Research Associates: Mathematics of Deep Learning
Department of Applied Mathematics and Theoretical Physics
Up to two positions exist for Research Associates to work on the EPSRC Programme Grant EP/V026259/1 on the "Mathematics of Deep Learning". Math4DL is a joint programme run by the University of Cambridge, University of Bath and University College London. The successful candidate(s) will join a close-knit team hosted by all three universities and led by Profs Carola-Bibiane Schönlieb and Richard Nickl (Cambridge), Prof Chris Budd OBE and Dr Matthias Ehrhardt (Bath) and Profs Simon Arridge and Bangti Jin (UCL). The successful candidate(s) will be based in Cambridge.
This is an exceptional opportunity to conduct ambitious research at the forefront of mathematics, statistics and machine learning. There are generous funds available for conference attendance, travel, computer equipment, training and career development.
The Vision and Ambition of Maths4DL is to develop a comprehensive mathematical, statistical and computational framework that addresses i) the foundational challenges on interpretability of the trained Deep Neural Networks (DNN), ii) robustness and generalisation of DNN on test data, iii) statistical confidence in DNN-outputs and iv) computational challenges with large-scale models and high-dimensional training data.
The successful candidate(s) will become part of the Math4DL team who are leading the expertise in deep learning, interdisciplinary mathematics, statistics, Artificial Intelligence (AI), Partial Differential Equations (PDE's), optimal transport, optimisation, environmental modelling, numerical analysis, geometric integration, scientific computing, image processing, probability, and inverse problems.