Obituray

Vidmantas Kastytis Bentkus, the head of the Mathematical Statistics Department in the Institute of Mathematics and Informatics, Vilnius, died on  June 3, 2010 from a heart attack. 

Vidmantas Kastytis Bentkus was born in Šilute (Lithuania) in 1949. In 1967-1970 he was a student of the Mathematical department at the Vilnius University; in 1970-1976 a student of the Mathematical- Mechanical Department at Moscow State Lomonosov University. In 1977 he received a doctor's degree from the Moscow University. In 1986 received the Habilitated Doctor of Sciences degree from Vilnius University. He was working at Mathematical Statistic Department of the Institute of Mathematics and Informatics as a principal research associate.

Vidmantas’ research interests and achievements were impressive. He started his scientific carrier in the area of functional analysis and infinite dimensional partial differential equations and latter moved to the area of limit theorems of mathematical statistics and probability theory. The last 10-15 years he was working on finding an optimal and most precise form of inequalities for sums of independent random variables. He is an author or a coauthor of more than 90 scientific papers published in most prestigious science journals worldwide.

In 1987 and 1999  he was awarded the Lithuanian National Prize of Science. In 1990-1991 he was a fellow of the Alexander von Humboldt Stiftung in Germany.

 

Obituary for Lester Eli Dubins, 1920-2010, written by David Gilat, Ted Hill, Bill Sudderth

Lester E. Dubins, a distinguished probabilist and Professor Emeritus of Mathematics and Statistics at UC Berkeley, passed away peacefully in his home on February 11, 2010. He is survived by his son Benjamin, his sister Blanche, his nephews Aaron, Mathew, Michael and Ross and his nieces Marylyn and Sarah.

Lester grew up in New York City. His college education was interrupted by World War II in which he served as an Air Force officer stationed at a radar installation in Iceland. After the war, in the late 1940s, Lester continued working (as a civilian) for the radar research & development branch of the US Air Force. In 1951 he resumed his studies as a graduate student of mathematics at the University of Chicago where he obtained his Ph.D. degree in 1955 submitting a thesis on Generalized Random Variables under the guidance of Irving E. Segal. In those days Chicago's mathematics department was fortunate to have an excellent cohort of graduate students, among them Paul Cohen, Don Ornstein and Jack Feldman.

After a year at the Princeton Institute for Advanced Study and a few years at Carnegie Institute of Technology in Pittsburgh, Lester joined the Mathematics and Statistics departments at UC Berkeley in 1962, where he remained for the rest of his life. Lester was forced to retire at 70, shortly before the abolition of mandatory retirement age in public institutions in the USA went into effect. He challenged the university, won his case in court and was reappointed. He then continued teaching until retiring in 2004 by his own volition.

While still in Chicago, Lester met Jimmie Savage (then on the mathematics faculty there) and surprised him by showing that bold play is not uniquely optimal in classical Red & Black (roulette). Jimmie was impressed and invited Lester to join him in trying to better understand the probabilistic structure of gambling situations. This encounter developed into a collaboration generating several key papers and culminating (in 1965) in the ground-breaking monograph How to Gamble if You Must (Inequalities for Stochastic Processes) which presented a coherent mathematical theory of gambling processes and optimal behavior in gambling situations, pointing out their relevance to traditional approaches to probability. In consultation with Bruno de Finetti and under his influence, Dubins and Savage presented their theory in the finitely additive framework in order to bypass measurability technicalities inherent in maximizing an uncountable set of functions in searching for optimal strategies. Lester paid tribute to Jimmie Savage - to his remarkable intellect and scholarship, and to their mutual friendship - in a beautifully phrased preface to the Dover edition of the book which appeared in 1976, five years after Jimmie's untimely death at the age of 54.

Lester continued to contribute to finitely additive probability throughout his entire career. As late as 1999 he published a paper entitled Paths of Finitely Additive Brownian Motion need not be Bizarre, demonstrating that those paths can be, for example, continuous piecewise linear or piecewise polynomial functions, etc. Not unrelated to his fondness of finite additivity, was Lester's great admiration for Bruno de Finetti and his firm belief in the subjective nature of probability according to which probabilities should not be thought to be inherent in objects, but instead in one’s ideas and expectations about those objects.

Although primarily a probabilist, Lester had deep interest in mathematics in general. Thus scattered among his close to 100 publications (the latest of which appeared in 2009) are papers on a variety of mathematical themes, such as curves of minimal length (under some constraints), Tarski's Circle Squaring problem, convex analysis and geometry.

Lester was an unconventional teacher. He seemed unable to bore himself and his audience by a systematic recitation of a textbook. Even the most elementary material would receive a fresh and original treatment under his hands. In teaching as in research, Lester would never be satisfied with just having a logically correct proof of a statement, but would always strive for a natural, simplest possible argument. He was at his best in advanced graduate seminars and informal discussions. In front of small groups of colleagues, international visitors and graduate students he challenged his audience, exposed and explored his own thoughts and still-evolving ideas, and inspired them with unexpected basic questions and insights. He would often open a session asking if anyone had anything interesting to tell the class, and his seminars were fascinating Socratic dialogues that continued informally between regular sessions.

As much as he lived for mathematics, Lester also cared deeply about people, especially his students and friends, who came from all walks of life. He was always a champion of the underdog, whether they were California farmworkers or oppressed people anywhere, and was an outspoken advocate of civil liberties. He loved words and stories and humor, and led a vigorous life, traveling and hiking and mountain-biking into his eighties, roaring with laughter or immersed in explorations of mathematical subtleties as he went. He was a person of the highest integrity in everything he did, scientifically and socially, and did not bend to political or peer pressures. You did not ask Lester how he liked the restaurant you were in or how good a logical argument was, unless you were prepared for the polite but unadulterated truth. For those of us who were fortunate enough to know him well, hardly a day passes when we do not ask ourselves “What would Dubins have done in this situation?” or “What would Lester have said?” Lester Dubins was the quintessential maverick mathematician, and above all, an absolute gentleman and scholar.

Obituray for Erich Lehmann by his family and Peter Bickel:

Professor Erich Lehmann, a major figure in our field of statistics, died on September 12, 2009, at the age of 91. Lehmann, who taught in Berkeley’s Statistics Department, touched the lives of many people in statistics and beyond. He was a leading figure in the second generation of statisticians, following the establishment of the modern field by Neyman, Fisher and Wald before and shortly after the Second World War. As is usual after a period of explosive innovation, confusion reigned. It was Lehmann’s great talent to clear the fog and build a coherent theoretical structure. This was reflected in his great books, Testing Statistical Hypotheses (1959) and Theory of Point Estimation (1983), which were the centerpieces of graduate statistical education for most of the last half of the century, and have been translated into many languages. The structures he built incorporated much of his own research into decision theoretic questions, the theory of unbiased estimation, rank based methods including their asymptotic theory and many other areas of  statistics, including concepts of dependence, multiple comparison procedures, the history of statistics and much else.

Lehmann achieved all the major honors awarded in the field and beyond: the prestigious Wald and Fisher lectureships , the presidency of the Institute of Mathematical Statistics and the editorship of its main journal, The Annals of Mathemtical Statistics. He was granted a remarkable three Guggenheim Fellowships in 1955, 1966 and 1980 and was elected to the American Academy of Arts and Sciences in 1975 and to the National Academy of Sciences in 1978. The Universities of Leiden and Chicago awarded him honorary doctorates. At Berkeley he held Miller Professorships twice and served reluctantly but very effectively as Department Chair.

Born in Strasbourg, France in 1917,  Lehmann was raised in Frankfurt am Main, where his family had deep roots. . Fleeing the Nazis with his family in 1933, he graduated from high school in Switzerland and attended college in Cambridge, England. He enrolled in Berkeley as a graduate student in 1940 and never left, save for stints in the Air Force during World War II, when he was stationed in Guam, and leaves at Columbia, Princeton and Stanford. Obtaining his Ph.D. in 1946, he embarked on a teaching career that included the supervision of more than 40 doctoral students of his own, several of whom became leaders in the next generation of statisticians. This achievement was due not only to his great scientific stature but also to remarkable personal qualities. He was kind and generous of spirit, had an unusual sensitivity to the feelings of others and a great astuteness about the world, what could be achieved, and how to do it. As a consequence his impact on his students and colleagues went well beyond the scientific. They honored him with a Festschrift (1983) for his 65th birthday, a series of three Lehmann Symposia (2002,2004,2007), and a forthcoming volume of selected works.

In addition to his masterpieces, Lehmann published three important, less advanced  texts: Basic Concepts of Statistics (with his longtime collaborator and friend J. L. Hodges, Jr.) Nonparametrics: Statistics Based on Ranks, and Elements of Large Sample Theory. After a second edition of his classic Testing Statistical Hypotheses in 1986, he recruited young collaborators for further editions of his major texts - George Casella for Estimation in 1998 and Joe Romano for a third edition of Testing in 2005. These were major revisions that brought the books back to the frontiers of research. In his last decade he turned his energies to the history of the field in whose development he played such an important part, publishing his professional autobiography, Reminiscences of a Statistician: The Company I Kept, and an account of the productive rivalry between Fisher and Neyman tentatively titled Fisher, Neyman, and the Creation of Classical Statistics, completed shortly before his death, to be published by Springer. He also enjoyed a lifelong passion for literature and in retirement translated stories by favorite authors such as Adalbert Stifter and Wilhelm Raabe, seeking to give them a wider audience than they previously enjoyed. At the time of his death he was working with Fritz Scholz, a former student, on a new edition of his Nonparametrics to be used in conjunction with the popular “R” statistical language.

He is survived by his wife, Juliet Popper Shaffer, herself an accomplished member of our field and a collaborator during the later part of his life and a loving blended family that includes his three children, Stephen, Barbara and Sophia; three step-children, Ron, Len and Tanya; eight grandchildren, and two great-grandchildren with a third on the way.

Written in memory of Miguel by Evarist Giné, David Mason, Francisco Samaniego and Anton Schick: Miguel Angel Arcones, 46, Professor of Mathematical Sciences at Binghamton University, died on December 30, 2009, after a long battle with cancer. A native of Segovia, Spain, Arcones began his graduate studies in the US in 1987 and earned his PhD in 1991 from the CUNY, working with Evarist Giné. After several postdoctoral appointments, Miguel joined the faculty of Binghamton University in 1998. In spite of periods of illness and hospitalization, he carried out his duties as a teacher and scholar with determination and grace, and he distinguished himself in both. He was well known for the depth and breadth of his research and was an elected member of the ISI and a Fellow of the IMS. He was highly appreciated by his students for his passion for the discipline and for his devotion to their learning. While serving as a Wylie Instructor at the University of Utah early in his career, he received an outstanding instructor award. Later, he enthusiastically served as the faculty mentor of the Mathematics Club at Binghamton University. He was the driving force behind the establishment of a thriving actuarial degree program at BU.

Miguel's research interests and achievements were impressive, both broad in range and marked by conceptual and technical sophistication. His work in Probability Theory includes notable contributions to the theory of U-processes, limit theory under dependence conditions and large deviations theory. Particularly influential were his series of papers published in 1993-4 on the asymptotic theory of U processes, his celebrated 1994 paper on the limit distribution of functionals of stationary sequences of Gaussian vectors and his definitive treatment, in 2003-4, of the large deviation principle for stochastic processes. His work in Mathematical Statistics was equally broad, including contributions to bootstrap theory, the asymptotics of U-Statistics and M-estimators, the Bahadur-Kiefer representation in a wide array of statistical contexts and order-restricted inference in Reliability. His contributions to the theory of U-statistics, beginning with his 1992 Annals of Statistics paper on bootstrapping U-Statistics, and including subsequent papers on large deviations, inequalities, the LLN, the CLT and the LIL for U-Statistics, were especially notable. He also wrote a diverse stream of research papers M-estimator asymptotics. His contributions to reliability were focused on constrained estimation of a survival function. His work on estimation under a uniform stochastic ordering constraint appeared in the AoS in 2000, and his work on estimation under a "stochastic precedence" constraint was published in JASA in 2002.

For seven years, Miguel served as an Associate Editor for JASA's Theory and Methods Section. The Editors he served for will readily attest to the outstanding service he rendered in this capacity. His AE reports were, without exception, timely and thorough, always containing insightful commentary on the reviews he commissioned and always including an independent analysis that added value. Additional evidence of the generosity of his professional service is the fact that he published over 130 articles in Mathematical Reviews. He was the founding Editor of the International Journal of Statistics and Management Systems and nurtured the fledgling journal for its first four years. His high standards and breadth of expertise are readily apparent from the volumes of the IJSMS that he saw through to publication. He passed away just weeks after having participated in the search for his replacement.

Miguel Arcones was a gentle giant in the statistical sciences, shy and sometimes unnoticeable in social settings, yet animated, stimulating and highly creative in his professional interactions. To his research visitors, he was a warm and gracious host. His colleagues and collaborators will not soon forget his generous nature, the excitement that he felt and shared when thinking about research questions, his sense of humor and hearty laugh and the friendship that he offered so fully and sincerely. His premature passing is a great loss to our discipline. While his work will continue to influence students and researchers in Probability and Statistics for years to come, the role he played in the lives of his friends, students, colleagues and collaborators is irreplaceable. He is a man who left a strong positive imprint, both personally and professionally. May he rest in peace.
© Bernoulli Society