Gary Bécigneul joined CLS in the 2016 selection round. At this stage, the Center was a year old and recruiting for the first time “full” doctoral fellows who would test out our doctoral training concept of co-supervision by MPI and ETH scientists, with a home base in the group of the main advisor and an exchange year spent with the co-advisor.
Gary’s Ph.D. within the CLS framework was supervised by Prof. Thomas Hofmann in the Data Analytics Laboratory, Department of Computer Science at ETH Zurich and co-supervised by Prof. Bernhard Schölkopf at the Max Planck Institute for Intelligent Systems.
Gary joined the program with a very strong mathematical background: he had read Part III of the Mathematical Tripos at the University of Cambridge (Queens' College), and prior to that obtained a B.Sc. in pure mathematics from University Pierre & Marie Curie (Paris VI). This led naturally to the goal of understanding deep learning from a mathematical perspective, using mostly differential geometry and group theory, as well as in deriving new algorithms based on this understanding. Gary’s work in these areas led to his Ph.D. dissertation entitled “On the Geometry of Data Representations”, which he defended successfully on September 29, 2020.
While being based mostly in the Data Analytics Laboratory in Zurich, Gary spent time in the Tübingen department of Prof. Schölkopf and also undertook a 6-month internship at MIT in Cambridge, USA.
CLS wishes Gary every success in the future.