Spring School on Mathematics of Data Science


School: 23 – 27 September 2019
Pre-course 16 – 20 September 2019


AIMS South Africa, Muizenberg, Cape Town

Data science has become very popular in the last decade due to spectacular successes in many signal processing and machine learning tasks. Mathematics plays a key role for the understanding and development of corresponding computational methods. A particular striking sub-area is compressive sensing which predicts that many types of signals can be recovered from seemingly incomplete data. This discovery has led to huge research activities on many applications and has triggered new developments in mathematics including optimization, approximation theory, statistics and high-dimensional probability theory.

Deep learning outperforms traditional methods in image classification, computer vision, natural language processing, and even in playing games. However, mathematical theory supporting the mostly empirical work in data science is largely missing and mathematicians are just starting to consider this challenging problem of major practical importance. The school aims to introduce participants to some the underpinning mathematical foundations (mainly methods developed in, and/or motivated by, compressed sensing) of data science. By the end of the school participants would be sufficiently armed with the mathematical tools and techniques to potentially undertake theoretical research in data science.

Target Audience:

PhD and Postdoctoral fellows


Bubacarr Bah (AIMS South Africa and Stellenbosch University)
Holger Rauhut (RWTH Aachen University)


Stephen Becker (University of Colorado, Boulder)
Massimo Fornasier (Technical University of Munich)
Felix Krahmer (Technical University of Munich)
Nathan Srebro (University of Chicago – to be confirmed)
Clayton Webster (University of Tennessee, and Oak Ridge National Laboratory)


Regisration form

Closing date for applications: 30 June 2019

Administration and logistics: data-workshops@aims.ac.za

Please note: Travel funding is limited and applicants are required to seek alternative sources of funding