the laws o the

society.

Te role o computational social science is a leading one in addressing

the Big Problems o society, avoiding crises and threats to its stability

and healthy development.

Most o us will agree with the first two statements, but I have somedoubts about the other two. O course, picking out a ew sentencesrom a -page article is not air, but I believe the authors should dis-cuss much more extensively questions such as: What are these laws othe society, what distinguishes them rom the laws o physics, how canthe idea o social laws coexist with concepts o individuality and reechoices, and what are the limits o predictability and controllability o

phenomena and processes at the societal level?Neuroscience is another field where hope is high that by exploit-

ing available data we can gain insight, discover undamental laws andfind cures or diseases like Alzheimers and Parkinsons.

Left: a first editionof Jacob Bernoullisposthumously pub-lished 1713 book ArsConjectandi sold for

$20,000 at auction in2008. It had been partof the Richard Greenlibrary, a private col-lection of hundredsof rare scientific andmathematical works.The Christies listingis at http://www.christies.com/salelanding/index.aspx?intSaleID=21644
http://www.christies.com/salelanding/index.aspx?intSaleID=21644http://www.christies.com/salelanding/index.aspx?intSaleID=21644http://www.christies.com/salelanding/index.aspx?intSaleID=21644http://www.christies.com/salelanding/index.aspx?intSaleID=21644http://www.christies.com/salelanding/index.aspx?intSaleID=21644http://www.christies.com/salelanding/index.aspx?intSaleID=21644http://www.christies.com/salelanding/index.aspx?intSaleID=21644http://www.christies.com/salelanding/index.aspx?intSaleID=21644http://www.christies.com/salelanding/index.aspx?intSaleID=21644http://www.christies.com/salelanding/index.aspx?intSaleID=21644


12/36


Te website o the Blue Brain Project, which recently wasapproved as an EU flagship project, describes their approach asollows:

Neuroscience:systematic, industrial-scale collection o experimental

data, making it possible to describe all possible levels o structural and

unctional brain organization fom the sub-cellular, through the cellular,

to the micro-circuit, meso-circuit and macrocircuit levels;

Neuroinformatics:automated curation and databasing o data,

use o Predictive Reverse Engineering to predict unknown data fom a

smaller sample o known data or fom data describing other levels o

brain organization;

Mathematical abstraction:definition o parameters, variables,

equations, algorithms and constraints representing the structure and

unctionality o the brain at different levels o organization.

Again, I am a bit skeptical that systematic collection o data willallow us almost automatically to iner the unctioning o the brainand to predict what happens at a different levels.

Most o us are presumably also surprised that the document usesthe term Predictive Reverse Engineering or tasks that we would con-sider as the central domain o statistics. Indeed, there is some concernthat statistical knowledge is not ully recognized by other branches o

science. I have heard this concern repeatedly in the + years o mycareer, the first time being when uzzy systems and neural networkswere the hot topics in science. I am thereore not too worried aboutthe uture o our fields, but I do believe that we should encouragemore people to reach out towards new disciplines, to develop andstudy new methods to address the needs o other disciplines. Weshould also have more appreciation or such interdisciplinary contri-butions and achievements.

In the remaining time, I dont want to discuss this point urther,but I would like to describe to you briefly how the knowledge aboutanother complex systemthe weather and climatehas progressed

in the past years. Tis topic fits well into this years celebrationo Mathematics o Planet Earth. My brie thoughts will also provideillustrations o the roles and interplays between data and theory,induction and deduction, stochastic and deterministic thinking.

I begin my story with Robert Fitzroy, who lived rom till. He is presumably unknown to you. He was the captain oHMS Beagle, the ship o Charles Darwins amous voyage, which gavehim lots o experience with weather and storms at sea. In he

was appointed as chie o a new government agency to deal with thecollection o weather data at sea, with the goal to make shipping lessdangerous.

Fifeen land stations were established to use the new telegraphto transmit daily reports o weather at set times. Fitzroy developedcharts to allow predictions to be made rom these data. Te first daily

weather orecasts were published in Te imes in . In analogyto inormation and communication technology today, he used thenewest technological means available in order to collect as much

inormation as possible in order to gain insight and to solve a practicaproblem.In he published Te Weather Book: A Manual o Practical

Meteorology,in which he attempted to uncover the laws governingthe weather rom his lie-long experience . Te success o his efforts

was however limited. Francis Galton, who had also an interest inmeteorology, wrote a devastating review o that book: It is a ault ina book intended to lay the oundations o a new experimental sciencethat it should be mainly occupied with deductions rom unprovenhypotheses, instead o the careul establishment o axioms by rigorousinduction rom observed acts.

IMS Presidential Address continued

Continues fom page

Incoming IMS President Bin Yu (left) takes the gavel from Hans Knsch, in the ceremonythat marks the passing on of the presidency, which takes place at the IMS AnnualMeeting, held this year at JSM in Montreal.


13/36


Despite Galtons assertion, big breakthroughs did not happenthrough a more sound induction rom observed acts. Tey wereachieved through an entirely different approach, initiated by theNorwegian Vilhelm Bjerknes (). He was the first to realizethat the undamental laws o fluid dynamics and thermodynamicscan be used to describe the large scale motions in the oceans and theatmosphere. In , he declared that weather orecasting is possiblein principle by solving deterministic partial differential equations.Tis was very much ahead o his time, and the first attempt to imple-ment the idea ended in a ailure. Computing power was simply notsufficient.

Te situation changed with the advent o computers in the s.Since then, deterministic differential equation models are at the coreo all scientific weather prediction services. Moreover, basically thesame models can be used also or climate studies, provided one takesinto account the interactions o the atmosphere and the oceans withthe biosphere and the cryosphere, the part o the earth where water isrozen.

However, as we all know, despite the huge progress in computingpower, both weather and climate predictions are still uncertain.Atmospheric physicists are nowadays aware that they have to quantiy

this uncertainty. Te uncertainty has two main reasons. Firstly, inview o the large range o space and time scales o processes in theatmosphere, the problem o how to deal with phenomena at smallerscales than the numerical resolution allows will not disappear.Secondly, in Lorenz discovered the chaotic nature o the

weather: sensitivity to initial conditions limits its predictability, eveni we could solve the equations exactly.

Te solutions to these difficulties are parametrizations and dataassimilation. Without going into technical details, parametrizationsare methods to approximate the effect o unresolved scales asunctions o resolved variables. Data assimilation is the sequentialadjustment o the current computed state o the atmosphere, based onobservations. In the next prediction step, this current state is used asnew initial condition.

Te relevant point or this talk is that or both parametrizationand data assimilation, advanced methods are stochastic: Stochastic

parametrization is a hot topic in the field, and ensemble methodswhich replace a single current state o the atmosphere by a samplein order to quantiy the uncertainty o predictions are nowadays

widespread. Tese approaches then turn weather prediction into afiltering problem in high dimensions and thus they offer interesting

possibilities or our fields, in combination with physics and numericalmathematics.

So the first attempts with a purely inductive approach to under-stand weather and climate ailed. Tey were replaced by a purelydeductive and deterministic approach. However, also that approachreached its limit. Uncertainty is unavoidable and stochastic methodshave re-entered the scene. What the story will look like or social

sciences or neuroscience in the uture, how progress will be made, andwhat the contributions o our fields will be, I dont know, but I wouldlike to live long enough to see at least part o it. I encourage you toactively participate in these research efforts and I hope that IMS willcontribute through its journals, meetings and the contacts it makes

possible.

Association for Women in Mathematics offers Travel Grants

Te next deadline or the AWM/NSF ravel Grant is October , . Te Association or Women in Mathematics travel program supportsthese opportunities or women:. ravel grants in mathematics, or women mathematicians attending mathematics conerences.. ravel grants or women mathematicians attending a mathematics education research conerence.. ravel grants or women mathematics education researchers attending a mathematics conerence.For complete eligibility and application details please visit http://www.awm-math.org/travelgrants.html. All applications are now submitted viaan online system. Te ravel Grant Program is sponsored by the National Science Foundation Division o Mathematical Sciences (DMS).

Please consider taking advantage o this opportunity yoursel, i you are eligible, and encourage your eligible colleagues to apply.Association or Women in Mathematics, Waples Mill Rd., Suite , Fairax, VA

Travel grants for women mathematicians
http://www.awm-math.org/travelgrants.htmlhttp://www.awm-math.org/travelgrants.htmlhttp://www.awm-math.org/travelgrants.html


14/36


ANNUAL REVIEWSCelebrating 80 Years of Publishing Excellence

New in 2014 From Annual Reviews:

Annual Review of Statistics and Its ApplicationVolume 1 January 2014 http://statistics.annualreviews.org

Access available online, via your mobile device, or in print.

Editor: Stephen E. Fienberg, Carnegie Mellon UniversityAssociate Editors: Nancy Reid, University of Toronto

Stephen M. Stigler, University of Chicago

TheAnnual Review of Statistics and Its Applicationaims to inform statisticians, quantitative

methodologists, and users of statistics about major methodological advances and the

computational tools that allow for their implementation. It will include developments in the field

of statistics, including theoretical statistical underpinnings of new methodology, as well as

developments in specific application domains such as biostatistics and bioinformatics, economics

machine learning, psychology, sociology, and aspects of the physical sciences.

Complimentary online access to the first volume will be available

for the first year.

Access this and all other Annual Reviews journals via your institution

at www.annualreviews.org.

Also presented online:Annual Reviews Conversations, enlightening interviews with todays

foremost scholars. www.annualreviews.org/page/audio

ANNUAL REVIEWS |Guiding Scientists to Essential Research Since 1932.

Tel: 800.523.8635 (US/CAN)| Tel: 650.493.4400 (WORLDWIDE)| Fax: 650.424.0910 (WORLDWIDE)| Email: [email protected]
http://statistics.annualreviews.org/http://www.annualreviews.org/http://www.annualreviews.org/http://www.annualreviews.org/page/audiomailto:[email protected]:[email protected]://www.annualreviews.org/page/audiohttp://www.annualreviews.org/http://statistics.annualreviews.org/


15/36


Statistics2013 workshop webcast

Earlier this year, we collected input rom IMS members or a subcommittee toexamine the current structure o support or the statistical sciences within the

US National Science Foundation (NSF) and to provide recommendations orNSF to consider.An update o the committee work is now available rom the website or the

NSF Advisory Committee or the Directorate o Mathematical and PhysicalSciences. Te report is rom the StatsNSF Subcommittee, by Iain Johnstoneand Fred Roberts, at http://www.nsf.gov/attachments/128225/public/MPSAC-July18-2013-StatSNSF-Johnstone-Roberts_Final.pdf)

Tese are draf recommendations at this point, pending eedback rom thevarious advisory committees at NSF.

NSF support for the statistical sciences: latest

Future of the Statistical Sciences Workshop

http://www.statistics2013.org/introduction-to-the-future-of-statistical-sciences-workshop/

One o the primary goals o the International Year o Statistics(Statistics) is to promote creativity and development in thesciences o probability and statistics. o that end, the six societiesthat ounded Statisticsthe American Statistical Association,International Biometric Society, Institute o Mathematical Statistics, International Statistical Institute (and the Bernoulli Society), and theRoyal Statistical Societyare organizing a cross-discipline workshop that will examine the uture o the statistical sciences, the capstone evento the International Year o Statistics.

Organizers are hard at work on preparations or the Future o the Statistical Sciences Workshop, that will be held November and inLondon, United Kingdom. Te workshop will bring together statisticians rom around the world, scientists who collaborate with statisticians,science writers and representatives rom unding agencies.

Te event will showcase the breadth and importance o statistics and highlight the extraordinary opportunities or statistical research in thecoming decade. More specifically, the workshop will tell an important story to a number o audiences: to scientists in data rich fields:statisticians are ready, willing and able to engage in deep, long-term collaborations; to unding agencies:statistics is the data science and can (and does) provide tools to enable progress across a huge range o human endeavors; to statisticians: dont let the opportunities or high-impact research pass you by;and to uture statisticians: engage with a discipline with a huge uture.

Participation is by invitation only and will involve approximately people. In addition to statisticians, the workshopwill involve representatives rom other areas o science and practice that involve problems o a statistical nature.Te workshop will be organized around presentations and panels. A list o the invited speakers is at http://www.statistics2013.org/workshop-invited-participants/

However,all will be able to view the workshop live as well as recordings o the proceedings via the web thanks to WileyPublishers.Afer the workshop, with input rom participants, the organizing committee will publish a summary document

that will be widely distributed to the larger scientific community.
http://www.nsf.gov/attachments/128225/public/MPSAC-July18-2013-StatSNSF-Johnstone-Roberts_Final.pdfhttp://www.nsf.gov/attachments/128225/public/MPSAC-July18-2013-StatSNSF-Johnstone-Roberts_Final.pdfhttp://www.nsf.gov/attachments/128225/public/MPSAC-July18-2013-StatSNSF-Johnstone-Roberts_Final.pdfhttp://www.nsf.gov/attachments/128225/public/MPSAC-July18-2013-StatSNSF-Johnstone-Roberts_Final.pdfhttp://www.statistics2013.org/introduction-to-the-future-of-statistical-sciences-workshop/http://www/http://www.nsf.gov/attachments/128225/public/MPSAC-July18-2013-StatSNSF-Johnstone-Roberts_Final.pdfhttp://www.nsf.gov/attachments/128225/public/MPSAC-July18-2013-StatSNSF-Johnstone-Roberts_Final.pdfhttp://www.nsf.gov/attachments/128225/public/MPSAC-July18-2013-StatSNSF-Johnstone-Roberts_Final.pdfhttp://www/http://www.statistics2013.org/introduction-to-the-future-of-statistical-sciences-workshop/


16/36


JSM2013 in pictures

EricSa

mpson/ASA


17/36



18/36


Forthcoming

A Basic Course in

Measure and ProbabilityTheory for Applications

Ross Leadbetter, Stamatis Cambanis,

andVladas Pipiras

$115.00: Hardback: 978-1-107-02040-5: 370 pp.

$50.00: Paperback: 978-1-107-65252-1

Bayesian Filtering and Smoothing

Simo Srkk

Inst itut e o f Mathematical Statisti cs Textbooks

$99.00: Hardback: 978-1-107-03065-7: 256 pp.

$36.99: Paperback: 978-1-107-61928-9

Stochastic Networks

Frank Kelly andElena Yudovina

Inst itut e o f Mathematical Statisti cs Textbooks

$99.00: Hardback: 978-1-107-03577-5: 160 pp.

$36.99: Paperback: 978-1-107-69170-4

Now in Paperback!

Large-Scale InferenceEmpirical Bayes Methods for

Estimation, Testing, and Prediction

Bradley Efron

Inst itut e o f Mathematical Statisti cs

Monographs

$39.99: Paperback: 978-1-107-61967-8: 280 pp.

Actuarial Mathematics for

Life Contingent RisksSecond Edition

David C. M. Dickson, Mary R. Hardy,

andHoward R. Waters

Inte rnational Series on Actuarial Science

$89.00: Hardback: 978-1-107-04407-4: 632 pp.

Introduction to Mathematical

Portfolio Theory

Mark S. Joshi andJane M. Paterson

Inte rnational Series on Actuarial Science

$70.00: Hardback: 978-1-107-04231-5: 325 pp.

Analysis of Multivariate and

High-Dimensional Data

Inge Koch

Cambridge Series in Statistical and

Probabili stic Mathematic s

$89.00: Hardback: 978-0-521-88793-9: 400 pp.

Statistical Principles

for the Design of ExperimentsApplications to Real Experiments

R. Mead, S. G. Gilmour, and

A. Mead

Cambridge Series in Statistical and

Probabili stic Mathematic s

$95.00: Hardback: 978-0-521-86214-1: 586 pp.

Advances in Statistical

BioinformaticsModels and Integrative Inference for

High-Throughput Data

Edited byKim-Anh Do,

Zhaohui Steve Qin, and

Marina Vannucci

$120.00: Hardback: 978-1-107-02752-7: 511 pp.

Brownian Models of

Performance and Control

J. Michael Harrison

$60.00: Hardback: 978-1-107-01839-6: 200 pp.

Handbook on Systemic Risk

Edited byJean-Pierre Fouque and

Joseph A. Langsam

$160.00: Hardback: 978-1-107-02343-7: 992 pp.

Measuring and ReasoningNumerical Inference in the Sciences

Fred L. Bookstein

$65.00: Hardback: 978-1-107-02415-1: 480 pp.

Whos Bigger?Where Historical Figures Really Rank

Steven Skiena andCharles Ward

$27.99: Hardback: 978-1-107-04137-0: 408 pp.

New Titles from Cambridge University Press!

Prices subject to change.

www.cambridge.org/statistics

800.872.7423

@cambUP_maths
http://www.cambridge.org/statisticshttp://www.cambridge.org/statistics


19/36


Annals of ProbabilityVolume 41, issue 4: July 2013

TheAnnals of Probability publishes research papers in modern probability theory, its relations to other areas of mathematics, and

its applications in the physical and biological sciences. Emphasis is on importance, interest, and originalityformal novelty and

correctness are not sufficient for publication. The Annalswill also publish authoritative review papers and surveys of areas in vigorous

development. The editor is Krzysztof Burdzy (from 20122014).

Access papers at http://projecteuclid.org/aop

Busemann functions and the speed of a second class particle in the rarefaction fan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ERIC CATOR AND LEANDRO P. R. PIMENTEL

Optimal transport from Lebesgue to Poisson . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . MARTI N H UE SMAN N AN D K ARL THE ODOR STU RM

Rate of convergence and Edgeworth-type expansion in the entropic central limit theorem . . . . . . . . . . . . . . . . . SERGE Y G. BOBKOV, GENN ADIY P. CHIST YAKOV AND FRIEDR ICH GTZE

Determinantal point processes with J -Hermitian correlation kernels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . EUGENE LYTVYNOV

Regularity of laws and ergodicity of hypoelliptic SDEs driven by rough paths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . MARTIN HAIRER AND NATESH S. PILLAI

Nonintersecting random walks in the neighborhood of a symmetric tacnode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . MARK ADLER, PATRIK L. FERRARI AND PIERRE VAN MOERBEKE

Extreme gaps between eigenvalues of random matrices. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G RAR D B EN AROU S AN D PAUL BOU RGAD E

Random walks at random times: Convergence to iterated Lvy motion, fractional stablemotions, and other self-similar processes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . PAUL JUNG AND GREG MARKOWSKY

Poisson approximations on the free Wigner chaos. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IVAN NOU RDI N AN D G IOVAN NI PECCAT I

Random fields and the geometry of Wiener space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . JONATHAN E. TAYLOR AND SREEKAR VADLAMANI

Large deviations for solutions to stochastic recurrence equations under Kestens condition . . . . . . . . . . . . . . . . . . . . . . D. BURACZEWSKI, E. DAMEK, T. MIKOSCH AND J. ZIENKIEWICZ

Reflecting random walk in fractal domains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . K RZYSZTOF BU RD ZY AN D ZHE NQI NG CH EN

Central limit theorem for a Stratonovich integral with Malliavin calculus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .DANIEL HARNETT AND DAVID NUALART

Uniqueness and universality of the Brownian map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . JEAN FRA NOIS LE GALL

The Burgers equation with Poisson random forcing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . YURI BAKHTIN

Inhomogeneous bond percolation on square, triangular and hexagonal lattices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G EO FF RE Y R. GR IM ME TT AN D I OA N M AN OL ES CU

Integrability and tail estimates for Gaussian rough differential equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . THOMA S CASS, CHRISTIA N L ITTERER A ND TERRY LYONS

Recent papers

IMS Textbooks

Cambridge University Press, in conjunction with the Institute oMathematical Statistics, publishesIMS extbooks. Tese are compactbooks that give introductory accounts o topics rom statistics or

probability o current interest, suitable or advanced courses at

masters level, or doctoral students andor individual study. Tey are shorter thana ully developed textbook, and typicallycontain exercises. Te Series Editors areSusan Holmes, Ben Hambly, Alan Agresti,D. R. Cox, and Xiao-Li Meng. See http://www.cambridge.org/ar/academic/subjects/

statistics-probability/statistics-and-

probability-general-interest/series/institute-

mathematical-statistics-textbooks

Forthcoming Title in IMS Textbooks: PRE-ORDER NOW

Stochastic Networks by Frank Kelly, University o Cambridge, andElena Yudovina, University o Michigan, Ann Arbor.

Tis textbook is derived rom a graduate-level course taught atthe University o Cambridge. Sel-contained appendices contain dis-

cussion o relevantmathematical tech-niques, e.g. Lagrangemultipliers.

Priced at hardback, .

paperback, but IMSmembers will receive discount. Orderdetails to ollow.
http://projecteuclid.org/aophttp://www.cambridge.org/ar/academic/subjects/statistics-probability/statistics-and-probability-general-interest/series/institute-mathematical-statistics-textbookshttp://www.cambridge.org/ar/academic/subjects/statistics-probability/statistics-and-probability-general-interest/series/institute-mathematical-statistics-textbookshttp://www.cambridge.org/ar/academic/subjects/statistics-probability/statistics-and-probability-general-interest/series/institute-mathematical-statistics-textbookshttp://www.cambridge.org/ar/academic/subjects/statistics-probability/statistics-and-probability-general-interest/series/institute-mathematical-statistics-textbookshttp://www.cambridge.org/ar/academic/subjects/statistics-probability/statistics-and-probability-general-interest/series/institute-mathematical-statistics-textbookshttp://projecteuclid.org/aophttp://www.cambridge.org/ar/academic/subjects/statistics-probability/statistics-and-probability-general-interest/series/institute-mathematical-statistics-textbookshttp://www.cambridge.org/ar/academic/subjects/statistics-probability/statistics-and-probability-general-interest/series/institute-mathematical-statistics-textbookshttp://www.cambridge.org/ar/academic/subjects/statistics-probability/statistics-and-probability-general-interest/series/institute-mathematical-statistics-textbookshttp://www.cambridge.org/ar/academic/subjects/statistics-probability/statistics-and-probability-general-interest/series/institute-mathematical-statistics-textbookshttp://www.cambridge.org/ar/academic/subjects/statistics-probability/statistics-and-probability-general-interest/series/institute-mathematical-statistics-textbooks


20/36


Annals of Applied ProbabilityVolume 23, issue 4:August 2013

TheAnnals of Applied Probability aims to publish research of the highest quality reflecting the varied facets of contemporary Applied

Probability. Primary emphasis is placed on importance and originality. The editor (2013 to 2015) is Timo Sepplinen.

Access papers at http://projecteuclid.org/aoap

Positive recurrence of piecewise OrnsteinUhlenbeck processes and common quadratic

Lyapunov functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. B. DIEKER AND XUEFENG GAO

Fluid limits to analyze long-term flow rates of a stochastic network with ingress

discarding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . JOHN MUSACCHIO AND JEAN WALRAND

On the closure in the Emery topology of semimartingale wealth-process sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CONSTANTINOS KARDARAS

Upper bound on the rate of adaptation in an asexual population. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . M ICHAEL KELLY

Randomized urn models revisited using stochastic approximation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . SOP HIE LARU ELLE AN D GILLES PAG S

Dynamics of cancer recurrence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . JASMINE FOO AND KEVIN LEDER

The cut-tree of large GaltonWatson trees and the Brownian CRT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . J EAN BE RTOI N AN D G R GORY MIE RM ON T

A trajectorial interpretation of Doobs martingale inequalities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. ACCIAIO, M. BEIGLBCK, F. PENKNER, W. SCHACHERMAYER AND J. TEMME

Convolution equivalent Lvy processes and first passage times . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . PHILIP S. GRIFFIN

Quantitative version of the KipnisVaradhan theorem and Monte Carlo approximation of

homogenized coefficients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ANTOINE GLORIA AND JEANCHRISTOPHE MOURRAT

Mimicking an It process by a solution of a stochastic differential equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . GERA RD BRUN ICK AND STEV EN SHRE VE

Kalikow-type decomposition for multicolor infinite range particle systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. GALVES, N. L. GARCIA, E. LCHERBACH AND E. ORLANDI

Weak approximations forWiener functionals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .DORIVAL LEO AND ALBERTO OHASHI

Analysis of casino shelf shuffling machines. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . P ER SI DI ACON IS, JASON FU LMAN AN D S USAN HOLME S

Filtering and smoothing methods are used to produce an accurate estimate o the state o atime-varying system based on multiple observational inputs (data). Interest in these methodshas exploded in recent years, with numerous applications emerging in fields such as navigation

aerospace engineering, telecommunications and medicine. Tis compact, inormal introduc-tion or graduate students and advanced undergraduates presents the current state-o-the-artfiltering and smoothing methods in a unified Bayesian ramework. Readers learn what non-linear Kalman filters and particle filters are, how they are related, and their relative advantagesand disadvantages. Tey also discover how state-o-the-art Bayesian parameter estimationmethods can be combined with state-o-the-art filtering and smoothing algorithms. Te book

practical and algorithmic approach assumes only modest mathematical prerequisites. Exampleinclude MALAB computations, and the numerous end-o-chapter exercises include compu-tational assignments. MALAB/GNU Octave source code is available or download at www.cambridge.org/sarkka, promoting hands-on work with the methods.

PRE-ORDER NOW

Bayesian Filtering and SmoothingSimo Srkk, Aalto University, FinlandPriced at hardback, . paperback;

IMS members will receive discount offthese prices. Ordering details to ollow.

Recent papers

Forthcoming Title in IMS Textbooksseries
http://projecteuclid.org/aoaphttp://www.cambridge.org/sarkkahttp://www.cambridge.org/sarkkahttp://www.cambridge.org/sarkkahttp://www.cambridge.org/sarkkahttp://www.cambridge.org/sarkkahttp://projecteuclid.org/aoap


21/36


Terences Stuff: Statistics and The War

Terry Speed reflects

on the significant

advances made in

statistics during, and

because of, the Second

World War.

George Boxs passing (see his obitu-ary in the previous issue) broughtto my mind something that haslong interested me: the enormous boost that

World War (hereafer the War) gave toStatistics in the United Kingdom, the USA,and elsewhere. Tis is one o the many par-adoxes o war. We all agree that it is terribleand to be avoided to the greatest extent pos-sible, yet it is hard not to concede that warscan bring scientific, technological, industrial,cultural, political, even economic benefits,over and above the purely strategic goals or

which wars are ought. Not only was thereextremely rapid development o some areaso statistics, especially industrial statistics, but

also a large proportion o the leaders in oursubject in the years ollowing the War metit or the first time during the War. It seemsto me likely that a large proportion o these

people, mostly men, would not have becomestatisticians but or the War. Indeed, Boxtitled his memoirsAn Accidental Statistician,the accident being that he was assigned therole o statistician when he was a soldier inthe War working on chemical deence.

I can remember being enormously

impressed that such an elegant, poweruland eminently practical theory as AbrahamWalds sequential analysis was developed inwartime. Tis work ascinated me or severalyears, particularly the extensions using mar-tingales o Walds identities. W. Allen Wallis

was the leader o the U.S. east coast StatisticalResearch Group (SRG) within which Wald

worked. In a JASApaper which I oundinspiring, he gave a history o this groupand a first-hand account o the origins o

Walds work. About the our books resultingrom the SRGs work, Wallis wrote, Allhave proved influentialWalds [Sequential

Analysis, ] ar more than the others, ocourse.

Many other wartime discoveries on bothsides o the Atlantic have proved influential.In the UK Ministry o Supplys ScientificResearch group SR led by George Barnard,one project concerned the operation oa weapon that contained components(actors). Wanted was an experiment with arealistic number o combinations o levels othese actors that was optimum in a suitablesense or estimating the coefficients in amain-effects only model. Te result was the

path-breaking work o R.L. Plackett and J.P.Burman utilizing the mathematical researcho R.E.A.C. Paley on Hadamard matrices.oday their work is seen as the beginning oscreening designs, which have become widelyused in industry, and are now vigorously

promoted as multivariable testing (MV).

Interestingly, writing about this work morethan years aferwards, Plackett noted, Anexperiment was carried out on the lines pro-

posed, and the interpretation o the data wasound to be somewhat difficult. Presumablythe interactions between the actors werenot all negligible!

Alan uring was not a statistician eitherbeore or afer the War, but during it madeseveral important contributions to our sub-

ject, in particular to the use o Bayes actors,

log Bayes actors and sequential analysis.Tese have already become part o Bayesianolklore, and have recently been summarizedalong with much else in the book Te TeoryTat Would Not Dieby Sharon BertschMcGrayne. urings cryptanalytic methodknown asBanburismus involved a Bayesianorm o sequential analysis pre-dating Walds,but details were not made public untildecades afer the War. It is not easy to getan accurate picture o who knew what in

the period , but much relevantmaterial is now available, particularly inbooks about uring, articles by I. J. Good,and on the websitealanturing.net.

Why should we care about these events years ago? Reflecting on them, two keyBritish participants, Barnard and Plackett

wrote in Peace finally returned, andthe statistical scene in the United Kingdomhad been completely transormed Noother method would have produced thesechanges in only six yearsTere was a largeincrease in

ims bulletin

Documents