Rankin lectures 2006

The Rankin Lectures 2006

Harnessing Chance

by

Professor Persi Diaconis.

The Rankin Lectures are an annual series of mathematical lectures. This year’s series of five lectures will be given by Professor Persi Diaconis, the Mary V. Sunseri Professor of Statistics and Mathematics at Stanford University. It will take place at the University of Glasgow from the 6^th to the 10th of November 2006. The lectures are intended to be of interested to the general scientific public, with some being particularly accessible to school pupils. Professor Diaconis is an acclaimed scientist, and famed in the public arena for his mathematical and statistical work which he applies to everyday activities, including card shuffling, coin tossing and code breaking. He is well-known as an entertaining and inspiring speaker.

Details of the Lectures.

Here is Professor Diaconis's summary of his lectures. All talks will be in the Mathematics Building.

The twentieth century has been called 'the time of taming chance'. Now, we are harnessing chance to make and break codes, find patterns in DNA, and even do basic counting. These themes will be explored in a sequence of lectures aimed at a general scientific public.

1. The Search for Randomness (particularly suitable for a general audience) 2pm - 3pm Monday 6th November, Room 516

I will explore some of our most primitive images of random phenomenon - flipping a coin, shuffling cards, and throwing a dart at the wall. While randomness can be achieved, usually we are lazy.

2. The Mathematics of Making a Mess 4pm - 5pm Tuesday 7th November, Room 515

I will show it takes seven ordinary riffle shuffles to mix up 52 cards. The analysis leads to a useful understanding of problems in computer science and phylogony.

3. What Do We Know About the Metropolis Algorithm? 4pm - 5pm Wednesday 8th November, Room 515

The Metropolis algorithm is one of the great workhorses of scientific computing. I will explain the algorithm, illustrate its use in breaking codes, finding motifs in proteins and physics. Analysis of the algorithm lies mostly in the future. There are fascinating connections with symmetric function theory and micro-local analysis.

4. Hit and Run as a Unifying Device 4pm - 5pm Thursday 9th November, Room 515

A bewildering variety of simulation schemes are used by practitioners. Recently, some order has emerged - many schemes are special cases of the same algorithm. Known variously as data augmentation or auxiliary variables, I will explain the geometric version - hit and run. As ever, it leads to mathematics problems at the edge of our understanding.

5. From Characterization to Algorithm 1pm - 2pm Friday 10th November, Room 516

I will explain a new scheme for taking mathematical characterizations, seemingly useless and studied only for their elegance, and turning them into useful algorithms. The algorithms allow study of the structure of large networks, co-occurrence data in ecology, and statisticians' contingency tables. The mathematics involves operations research and algebraic geometry.

For further details please e-mail us at rankin@maths.gla.ac.uk

Professor Robert A. Rankin (1915-2001) was an eminent Scottish mathematician who spent a large part of his career in the Department of Mathematics at the University of Glasgow. These lectures are in celebration of Professor Rankin’s mathematical achievements and influence. They are funded by the Glasgow Mathematical Journal Trust Fund.