Peter Bickel writes:
Of the nine National Medals of Science awarded by President Obama this year, three were in what I view, ecumenically, as the mathematical, information and data sciences. The medals are awarded to David Blackwell, posthumously, to Alexandre Chorin and to Thomas Kailath. Of these, David Blackwell’s is close to the hearts of our community. Blackwell’s National Medal of Science is apparently the very first such medal awarded after the recipient’s death was known.
David Blackwell in 1989.
Photo: George M. Bergman, Berkeley/Oberwolfach
Blackwell died July 8, 2010 at the age of 91. At the time of his death he was Professor Emeritus of Statistics and Mathematics at the University of California, Berkeley. He was one of the foremost statisticians and probability theorists in the world—contributing not only to mathematical statistics and probability theory but also to game theory, information theory and mathematical logic. His work made a significant impact on these disciplines, as well as on economics, engineering, medicine and other sciences.
In statistics he is known for the Rao-Blackwell improvement scheme, for helping lay the foundations of dynamic programming, and, with Arrow and Girshick, for applying the backward induction method to prove the fundamental theorem of sequential analysis. In a final major contribution he asked: when can one statistical experiment be more informative than another? He defined this concept and provided a simple, necessary, and sufficient condition for experiments to satisfy this definition. This beautiful piece of mathematics has become one of the pillars of the decision-theoretic approach to mathematical statistics.
In probability, Blackwell is best known for the Blackwell renewal theorem, a key tool in queueing theory. In information theory, he contributed among other things the Blackwell channel. In game theory, he was among the first to deal with games with imperfect information. This led to his interest in mathematical logic to which he added Blackwell games.
Blackwell’s active career, 1941-1988, was recognized with well-deserved honors: election to the National Academy of Sciences [in 1965], the American Academy of Arts and Sciences [in 1968], twelve honorary doctorates and other honors and prizes. His talents were recognized early but, because he was African-American, the beginning of his career was harder than it should have been. To take one example of many, when he was a Fellow at the Institute for Advanced Studies, his thesis advisor, Joe Doob, had to intervene to ensure him privileges at Princeton University that were normally granted to fellows of the Institute.
Blackwell was a marvelous teacher. His lucid expositions made difficult ideas seem simple and clear. His students testified to this in “A Tribute to David Blackwell,” Notices of the AMS 58 (2011), 912–928. With his students, colleagues, family and friends, he was a man of exceptional kindness, wit, charm and playfulness. Philosophically he viewed himself as a Bayesian, but the variety of his contributions showed that, in practice, he was ecumenical. David Blackwell deserved the National Medal of Science in his lifetime. We celebrate the man himself and his work. We regret that he is not here with us to enjoy the honor.
IMS Fellow Thomas Kailath also received a Medal of Science this year. Kailath is an electrical engineer known for his contributions to the information and system sciences. He is the Hitachi America Professorship of Engineering, Emeritus, at Stanford University.