


Terence Speed
Professor, Head of the Bioinformatics Division at the Walter and Eliza Hall Institute of Medical Research, in Melbourne, Australia, and Department of Statistics, University of California, Berkeley
Doctor of Science
Five Hundred Nineteenth Convocation
June 14, 2014
Click for additional photos
Citation:
Presenter: Stephen Stigler
Terry Speed, a statistician and statistical geneticist, is a pioneer in the development and application of statistical methods for the analysis of biomedical and genomic data. His work is considered to exemplify the best of applied statistics in crossdisciplinary research. Initially a pure mathematician specializing in algebra, he turned to the study of probability and mathematical statistics, and later to genomics. His expertise is in developing novel statistical and computational methods to extract the key signals of interest from the inherently large, complex, noisy datasets that arise from emerging genomics technologies.
He and his collaborators have developed innovative statistical methods for addressing key practical issues in microarray data analysis. These methods have become standards that have been implemented in widely used opensource software, thereby transforming the field. His earlier work on Markov random fields and loglinear models was similarly influential, laying the groundwork for the modern analysis of graphical models. He has also made major contributions to the understanding of recombination, the fundamental biological process responsible for shuffling genetic material between chromosomes as it passes from generation to generation. In addition, he has contributed to various aspects of genetic sequence analysis, including transcription factor binding site prediction, as well as the analysis of data from highthroughput sequencing assays, for example, in his recent work with the Cancer Genome Atlas Project.


David Donoho
Anne T. and Robert M. Bass Professor in the Humanities and Sciences, and
Professor of Statistics
Doctor of Science
Five Hundredth Convocation
October 9, 2009
Presentation Statement: David L. Donoho has pioneered the application of
principled methods in
mathematical statistics to address a great modern scientific challenge, the
problem of "sparsity" in high dimensional data sets. A data set of a
million cases is not large if ten dimensions are recorded for each case and
there are a huge number of potential interrelations or interactions among
the characteristics. An unknown very few of these are likely to be
important, and in that sense the structure is sparse. Determining ways to
reveal that structure is a daunting challenge.
Donoho draws upon classical statistics and crafts both elegant theory and
novel algorithms to overcome the dimensional complexity. He devises methods
that include the use of wavelets and what he terms "compressed sensing" to
recover sparse relationships with a fraction of the number of observations
needed by other methods. This methodology is enormously influential in
astronomy, genetics, geophysics, signal processing, financial analysis, and
medical imaging.
At one level this work is only Occam's razor, but the technical complexities
are immense and the clarity and mathematical rigor Donoho brings to this
area of analysis are nothing short of extraordinary.
Citation: David L. Donoho is a mathematical statistician, and also one of
the more
influential applied mathematicians of his generation. Building upon the
discipline of statistics, Donoho has developed effective new approaches to
constructing low dimensional representations for modern highdimensional
data problems. His work provides new insight into some of the most pressing
scientific questions of the present day.


Grace Wahba
I. J. SchoenbergHilldale Professor of Statistics,
Professor of Biostatistics and Medical Informatics,
and Professor of Computer Sciences (by courtesy), University of Wisconsin, School of Medicine and Public Health
Doctor of Science
Four Hundred Nineth Convocation
June 8, 2007
Grace Wahba, the I. J. Schoenberg professor of statistics,
University of WisconsinMadison, represents the very best of the modern synthesis
of applied statistical, mathematical and computational science. Her most influential
work has concerned problems in the estimation of curves and surfaces from
large, highdimensional data sets, such as occur frequently in geophysics.
Wahba has introduced the use of reproducing kernel Hilbert spaces in the
formulation of nonparametric smoothing problems to reveal general patterns
without obscuring local features. Her pioneering methods include the introduction
of Generalized CrossValidation, now a generally adopted approach to making
a principled tradeoff between smoothness and attention to detail.
In recent years she and her students have applied these same statistically
based theories to a diverse group of classification problems known in computer
science as “machine learning.” Diverse areas in applied science
have benefited from Wahba’s research, including satellite imaging, magnetic
resonance imaging, meteorology, climatology and DNA microarrays.
Source: http://chronicle.uchicago.edu/070607/honorarydegrees.shtml


Persi Diaconis
The Mary V. Sunseri Professor and Professor of Mathematics, Stanford University
Doctor of Science
Four Hundred Seventy Third Convocation
June 13, 2003
Diaconis, the Mary V. Sunseri professor and professor
of mathematics at Stanford University, has over the past 20 years had a major
influence on the development of probability theory. He and his collaborators
have created diverse and recondite mathematical tools to analyze games of
chance and their associated accouterments, such as cards, dice, coins and
roulette wheels, and for other statistical investigations.
Though the mathematical theory of probability was born
some 350 years ago when the Chevalier de Mere brought to the attention of
Pierre de Fermat and Blaise Pascal the problem of calculating odds in a dicerolling
game played in certain French casinos, Diaconis' contributions have furthered
the study of probability theory.
He is perhaps best known for his discovery with David
Bayer that "seven shuffles suffice" to "randomize" a deck of cards. The problem
of ascertaining how long a random process must run before reaching equilibrium
recurs in almost every area of science where random processes arise. Accordingly,
Diaconis' work and ideas have had ramifications throughout the sciences.
Before he embarked on a career in mathematics, Diaconis
was a professional magician with an expertise in card tricks. By bringing
the power of abstract mathematical reasoning to bear on the concrete problems
of chance that arise from the milieu of gambling, magic and everyday coincidence,
Diaconis followed the best and oldest tradition of probability.
Source: http://chronicle.uchicago.edu/030612/hondegrees.shtml


David Aldous
Professor of Statistics, University of California, Berkeley
Doctor of Science
Four Hundred Sixty Second Convocation
November 2, 2000
David
Aldous, a professor of statistics at the University of California, Berkeley,
will be awarded a Doctor of Science degree. Aldous’ contributions
place him among the world’s leading practitioners of both mathematical
probability and in the theory of computing.
His 1989 monograph, Probability Approximations via the
Poisson Clumping Heuristic, concisely shows how to solve more than 100 challenging
and wideranging problems in probability theory.
In 1993 he became the first recipient of the Line and
Michel Lo‘ve International Prize in Probability. The prize recognizes
the outstanding contributions of probability researchers under the age of
45.
Reference: http://chronicle.uchicago.edu/001102/honorarydegrees.shtml


Bradley Efron
Professor in the Departments of Statistics and Health Research and Policy,
Stanford University
Doctor of Science
Four Hundred Thirtyninth Convocation
June 9, 1995
Efron, widely regarded as the most original and innovative
statistician of his generation, will receive the Doctor of Science degree.
He is a professor in the department of statistics and the department of health
research and policy at Stanford.
Efron is regarded as the inventor of the bootstrap form of computerintensive
resampling for solving distributional problems connected with statistical
inference, and his article on the topic is among the most influential papers
on statistics in the 20th century. His work on the relationship between scientific
inference and statistical models has set the research agenda for much of current
statistics and has had a significant impact on a wide variety of sciences,
from astronomy and paleontology to demography and medicine.
A faculty member at Stanford since 1966, Efron received his B.S. in 1960
from Caltech and his M.S. in 1962 and his Ph.D. in 1964 from Stanford. He
has received numerous honors and awards, including a MacArthur Fellowship
in 1983 and the American Statistical Association's Wilks Medal in 1990. He
is a member of the American Academy of Arts and Sciences and the National
Academy of Sciences.
Reference: http://chronicle.uchicago.edu/950525/honorary.shtml


Ulf Grenander
L. Herbert Ballou University Professor Emeritus, Division of Applied Mathematics,
Brown University
Doctor of Science
Four Hundred Thirtyfifth Convocation
June 10, 1994
Citation:
Professor Ulf Grenander is an extraordinarily innovative statistician and
applied mathematician. He has a rare ability to integrate elements and structures
from various fields of statistics and mathematics and thereby to create new
paradigms of the most original, profound and esthetic nature. His work in
the 1950s was a cornerstone in the development of modern time series analysis.
In the 1960s he turned his attention to probabilities on algebraic structures,
or probabilities on groups. This work, summarized in his 1963 book, was ahead
of its time. Grenander's work in abstract inference, or inference about random
processes, spans three decades, from his 1950 Ph. D. thesis to his 1981 book
on the topic. The method of sieves, central to nonparametric regression and
density estimation, is Grenander's creation. Starting in the early 1970s,
he turned his attention to the immensely challenging problems connected with
pattern theory, pattern recognition, and image analysis. This was unexplored
intellectual territory at the time, and Grenander's clear conceptual framework
charted the way for the mathematical discussion of pattern over the next two
decades. He continues to set the pace in formulating problems and inventing
solutions in the area of pattern analysis.


Charles M. Stein
Professor Emeritus of Statistics, Stanford University
Doctor of Science
Four Hundred Twentysixth Convocation, Second Session, Celebrating the Centennial
Year
June 12, 1992
Presentation Statement: Professor Charles M. Stein obtained
a B.S. Degree at the University of Chicago in 1940 and taught at the University
from 19511953, but for most of his career he has been in the Department of
Statistics at Stanford University. Through
a series of brilliant discoveries of striking originality, Stein has reoriented
the way theoretical statisticians view their field.
Stein's work is concerned with deep problems in the foundations of statistical
inference. Some of this has to do with the attainability of optimum
results with statistical data assumed to be generated within a limited class
of mechanisms (a parametric family of distributions), and some has to do with
the radically different nature of multidimensional problems in inference. His
most remarkable discovery, made in the 1950s, was of a phenomenon that was
totally at odds with intuition: if you are faced with a problem involving
3 or more equally uncertain simultaneous estimates, you can improve the result
by "shrinking" the separate averages towards a common value, such
as zero. In a prosaic example, if you wish to simultaneously estimate
the average wholesale prices of bushels of apples in Wenatchee, Washington,
of oranges in Orlando, Florida, and of grapes in Modesto, California, and
you will judge their performance by a combined measure of overall accuracy,
you can expect to improve (on average) overall performance by decreasing all
three separate estimates, despite the fact that the problems are in an important
sense unrelated! Of course some assumptions are involved, but they are
surprisingly benign. When Stein first advanced this, it made many good
scholars acutely uncomfortable, and the discovery disturbed a whole community's
complacency. Now, 35 years later, nobody questions the result, and although
there are debates about its practical relevance, the phenomenon (variously
called the "Stein Paradox" or the "Stein Phenomenon")
has had a profound impact on the field.
Stein has made several other discoveries of major importance. In particular,
he discovered new possibilities of attaining asymptotic efficiency with nonparametric
estimates (1956), and he invented a radically new way of obtaining normal
approximations (1972). Both of these works have been increasingly influential
since they were published. Charles Stein is a brilliant dedicated scholar;
personally, he is unusually modest and his prominence in the field is due
entirely to the force of his ideas.
Citation: Whose research in mathematical statistics
and probability theory, has produced a series of brilliant discoveries of
remarkable force and originality. His work on decision theory first startled
and then inspired an entire community of scholars, changing forever the way
statisticians view multidimensional problems of inference, and his novel methods
for deriving approximations in probability theory have been adopted internationally.


Erich L. Lehmann
Professor of Statistics, University of California, Berkeley
Doctor of Science
Four Hundred Twentythird Convocation, Celebrating the Start of the Centennial
Year
October 3, 1991
Presentation Statement: Erich Lehmann is a mathematical
statistician. Mathematics flourishes
in the detailed exploration of constrained structures with widely accepted
rules; statistics flourishes with the flexibility to treat the infinite variety
of problems in the real world. Lehmann’s genius has been his ability
to reconcile these divergent goals and enrich both sides. During his
long and distinguished career, Erich Lehmann has played a crucial role in
the flourishing of the Neyman/Wald paradigm in theoretical statistics. Lehmann
organized the theory, led in developing new concepts, methods, and results,
taught generations of students, and wrote the definitive books on the subject.
The focus of the approach that is so closely associated with Lehmann is the
application of decision theory to statistical problems, the construction of
a calculus of optimal statistical procedures. The success of Lehmann
and his school has come from the balance they have maintained in creating
a system of mathematical structures of sufficient richness to encompass a
large range of practical problems, yet sufficiently concise that a true discipline
could be constructed around them.
Erich Lehmann has emphasized optimality as the unifying principle for theoretical
statisticians. The optimality principle can be viewed in two ways: Given
a criterion, find the best statistical procedure according to the criterion;
given a statistical procedure, find criteria for which it is optimal. Lehmann’s
books and papers have developed both views, thus throwing new light on established
procedures. He is (with Henry Scheffé) responsible for the concept
of completeness and its use to find optimal estimates; he has explored the
concept of unbiased tests of statistical hypotheses and showed that many standard
statistical tests are uniformly most powerful unbiased. He has also
played a key role in the development of mathematical theory for nonparametric
inference, showing how this amorphous subject could be treated in a rigorous
disciplinary framework.
Erich Lehmann’s texts have shaped the paradigm he is associated with. Not
all of the results are his (though many are), but the arrangement, the elegant
seamless presentation, and the coherence of the whole, are his in a way that
is seldom seen. To a large degree, Erich Lehmann created the curriculum
of the world’s graduate programs in mathematical statistics from midcentury
onward.
Citation: Your research on the applications of statistical
theory to the construction of a calculus of optimal statistical procedures
has helped create and organize modern mathematical statistics. You and your
school have succeeded in maintaining a remarkable balance in creating a system
of mathematical structures of sufficient richness to encompass a large range
of practical problems, yet sufficiently concise that a true discipline could
be constructed around them. Your several elegant treatises have guided the
curricula of graduate programs in statistics and thus given shape to this
discipline, and your teaching has inspired a generation of scholars.


Frederick Mosteller
Professor of Mathematical Statistics, Harvard University
Doctor of Science
Three Hundred Fortysixth Convocation, In Celebration of the Opening of the
Harper Memorial Library College Center
October 26, 1973
Citation: Masterful teacher, distinguished scholar,
and imaginative leader, whose exemplary work has enriched and advanced the
teaching and the practice of statistical and quantitative inquiry.


John Wilder Tukey
Professor of Mathematics and Statistics, Princeton University, and Associate
Executive Director of Research, Bell Telephone Laboratories
Doctor of Science
Three Hundred Twentyeighth Convocation
June 13, 1969
Citation: Illustrious scientific generalist, who has
redirected the development of statistics by his farreaching contributions
to theory and methods, and by his creative use of statistics and data analysis
in the pure and applied sciences.


Maurice Stevenson Bartlett
Professor and Head, Department of Statistics, University College, London,
England
Doctor of Science
Three Hundred Fourteenth Convocation
June 10, 1966
Citation: Distinguished innovator in statistical theory
and application, whose wisdom has guided the balanced development of statistical
inference.


Jerzy Neyman
Director of the Statistical Laboratory, and Research Professor in the
Institute for Basic Research in Science, University of California, Berkeley
Doctor of Science
Two Hundred Eightysecond Convocation
June 12, 1959
Citation: Illustrious creator and analyst of statistical
methods, whose theoretical studies and wise applications of statistics have
provided insight into the nature of inductive behavior.


Harold Hotelling
Professor of Statistics, University of North
Carolina
Doctor of Laws
Two Hundred Sixtyseventh Convocation, In Commemoration of the Twentyfifth
Anniversary of the Social Science Research Building of the University of
Chicago
November 11, 1955
Citation: Foremost contemporary contributor of quantitative
methods to the social sciences, who by mathematical analysis has notably advanced
our understanding of fundamental problems in economics and in statistics. 

Sir Ronald Aylmer Fisher
Arthur Balfour Professor of Genetics, University of Cambridge, England
Doctor of Science
Two Hundred Fiftyfirst Convocation
June 13, 1952
Citation: The greatest
figure in the history of statistics
and
one of the greatest in the history
of scientific
method generally, whose creation
of the science of experimental
design and
formulation of the principles of
interpreting evidence have opened
the way and carried
us far toward an understanding
of the logic of inductive reasoning. 
