Persi Diaconis

Persi Diaconis
Steve Castillo Photos · Public domain · source
Name	Persi Diaconis
Birth date	1945-01-06
Birth place	New York City
Nationality	United States
Fields	Mathematics, Statistics, Probability theory, Combinatorics
Alma mater	Kutztown University of Pennsylvania, University of Pennsylvania
Doctoral advisor	William Feller
Known for	Card shuffling, Randomness, Bayesian statistics, Markov chain Monte Carlo

Persi Diaconis is an American mathematician and statistician noted for rigorous studies of card shuffling, randomness, and the mathematics underlying magic (illusion) and coin flipping. He has held research and faculty positions at institutions such as Stanford University and Harvard University, and collaborated widely across mathematics, statistics, and computer science communities. His work bridges theoretical results in probability theory and practical questions in statistical physics, cryptography, and machine learning.

Early life and education

Born in New York City to immigrant parents, Diaconis dropped out of high school and became a professional magician, performing in venues linked to Las Vegas, Broadway, and touring circuits associated with entertainers like Jack Benny and Ed Sullivan. After returning to formal study, he attended Kutztown University of Pennsylvania before transferring to University of Pennsylvania, where he studied under probabilists connected to traditions from Norbert Wiener and Andrey Kolmogorov. He completed a Ph.D. at Harvard University under the supervision of William Feller-influenced faculty and worked at places including Bell Labs during formative years that connected him with researchers from Princeton University and Institute for Advanced Study.

Mathematical career and contributions

Diaconis's career spans appointments at Stanford University, Harvard University, and research affiliations with Microsoft Research and the Santa Fe Institute. He developed influential results in group theory-based analyses of random processes, building on work by Paul Erdős, John von Neumann, and Mark Kac. His collaborations include partnerships with Dave Bayer on mixing times for shuffling, joint work with Ronald Graham-adjacent combinatorialists, and interactions with statisticians in the tradition of Jerzy Neyman and Ronald Fisher. He contributed to rigorous analyses of Markov chains employing techniques related to Fourier analysis on finite groups and methods resonant with the Central Limit Theorem lineage of Andrey Kolmogorov and P. Lévy.

Work on probability, statistics, and randomness

Diaconis's probabilistic research addresses foundational questions about randomness posed by practitioners such as Claude Shannon and theoreticians like Alan Turing. His work on card shuffling established precise mixing times—famously quantifying the number of riffle shuffles needed—drawing on the analytic tools used by Salvador Liaño-style harmonic analysts and probabilists in the lineage of William Feller and Kolmogorov. He investigated random permutations and exchangeability in the tradition of Bruno de Finetti, explored de-biasing mechanisms for coin flipping similar to problems studied by Joseph L. Doob, and analyzed convergence rates of Markov chain Monte Carlo methods influential for researchers at Brookhaven National Laboratory and in Bayesian statistics circles following Thomas Bayes and Harold Jeffreys. Diaconis also examined pseudorandomness and randomness extraction with implications for cryptography as studied by Whitfield Diffie and Ron Rivest.

Publications and books

Diaconis authored and coauthored numerous influential articles and books, often collaborating with figures such as Dave Bayer, David Freedman, Susan Holmes, and Ron Graham. Notable works include monographs and papers that intersect with the literatures of combinatorics and statistical theory alongside textbooks used at Stanford University and Harvard University. His publications appear in journals that publish work by peers like Persi Diaconis coauthors? and have influenced subsequent research by scholars connected to MIT, UC Berkeley, and Columbia University.

Honors and awards

Diaconis has received honors from organizations including the National Academy of Sciences, fellowships similar to those from the MacArthur Fellows Program, and prizes in the company of recipients from institutions such as American Mathematical Society and Institute of Mathematical Statistics. His recognition situates him among members of scholarly communities alongside figures like John Tukey, Bradley Efron, and David Mumford.

Personal life and outreach

Beyond academia, Diaconis has engaged with communities around magic (illusion) and public outreach venues such as TED Conferences-adjacent events and public lectures at museums like the Museum of Mathematics. He has mentored graduate students who later held positions at universities such as Princeton University, Yale University, and University of Chicago, and collaborated with practitioners from technology companies and research groups in the data science ecosystem.

Category:American mathematicians Category:American statisticians