Scientific computation using machine-learning algorithms

Recent mathematical advances and applications


Workshop date

25-26 April 2019


Teaching And Learning Building, Room C14,
University of Nottingham,
University Park Campus


Keynote speakers

  • Martin Eigel
    (WIAS, Germany)
    "A Statistical Learning Approach for Parametric PDEs"
  • Ahmed Elsheikh
    (Heriot Watt University, UK)
    "Machine Learning Approaches for Uncertainty Quantification of Subsurface Flow Models"
  • Jan Hesthaven
    (EPFL, Switzerland)
    "On the Use of Machine Learning in Computational Science and Engineering"
  • Desmond Higham
    (University of Strathclyde, UK)
    "Numerical Precision in Deep Learning"
  • Jakub Marecek
    (IBM Dublin, Ireland)
    "Scaling Up Deep Learning for PDE-based Models"
  • Kaj Nyström
    (Uppsala University, Sweden)
    "Some Thoughts on Neural Networks, PDEs and Data-Driven Discovery"
  • Philipp Petersen
    (University of Oxford, UK)
    "Challenges and Opportunities for Numerical Solvers of PDEs Based on Deep Neural Networks"

Contributed talks

  • Markus Geveler
    (TU Dortmund, Germany)
    "Machine Learning Approaches for the Acceleration of the Linear Solver in PDE Simulations"
  • Jim Magiera
    (University of Stuttgart, Germany)
    "Constraint-Aware Neural Networks for Riemann Problems"
  • Yufei Zhang
    (University of Oxford, UK)
    "Rectified Deep Neural Networks Overcome the Curse of Dimensionality forNonsmooth Value Functions in Zero-Sum Games of Nonlinear Stiff Systems"

Workshop programme

Thursday 25 April

(Teaching and Learning Building, Room C14) 
9:15 Registration open
9:50 Welcome 
10:00 Jan Hesthaven 
11:00 Philipp Petersen
12:00 -Lunch-
13:30 Jacub Marecek
14:30 Kaj Nyström
15:30 -Break-
16:10 Markus Geveler
16:35 Jim Magiera 
17:00 Yufei Zhang

Friday 26 April

(Teaching and Learning Building, Room C14)
9:00 Martin Eigel
10:00 Ahmed Elsheikh
11:00 -Break-
11:40 Desmond Higham
12:40 -Panel discussion-
13:30 Closing Remarks and Lunch

Workshop objective

The aim of this workshop is to discuss recent advances in mathematical foundations of machine learning and artificial neural networks, and their application in computational science.

The workshop will bring together experts in such diverse areas as approximation theory, PDEs, scientific computing, networks, computer science, and uncertainty quantification to facilitate the exchange of results and ideas as well as to initiate new collaborations.

The workshop is spread over two days and contains 7 keynote lectures, contributed talks, and a panel discussion. Registration includes refreshments and lunch on both days.

Workshop essentials

Online Registration





The University of Nottingham
University Park

telephone: +44 (0) 115 951 5151