Corcoran Memorial Prize Award and Lecture 2021
| 3.00 pm | Welcome from Professor Christl Donnelly, Deputy Head of Department |
| 3.05 pm | Introduction of Corcoran Lecture speaker by Professor Yee Whye Teh |
| 3.10 pm |
Corcoran Memorial Lecture Speaker: David Silver, UCL & Google DeepMind Title: The Pursuit of Reward: from Zero to Superhuman. Abstract: Could the objective of maximising reward be enough to drive intelligent behaviour? Reward-maximising systems have achieved remarkable success in several challenging problems for artificial intelligence, by combining reinforcement learning with deep neural networks. In this talk I describe the ideas and algorithms that led to AlphaGo: the first program to defeat a human champion in the game of Go; AlphaZero: which learned, from scratch, to also defeat the world computer champions in chess and shogi; and MuZero: which achieved superhuman performance without any knowledge of the rules. |
| 4.10 pm | Award of Corcoran Memorial Prize by Christl Donnelly |
| 4.20 pm |
Short talk by 2020 prize winner Speaker: Chris J. Maddison, University of Toronto Title: Lossy Compression for Lossless Prediction Abstract: Most data is automatically collected and only ever "seen" by algorithms. For example, the number of images in many sky surveys is so large that visual inspection is impractical. Despite this, classical lossy data compression methods are designed to store the information needed to guarantee perceptual fidelity rather than just the information needed by algorithms performing downstream tasks. So, we are likely storing vast amounts of unneeded information. In this talk, we characterize the minimum bit-rates required to ensure high performance on all predictive tasks that are invariant under a set of transformations. Based on our theory, we design unsupervised objectives for designing deep-learning-based compression methods, and we show that it is possible to achieve dramatic rate savings on standard image classification datasets. |
| 4.45 pm | Finish |
The Corcoran Memorial Lecture is an annual event with a Corcoran Memorial Prize awarded biennially to students of the Department of Statistics for outstanding graduate work.
The Corcoran Memorial Lectures
The Corcoran Memorial lectures are named in memory of Stephen Corcoran who was a graduate student in the Department of Statistics until his death in 1996. Stephen was a student of Wadham College, Oxford and graduated First Class Honours in Mathematics in 1991. He subsequently gained a Diploma in Mathematical Statistics from Cambridge University before returning to Oxford to study for a D.Phil in Statistics.
Stephen's research was in the field of empirical likelihood. He made substantial progress in this work but sadly his thesis remained unfinished at the time of his death from cancer. Part of Stephen's uncompleted thesis was edited by Professor A. C. Davison and published in Biometrika (1998, pages 967-972).
A family bequest has established an annual lecture in honour of Stephen in which distinguished guest lecturers are invited to deliver a lecture on important aspects of their work. In addition, the Corcoran Memorial Prize is awarded every two years to students of the Department of Statistics for outstanding graduate work. The prizewinners are also invited to give a lecture.
For full details of all prize award winners and lectures see here.