Mathematics of Machine Learning

The last decade has seen a huge uptake of the applications of machine learning methods, yet, the mechanisms behind the success of such methods is far from well understood. We will be running three events on the Mathematics of Machine Learning in the summer of 2020: an ICMS workshop at the University of Bath in in the week 27-31 July, followed by an LMS Symposium (3-7 August 2020), which in turn is followed by a Research Incubator in the week starting 10 August. The origin of this activity is that the LMS symposia, run at Durham since 1974, are moving to in 2020 to Bath.

ICMS Workshop on Analytic and Geometric Approaches to Machine Learning (27-31 July 2020)

This is the first ever ICMS workshop outside Scotland. The aim of the workshop is to bring together researchers that apply mathematical methodology to machine learning. The emphasis will be on how mathematical theory can inform applications and vice versa. On the theoretical side the event focuses on large data limits, approximation theory and probabilistic modelling. The theory will be connected to applications that either make use of these results or have the potential to do so. Topics include the geometry of data, total variation functionals, Gaussian process regression and neural networks. This is largely an event by invitation only, but the organisers would like to actively encourage locals to attend. Participation of students is particularly welcome.

Organisers are Patricia Vitoria Carrera, Bamdad Hosseini and Matthew Thorpe.

For more information, see link.

LMS-Bath Symposium on the Mathematics of Machine Learning (3-7 August 2020)

The aim of the Symposium is to bring together world leading mathematicians, theoretical physicists and statisticians who are working on the mathematics of machine learning, to promote dialogue between different groups and to identify new directions of research.

Throughout the week, there will be talks by 30 invited established and early career researchers who are currently at the forefront of theoretical and applications-based research in this field. They will deliver talks on a wide range of machine learning related topics. We also plan to organise a public lecture (speaker to be confirmed) and expert panel sessions. There is also an opportunity to contribute to the symposium by presenting a poster. A selection of these posters will be accompanied by spotlight talks. A few SAMBa and SAMBa-aligned students have already expressed interest in presenting a poster.

There are a number of ways how students and other members of the department can get involved. First, as mentioned before, everyone can apply for a poster presentation (deadline 27 March). This is a great way of starting a detailed discussion with experts in the field. Moreover, there will be space for local participants to attend the invited talks and ask questions. There will be plenty of opportunity to discuss and meet these researchers during the coffee breaks/lunch breaks/conference dinner.

Organisers are Philip Aston, Matthias Ehrhardt, Catherine Higham and Clarice Poon.

For more information, see link.

EPSRC research incubator (10-14 August 2020)

This event is similar in style to a SAMBa ITT but with the focus being on the Mathematics of Machine Learning. This mix of a summer school and brainstorming event will bring excellent early career and established researchers in the field to Bath to share their expertise with upcoming talent. Two days of lectures are followed by three days where the participants work on problems set by researchers and users of machine learning. There will be an open call (to be announced March/April) for participation, which is open to everyone, including SAMBa students.

Organisers are Neill Campbell and Matthias Ehrhardt.