From deep network mysteries to physics

The Royal Society Milner Award and Lecture 2023 given by Professor Stéphane Mallat.
 
The remarkable performances of deep neural network remains a mathematical mystery. How come similar network architectures can capture properties of data as different as languages, natural images or physical fields? Physics provides a rich framework to understand this mystery. This presentation bridges neural network and physics, by showing in what sense they rely on similar mathematical principles. It introduces models of deep neural networks and complex physical fields, by separating phenomena appearing at different scales. Interactions of structures across scales are learned with random weights and wavelets. These models are applied to image classification, as well as generation of fluid turbulence and cosmological fields.
 
The Award
 
The Royal Society Milner Award and Lecture, supported by Microsoft Research, is the premier European award for outstanding achievement in computer science. It is awarded to candidates at the peak of their career who have made a substantial contribution to computer science in Europe, with the strategic aim of supporting European researchers and institutes. The recipient is a European researcher or researcher who has been resident in Europe for 12 months or more, and is chosen by the Council of the Royal Society on the recommendation of the Milner Award Committee. The Committee is made up of Fellows of the Royal Society, Members of the Académie des sciences (France) and Members of Leopoldina (Germany). The award is named in honour of Professor Robin Milner FRS (1934-2010), a pioneer in computer science. The medal is of bronze, is awarded annually and is accompanied by a gift of £5,000.